Microsoft Word - INDEX EUR 21590 EN2 report.doc

Institute for Health and Consumer Protection

Physical and Chemical Exposure Unit

I-21020 Ispra (VA), Italy

Final Report

The INDEX project

Critical Appraisal of the Setting and Implementation of Indoor

Exposure Limits in the EU

Dimitrios Kotzias, Kimmo Koistinen, Stylianos Kephalopoulos,

Christian Schlitt, Paolo Carrer, Marco Maroni,

Matti Jantunen, Christian Cochet, Séverine Kirchner,

Thomas Lindvall, James McLaughlin, Lars Mølhave,

Eduardo de Oliveira Fernandes and Bernd Seifert

2005 EUR 21590 EN

European Commission

Directorate-General Joint Research Centre

http://www.jrc.cec.eu.int

Legal Notice

Neither the European Commission nor any person acting on behalf of the

Commission is responsible for the use which might be made of this publication.

A great deal of additional information on the European Union is available on the Internet.

It can be accessed through the Europa server

http://europa.eu

EUR 21590EN

ISSN 1018-5593

Luxembourg: Office for Official Publications of the European Communities

Reproduction is authorised provided the source is acknowledged

Printed in Italy

The INDEX project Final report

Legal notice

Neither the European Commission nor any person acting on behalf of the Commission is responsible for the use which

might be made of the following information.

Foreword

In the past decades a large number of studies have indicated the presence of different compounds belonging to a variety

of chemical classes in indoor environments (buildings, homes). The presence of these chemicals in indoor air is the

result of infiltration of polluted outdoor air and of emissions from various indoor sources, including building materials,

activities of the occupants, consumer products, smoking etc.

For many of these chemicals, the risk to human health and comfort is almost totally unknown and difficult to predict

because of lack of toxicological data and information on the dose-response characteristics in humans or animal models.

On the other hand, a full toxicological testing as requested by the “existing chemicals” legislation is difficult to

accomplish for these compounds, because it would involve investigation of acute and subacute toxicity, mutagenicity,

carcinogenicity and reproductive toxicity according to testing protocols that are complex, time-consuming and

expensive. Moreover, the EU policy on limitation of unnecessary animal testing further limits the possibility of

advocating a generalized animal testing of these chemicals.

The result of this situation is that there is an objective difficulty in regulating the presence of these substances in indoor

air principally because of the absence of adequate hazard and risk assessment. There is therefore an urgent need to

develop a strategy for the identification of priorities in testing, assessment and regulation.

In the frame of the INDEX project the existing knowledge worldwide has been assessed on

- type and levels of chemicals in indoor air and

- available toxicological information to allow the assessment of risk to health and comfort.

The collection and evaluation of the aforementioned information within the frame of the INDEX project shall contribute

to develop a strategy for prioritization in assessment and regulation of chemicals in indoor environments.

The Steering Committee

The INDEX project Final report

Organisation responsible of the project: European Commission, Joint Research Centre, Institute for Health and

Consumer Protection, Physical and Chemical Exposure Unit, Ispra, Italy (JRC/IHCP/PCE).

Project leader: Dr. Dimtrios Kotzias, JRC/IHCP/PCE

Sources of funding: DG Consumer Protection (SANCO) and JRC/IHCP/PCE

Sub-contractors: University of Milan, Department of Occupational Health, Unit Ospedale Luigi Sacco, Milan, Italy

National Public Health Institute (KTL), Kuopio, Finland

Preparation of documents: All working drafts and this final report were prepared by the Joint Research Centre

(Institute for Health and Consumer Protection, Physical and Chemical Exposure Unit, JRC), the University of Milan

(Department of Occupational Health, Unit Ospedale Luigi Sacco, UNIMI), the National Public Health Institute

(Department of Environmental Hygiene, Kuopio, Finland, KTL) and with the individual contributions and advice in the

course of the work from members of the Steering Committee.

Final drafting:

Hazard Identification / Exposure Assessment:

Dr. Kimmo Koistinen and Dr. Dimitrios Kotzias (JRC)

Dose/Response Assessment / Risk Characterization:

Christian Schlitt, Dr. Paolo Carrer and Prof. Marco Maroni (UNIMI)

Risk Management:

Prof. Matti J. Jantunen (KTL)

Date of last literature search: September 2004

Steering Committee Members:

Dr. Christian Cochet Centre Scientifique et Technique du Bâtiment, Division Santé–Bâtiment,

Marne la Vallée, France

Prof. Matti J. Jantunen KTL, Department of Environmental Hygiene, Kuopio, Finland

Dr. Stylianos Kephalopoulos European Commission, Joint Research Centre, Institute for Health and

Consumer Protection, Physical and Chemical Exposure Unit, Ispra, Italy

Dr. Séverine Kirchner Centre Scientifique et Technique du Bâtiment, Division Santé–Bâtiment,

Marne la Vallée, France

Prof. Thomas Lindvall Karolinska Institute, Institute of Environmental Medicine, Stockholm,

Sweden

Prof. Marco Maroni Università di Milano, Department of Occupational Health, Unit Ospedale

Luigi Sacco, Milan, Italy

Dr. James P. McLaughlin University College Dublin, Department of Physics, Dublin, Ireland

Prof. Lars Mølhave Aarhus Universitet, Institut for Miljø– og Arbejdsmedicin, Århus, Denmark

Prof. Eduardo de Oliveira Fernandes Universidade do Porto, Departamento de Engenharia Mecânica, Faculdade

de Engenharia, Porto, Portugal

Prof. Bernd Seifert Umweltbundesamt, Abteilung Umwelthygiene, Berlin, Germany

The INDEX project Final report

Contents

Foreword ........................................................................................................................................................................................................................... 2

Contents ........................................................................................................................................................................................................................... 4

Executive Summary ....................................................................................................................................................... 9

Introduction and objectives.......................................................................................................................................................................................... 9

Methodology................................................................................................................................................................................................................ 9

Risk assessment of the selected compounds ............................................................................................................................................................. 10

Risk management tools.............................................................................................................................................................................................. 14

Recommendations and management options ............................................................................................................................................................ 14

1. Introduction .............................................................................................................................................................. 17

2. Objectives ................................................................................................................................................................ 18

3. Methodology ............................................................................................................................................................. 19

3.1 Risk Assessment ....................................................................................................................................................................................................... 19

3.1.1 Selection of Indoor Air Chemicals for Consideration (Hazard Identification) ............................................................................................. 19

3.1.2 Exposure Assessment ..................................................................................................................................................................................... 22

3.1.3 Dose/Response Assessment............................................................................................................................................................................ 22

3.1.4 Risk Characterization and prioritisation of chemicals ................................................................................................................................... 26

3.2 Risk Management ..................................................................................................................................................................................................... 27

4. Risk Assessment of the selected chemicals ............................................................................................................. 29

Formaldehyde............................................................................................................................................................... 30

1. Compound Identification............................................................................................................................................................................................ 30

2. Physical and Chemical Properties .............................................................................................................................................................................. 30

3. Indoor Air Exposure Assessment ............................................................................................................................................................................... 30

Emission sources........................................................................................................................................................................................................ 30

Indoor air and exposure concentrations..................................................................................................................................................................... 31

4. Toxicokinetics............................................................................................................................................................................................................. 35

Absorption.................................................................................................................................................................................................................. 35

Distribution ................................................................................................................................................................................................................ 35

Metabolism and elimination ...................................................................................................................................................................................... 35

5. Health effects .............................................................................................................................................................................................................. 35

Effects of short-term exposure................................................................................................................................................................................... 35

Effects of long-term exposure (noncancer) ............................................................................................................................................................... 39

Carcinogenic effects .................................................................................................................................................................................................. 43

Interactions with other chemicals .............................................................................................................................................................................. 46

Odour perception ....................................................................................................................................................................................................... 47

Summary of Formaldehyde Dose Response Assessment ......................................................................................................................................... 47

6. Risk Characterization.................................................................................................................................................................................................. 49

Health hazard evaluation ........................................................................................................................................................................................... 49

Percentage of population exposed beyond given threshold levels............................................................................................................................ 49

Cancer risk evaluation ............................................................................................................................................................................................... 49

Comments .................................................................................................................................................................................................................. 49

Result ......................................................................................................................................................................................................................... 50

Carbon monoxide ......................................................................................................................................................... 51

1. Compound identification ............................................................................................................................................................................................ 51

2. Physical and Chemical properties............................................................................................................................................................................... 52

3. Indoor Air Exposure assessment ................................................................................................................................................................................ 52

Endogenous sources of carbon monoxide:................................................................................................................................................................ 59

Elimination................................................................................................................................................................................................................. 61

Effects of long-term exposure ................................................................................................................................................................................... 68

Carcinogenic and genotoxic effects........................................................................................................................................................................... 70

Summary of Carbon Monoxide Dose Response Assessment ................................................................................................................................... 70

Percentage of population exposed beyond given threshold CO levels ..................................................................................................................... 72

Nitrogen dioxide ........................................................................................................................................................... 75

The INDEX project Final report

Carcinogenic and mutagenic effects.......................................................................................................................................................................... 87

Interaction with other chemicals................................................................................................................................................................................ 87

Summary of Nitrogen dioxide Dose Response Assessment ..................................................................................................................................... 89

6. Risk characterization................................................................................................................................................................................................... 90

Health hazard evaluation of short- and long-term exposure ..................................................................................................................................... 90

Exposure data of relevance for the European population ......................................................................................................................................... 90

Benzene ................................................................................................................................................................ 93

Genotoxic effects ..................................................................................................................................................................................................... 107

Susceptible population............................................................................................................................................................................................. 112

Summary of Benzene Dose Response Assessment................................................................................................................................................. 114

Population cancer risk estimates of benzene-induced leukemia ............................................................................................................................. 116

Health hazard evaluation (noncancer) ..................................................................................................................................................................... 118

Naphthalene .............................................................................................................................................................. 121

Genotoxicity............................................................................................................................................................................................................. 133

Carcinogenic potential ............................................................................................................................................................................................. 134

Summary of Naphthalene Dose Response Assessment .......................................................................................................................................... 135

Health hazard evaluation of short-term exposure.................................................................................................................................................... 136

Health hazard and cancer risk evaluation of long-term exposure........................................................................................................................... 136

Relevance of EU-population exposure to naphthalene ........................................................................................................................................... 137

Acetaldehyde .............................................................................................................................................................. 138

Contribution of inhalation exposure to total exposure............................................................................................................................................ 138

The INDEX project Final report

Summary of Acetaldehyde Dose Response Assessment......................................................................................................................................... 151

Cancer risk and health hazard evaluation................................................................................................................................................................ 152

Percentage of population exposed beyond given threshold levels.......................................................................................................................... 152

Toluene .............................................................................................................................................................. 154

Summary of Toluene Dose Response Assessment.................................................................................................................................................. 170

Health hazard evaluation of short- and long-term exposure ................................................................................................................................... 171

Relevance for the EU-population exposure............................................................................................................................................................. 172

Xylenes (ortho-, meta- and para-) ..............................................................................................................................173

Differences among individual xylene isomers ........................................................................................................................................................ 181

Carcinogenic potential and genotoxicity................................................................................................................................................................. 188

Summary of Xylenes Dose Response Assessment.................................................................................................................................................. 190

Health hazard evaluation of short- and long-term exposure ................................................................................................................................... 191

Relevance of EU-population exposure to xylenes .................................................................................................................................................. 191

7. Aromatic compounds: Comparison of relevant data................................................................................................................................................ 193

The reciprocal calculation approach for mixtures................................................................................................................................................... 193

Styrene .............................................................................................................................................................. 195

Macromolecular adducts and genetic toxicity......................................................................................................................................................... 207

Summary of Styrene Dose Response Assessment .................................................................................................................................................. 211

Health hazard and cancer risk evaluation of short- and long-term exposure.......................................................................................................... 212

Relevance of EU-population exposure to styrene ................................................................................................................................................... 213

The INDEX project Final report

Ammonia .............................................................................................................................................................. 214

Reproductive, mutagenic and carcinogenic effects................................................................................................................................................. 224

Summary of Ammonia Dose Response Assessment............................................................................................................................................... 225

Health hazard evaluation of short-term exposure.................................................................................................................................................... 226

Health hazard evaluation of long-term exposure..................................................................................................................................................... 226

Relevance of EU-population exposure to Ammonia............................................................................................................................................... 226

Limonene .............................................................................................................................................................. 227

Contribution of inhalation exposure to total exposure............................................................................................................................................ 227

Terpene/ozone reaction products........................................................................................................................................................................... 228

Summary of Limonene Dose Response Assessment .............................................................................................................................................. 234

Health hazard evaluation of long-term exposure..................................................................................................................................................... 235

Relevance of EU-population exposure to limonene................................................................................................................................................ 235

alpha-Pinene .............................................................................................................................................................. 236

Carcinogenic and genotoxic potential ..................................................................................................................................................................... 240

Summary of alpha-Pinene Dose Response Assessment.......................................................................................................................................... 241

Health hazard evaluation of long-term exposure..................................................................................................................................................... 242

Relevance of EU-population exposure to α-Pinene ................................................................................................................................................ 242

5. Risk management tools .......................................................................................................................................... 243

6. Recommendations and management options....................................................................................................... 246

6.1 High priority chemicals .......................................................................................................................................................................................... 246

6.2 Second priority chemicals....................................................................................................................................................................................... 247

6.3 Chemicals requiring further research with regard to human exposue or dose response........................................................................................ 248

The INDEX project Final report

Annex 1. Phase 1: Exposure and dose-response data for indoor air pollutants collected in the literature review...................................................... 249

Annex 2. Phase 2: The selection of compounds to the further analysis....................................................................................................................... 258

Annex 3. Phase 3: Compounds selected into the detailed risk assessment.................................................................................................................. 265

Annex 4. The EXPOLIS study: parameters describing the indoor air concentration distributions plotted to the graphs of this report..................... 270

Annex 5. The National Survey of air pollutants in English homes: parameters describing the indoor air concentration distributions

plotted to the graphs of this report................................................................................................................................................................ 271

Annex 6. The French National Survey on Indoor Air Quality (preliminary results from the ongoing project): parameters describing the indoor

air concentration distributions plotted to the graphs of this report (Golliot et al 2003, Kirchner 2004).................................................... 272

Annex 7. Relevant National and International Guidelines and Recommendations .................................................................................................... 273

Bibliographic references.................................................................................................................................................................................. 275

References in introductory chapters (chapters 1 to 4).................................................................................................................................................. 275

References by compounds (chapter 4).......................................................................................................................................................................... 276

Formaldehyde .......................................................................................................................................................................................................... 276

Carbon monoxide..................................................................................................................................................................................................... 282

Nitrogen dioxide ...................................................................................................................................................................................................... 286

Benzene.................................................................................................................................................................................................................... 291

Naphthalene ............................................................................................................................................................................................................. 294

Acetaldehyde............................................................................................................................................................................................................ 298

Toluene..................................................................................................................................................................................................................... 303

Xylenes..................................................................................................................................................................................................................... 309

Styrene ..................................................................................................................................................................................................................... 315

Ammonia.................................................................................................................................................................................................................. 321

Limonene ................................................................................................................................................................................................................. 325

alpha-Pinene............................................................................................................................................................................................................. 328

References in Annex 1, 2 and 3 .................................................................................................................................................................................... 330

The INDEX project Final report

Executive Summary

Introduction and objectives

The INDEX project (Critical Appraisal of the Setting and Implementation of Indoor exposure Limits in the EU) started

in December 2002 and had a duration of two years, until December 2004.The project was financially supported by DG

SANCO and it was coordinated and carried out by the JRC in collaboration with a Steering Committee of leading

European experts in the area of indoor air pollution. Scope of INDEX was to identify priorities and to assess the needs

for a Community strategy and action plan in the area of indoor air pollution by:

- setting up a list of compounds to be regulated in indoor environments with priority on the basis of health

impact criteria

- providing suggestions and recommendations on potential exposure limits for these compounds, and

- providing information on links with existing knowledge, ongoing studies, legislation etc. at world scale.

Methodology

The main steps to be followed in the project as they have been defined by the Steering Committee were:

- literature review (step 1)

- setting up criteria to select compounds (step 2)

- review of exposure and dose/response data (step 3)

- risk characterization of the selected compounds (step 4)

- prioritisation of the selected compounds (step 5) and

- recommendations and risk management options on potential exposure limits (step 6)

In the first step the world literature was reviewed to collect background information on exposure concentrations and

dose/response data of the chemical pollutants that have strong indoor sources. In the second step, the Steering

Committee has defined the criteria for the selection of pollutants for further risk analysis. After careful examination of

exposure and toxicological dose response data (step 3), the Steering Committee ended up to a short list of pollutants that

were selected for the risk characterization (step 4). The compounds that probably cause the highest health risk in

European population according to the risk characterization were prioritized (step 5). Finally, recommendations and risk

management options on European exposure limits should be given for a few selected compounds (step 6).

In order to accomplish with the tasks mentioned above the Steering Committee was divided into two working groups,

one for exposure assessment and one for dose/response assessment.

Exposure to selected indoor air pollutants was estimated by reviewing exposure data collected from scientific literature,

from available databases, and by personal communications. The aim of this work was to summarise prevailing indoor

air and personal exposure concentrations of these compounds in Europe and also worldwide. Results from population-

based studies have been preferred to be able to generalise the results from studied individuals to larger populations,

targeting to assess exposures of all the Europeans. Considering the fact that no new data could be generated in this

project, the steering group defined the following criteria to be able to select the chemical compounds to the risk

analyses:

1. Only single compounds will be considered

2. The compound should have common indoor sources, which dominate the exposures of at least significant fraction

of the population

3. The compound should have known health effects.

It was also decided that compounds, which have been regulated by specific guidelines or regulations would be excluded

from theses analyses. For example, radon and tobacco smoke were excluded from the risk assessment process due to

these criteria.

In preparing the dose-response assessment fact sheets of the selected chemicals, information were retrieved from

scientific literature (mainly by electronic search), comprehending toxicological reviews of leading health organizations,

risk evaluation documents and available databases. In addition, Toxline and Medline were searched for relevant

scientific communications published up to September 2004.

Nearly all key-studies referred to in the present assessment, establishing effect levels for appropriate toxicological

endpoints, are those selected by health organizations for the derivation of health based limits of exposure or among risk

The INDEX project Final report

assessment requirements. Although not specifically addressing health hazards and risks associated with indoor air

exposure, i.e. not being designed for the expression of effects at lowermost exposure concentrations, nearly all studies

were aimed at identifying the most sensitive endpoint considered to be of relevance to humans. Where relevant, studies

conducted on susceptible sub-populations (e.g. asthmatics, infants, children, pregnant women etc.) were quoted and

taken into consideration in the risk characterization.

In the final step of the general risk assessment process, the incidence of health hazards and risks in the European

population, associated with indoor exposure to individual chemicals, was estimated. Limits of exposure (ELs, following

short- and long-term exposure) were derived for each chemical after selection of a critical study describing the

appropriate toxicological endpoint and by applying the “no-observed-adverse-effect level (NOAEL) / assessment factor

(AS)” approach. Where no NOAEL observation was documented, a lowest-observed-adverse-effect level (LOAEL) was

taken and an additional assessment factor of 10 used for EL derivation. Only for one compound (benzene) the

characterization was based on population cancer risk estimation. Where supported by scientific evidence, susceptible

subpopulations were accounted for, in particular: asthmatic individuals, infants, children, individuals with heart

diseases, pregnant women, individuals with enzyme deficiencies.

On the basis of the available information and after careful examination of the existing data, the steering committee

finally decided to include into a detailed assessment 14 out of initial 41 candidate compounds (phase 4) i.e.:

acetaldehyde, alpha-pinene, benzene, carbon monoxide, d-limonene, formaldehyde, meta-and para xylene, ortho-

xylene, naphthalene, ammonia, nitrogen dioxide, styrene and toluene.

Risk assessment of the selected compounds

Information from the exposure assessment (Chapters 1-3) and toxicity assessment (Chapters 4-5) were integrated and a

risk characterization (Chapter 6) performed on each chemical. Based on the conclusions of the assessments and on the

completeness of individual databases, a priority ranking was arranged with the 14 chemicals assigned to three groups as

given hereafter.

Group 1: High priority chemicals

Formaldehyde : Because of its high chemical reactivity, formaldehyde is the most important sensory irritant

among the chemicals assessed in the present report. Due to being ubiquitous pollutant in

indoor environments and to the increasing evidence indicating that children may be more

sensitive to formaldehyde respiratory toxicity than adults it is considered a chemical of

concern at levels exceeding 1 µg/m

, a concentration more or less corresponding with the

background level in rural areas. Results from available exposure data, although limited,

confirm that almost the entire population is exposed indoors at levels (Median level±sd: 26±6

µg/m

; 90th±sd: 59±7 µg/m

; N = 6) higher than this background level, here established as

the limit of exposure, with at least 20% of the European population exposed at levels

exceeding the no-observed-effect-level (NOAEL: 30 µg/m

). Within the concentration range

measured, mild irritation of the eyes could be experienced by the general population as well

as the odour perceived above 30 µg/m

Reported formaldehyde concentrations were lower (99

< 150 µg/m

)

than a presumed

threshold for cytotoxic damage to the nasal mucosa (about 1 mg/m

) and hence

considered low enough to avoid any significant risk of upper respiratory tract cancer in

humans. The last statement could be subjected to changes due to the current IARC

revision of the carcinogenicity of formaldehyde.

Carbon monoxide : Available exposure data confirm that Carbon Monoxide (CO) sources in EU-residences are

contributing to short-term rather than to long-term exposures. Personal exposure outcomes

averaged over 1-hour were considered of moderate concern even for the most susceptible

subpopulations. Nevertheless, uncertainties resulting from the predictive capabilities of the

CFK-model* in individuals exposed at low

CO concentrations and its applicability to

sensitive subpopulations, suggest that about 10% of the general non-smoking population

experience CO levels which could be hazardous for individuals with heart diseases. Increased

exposures could be expected for residences in the vicinity of busy city streets.

In addition, there is no evidence that long-term CO exposures in EU residences contribute to

carboxyhaemoglobin levels in blood higher than the baseline levels resulting from

endogenous production in normal, non-smoking individuals.

The INDEX project Final report

On the other hand and in contrast with all other chemicals assessed in the present report,

carbon monoxide causes a considerable number of deaths and acute poisonings in the general

population (with complications and late sequel). Also, individuals suffering from CO

poisoning are often unaware of their exposure because symptoms are similar to those

associated with viral illness or clinical depression. In indoor environments, these health risks

are nearly completely associated with the incorrect use of combustion devices or faulty

unvented gas appliances.

* The physiologically based pharmacokinetic (PBPK) model of Coburn, Forster, and Kane (CFK-model) is a reliable

method for predicting COHb blood levels for exposure to a given ambient carbon monoxide concentration. This

model has been extensively validated over many years. Precision is acceptable, providing that the original conditions

of use are rigorously applied.

Nitrogen dioxide : Reported maximum nitrogen dioxide (NO

) levels associated with the use of gas appliances

in homes (gas cooking and heating) are in the range 180-2500 µg/m

. Exposure at these

levels could generate effects in the pulmonary function of asthmatics, considered to be the

subjects most susceptible to acute NO

exposure, with the lower end of the range

approximating the WHO guideline (200 µg/m

, 1-hour average), established for the

protection of asthmatic individuals and the upper end starting to affect health in normal

individuals.

For long-term exposures, increased respiratory symptoms and lung function decreases in

children were documented to be the most sensitive effect in the general population. Measured

background

levels in European homes indicate that a remarkable portion of the population is

exposed at NO

levels higher than current guideline values protecting from respiratory effects

in children. In up to 25% of the investigated residences (45% in an Italian study) NO

levels

exceeded the German indoor-related guideline value (GV II: 60 µg/m

, 1-week average),

what would have resulted in immediate action i.e. the examination of the situation with

regard to a need for control measures.

On the other hand, safe levels in homes, i.e. < 40 µg/m

(following the WHO recommended

annual value), are not likely to be achievable everywhere (e.g. in areas with intense

automotive traffic) given that ventilation alone may introduce outdoor air containing such

concentrations.

Benzene : Benzene is ubiquitous in the athmosphere, mainly due to anthropogenic sources (90%), with

concentrations in the European continental pristine air ranging from 0.6 to 1.9 µg/m

. It is a

genotoxic carcinogen and hence no safe level of exposure could be recommended. Results

from nine monitoring surveys indicate that the European population is experiencing in their

homes an increased risk in contracting benzene induced leukaemia, with respect to the

estimated background lifetime risk of 7-8 cases per one million people (considering the WHO

unit risk factor). Based on the available exposure data (Median levels±sd: 4.2±3.2 µg/m

;

levels±sd: 11.5±11.1 µg/m

; N = 9) two main scenarios could be described as follows:

− People living in highly trafficked urban areas are expected, on average, to

experience an estimated 6 to 30-fold increase in contracting benzene induced

leukaemia during their life, the benzene levels encountered in these areas not being

expected to produce chronic effects other than cancer, in particular haematological

effects, nor acute sensory effects such as odour perception (odour threshold: 1.2

mg/m

) and sensory irritation. Also, a reduced contribution of specific indoor

sources is likely to be expected, given that ventilation alone may introduce increased

outdoor benzene levels.

− People living in rural areas or poorly trafficked towns were expected, on average, to

experience an estimated 1 to 5-fold increase in contracting benzene induced

leukemia during their life, this factor depending principally on the presence of

indoor sources.

Naphthalene : With regard to the general population a long-term exposure limit has been set at 10 µg/m

according to the assumption that nasal effects observed in mice are consistent with the health

effects reported among exposed workers. Available exposure data indicate that, on average,

the European population is exposed at naphthalene levels 10 times lower than this EL,

although an important exception resulted from a survey held in Athens, were levels

exceeding the EL were measured in nearly all residences. It is assumed that increased

residential exposures originate from the use of naphthalene based moth-repellents, a

The INDEX project Final report

widespread use occurring in certain countries of the Mediterranean area.

An important source of uncertainty in establishing safe exposure limits is the potentially

greater sensitivity of certain subpopulations to naphthalene toxicity, including infants and

neonates, and individuals deficient in glucose-6-phosphate dehydrogenase (G6PD), the

prevalence of this inherited deficiency reported to be 2 to 20% in defined Mediterranean

subpopulations. In these latter cases manifested effects are hemolytic anemia and its sequel.

In relation to carcinogenicity, naphthalene is not genotoxic in vivo and thus tumour

development, observed in rodents, is considered to arise via a non-genotoxic mechanism.

Also, the underlying mechanism for the development of nasal tumours in the rat is considered

to be the chronic inflammatory damage seen at this site. It follows that prevention of local

tissue damage would prevent subsequent development of tumours.

Group 2: Second priority chemicals

Acetaldehyde : The results from only three indoor air monitoring surveys allow a crude estimate of average

acetaldehyde concentrations in European residences. Median concentrations (10-20 µg/m

)

are one order of magnitude lower than the Exposure Limit set here at 200 µg/m

and are

within the same range of concentrations occurring in exhaled breath following its

endogenous production in the general population, not taking into account increases resulting

from the consumption of alcoholic beverages. Considering that exogenous acetaldehyde peak

exposures are mainly associated with tobacco smoke, concentrations in the order of the

Exposure Limit could be expected following intense cigarette consumption.

Assuming that the available exposure data are indicative of the population residential

exposure it is concluded that people in Europe do not experience increased health hazards

associated with acetaldehyde levels in their homes, although additional work should be

warranted for a better characterization of exposure and dose response.

Also, measured indoor levels are lower than a presumed threshold for cytotoxic damage to

the nasal mucosa, and hence considered low enough to avoid any significant risk of upper

respiratory tract cancer in humans.

Toluene : Human effects on the central nervous system are considered as the most sensitive effect in

both short- and long-term inhalatory exposure to toluene. Available exposure data indicate

that the European population is not experiencing health effects of concern resulting from the

exposure to toluene in their homes. Results from ten monitoring surveys show that toluene

levels in the order of the established exposure limit of 300 µg/m

could be reached under

worse-case conditions and in a limited number of urban residences. On average, median

concentrations (90th percentile) were found to be 16 (5) times lower than the EL. Also,

short-term exposures associated with human indoor activities are not expected to exceed the

acute EL set here at 15.000 µg/m

Xylenes : A chronic exposure limit of 200 µg/m

has been derived based on generally mild adverse

effects associated with CNS and increase in the prevalence of eye irritation and sore throat.

The results of eight monitoring surveys indicate that background levels of xylenes in

European residences are of no concern to human health since median (90th percentile) levels

are, on average, 20 (6) times lower than the EL established. Acute exposure data indicate that

it is very unlikely that xylenes emissions associated with human indoor activities would

generate levels in the order of the proposed short-term EL of 20 mg/m

, considered

protective for irritative effects in the general population.

Although human exposure most likely occurs to the mixture of xylene isomers, animal and

human toxicity data suggest that mixed xylenes and the different xylene isomers produce

similar effects.

Styrene : A long-term exposure limit (EL) of 250 µg/m

has been derived based on the assumption that

neurological effects are probably the most sensitive indicator of styrene toxicity. When

examining the results of eight monitoring surveys it can be concluded that background

styrene concentrations in European residences are of no concern to human health since

median levels are, on average, two orders of magnitude below the established EL. Although

no acute exposure data were available, it is unlikely that styrene emissions associated with

human indoor activities would generate levels up to the proposed short-term EL of 2000

The INDEX project Final report

µg/m

, considered protective for irritative effects in asthmatics.

Although genotoxic effects in humans have been observed at relatively low concentrations,

they were not considered as critical endpoints for the derivation of the exposure limit, in

view of the equivocal evidence for the carcinogenicity of styrene in humans (WHO).

Group 3: Chemicals requiring further research with regard to human exposure or dose response

Ammonia : There is a lack of knowledge concerning indoor concentrations and exposures of ammonia.

Exposure data are limited on only one monitoring survey describing concentrations of

ammonia in Finnish homes with and without known indoor air quality (IAQ) problems. In

both cases measured concentrations were within the same order of magnitude with both

exposure limits here established for short- and long-term effects (70 and 100 µg/m

respectively), relating on irritative effects and pulmonary functions and taking into account

the particular susceptibility of asthmatic subjects. It is assumed that exposure concentrations

in the order of the short-term EL could easily be attained during domestic activities making

use of ammonia containing household products.

Limonene : An attempt has been done in deriving an exposure limit (EL) for long-term effects associated

with limonene exposure by referring to a study on volunteers exposed at sub-acute (2 hours)

inhalation doses. When comparing this EL (450 µg/m

) with the results from seven indoor

surveys it is concluded that no neurological effects would be expected at background

limonene levels encountered in European homes, with median (90

percentile) levels at least

10 (3) times lower than the proposed EL. It is assumed that at 10-fold the level set as the EL,

health effects could be expected following acute exposure. Due to its widespread use as a

flavouring agent in numerous consumer products, short-term exposures at levels in the order

of some mg/m

could not be excluded, although significative exposure data are lacking.

An exacerbation of effects (not better defined) could be expected following the concomitant

presence in residences of ozone emitting sources, due to the formation of irritant reaction

products.

α-Pinene : An attempt has been done in deriving an exposure limit (EL) for long-term effects associated

with α-pinene exposure by referring to a study on volunteers exposed at sub-acute (2 hours)

inhalation doses. When comparing this EL (450 µg/m

) with the results from six indoor

surveys it is concluded that no irritative effects to the eyes, nose and throat would be

expected at background α-pinene levels encountered in European homes, with median (90

percentile) levels at least 40 (10) times lower than the proposed EL. It is assumed that at 10-

fold the level set as the EL, health effects could be expected following acute exposure. Due

to its widespread use as a flavouring agent in numerous consumer products, short-term

exposures at levels in the order of some mg/m

could not be excluded, although significative

exposure data are lacking.

An exacerbation of effects (not better defined) could be expected following the concomitant

presence in residences of ozone emitting sources, due to the formation of irritant reaction

products.

The INDEX project Final report

Risk management tools

The indoor air risk management tools include:

IAQ Standards and guidelines: The huge quantity and privacy of indoor spaces makes enforceable indoor air quality

standards impossible. Guideline values for key pollutants, however, are useful when matching potential sources,

occupant needs and ventilation rates in the design of new building or renovation of an existing one. Sometimes they can

also be applied in judging the acceptability of indoor air quality in a building.

Building codes and ventilation standards apply to every new building, and can therefore be powerful means of

preventing or controlling indoor sources and ensuring sufficient contaminant dilution.

Building and household equipment standards and permits cover new installations, and can therefore ensure that

improper - hazardous or avoidable - emissions are avoided or properly controlled, e.g. exhausted.

Mandatory maintenance and inspections of, e.g., gas appliances or fireplaces & chimneys, help avoid gradual

deterioration of the equipment, and accumulation of its risks, and ensure that safe operation requirements are met

through their lifetime.

Limits, labelling and reporting of the contents of or releases from building products, furnishing materials,

equipment and consumer products can be quite inflexible and heavy instruments when mandated by law, but flexible

and effective on voluntary basis, as long as commonly agreed (between consumers, builders, industry and public health

authorities), widely published and generally understood criteria are used. Clear product labels guide informed selections

of both builders and consumers.

Public awareness raising and information is the key to safe indoor environments, because, due to their sheer

numbers, a great majority of potential indoor pollution risks must be identified, assessed and managed by occupants

themselves. They are best helped by widespread and general awareness raising information, and easily found and

understood sources (leaflets, internet sites) for more specific technical information.

Recommendations and management options

The recommendations and management options proposed protect the general population and most individuals most of

the time, but they will not prevent all harm to every individual. Besides, the proposed measures will not prevent health

damage from outdoor air pollutants which penetrate indoors. Following general recommendations apply to many indoor

air contaminants:

• Restrict tobacco smoking in all indoor spaces.

• Restrict the construction of attached garages, or isolate them from living and working spaces.

• Ensure that ventilation dilutes predictable indoor emissions below the guideline levels.

• Raise public awareness about indoor air risks.

High priority chemicals

The high priority compound list consists of 5 chemicals, with potential of high indoor concentrations, uncontested

health impacts, and effective risk management.

Formaldehyde: The non-carcinogenic no-effect level is 30 µg/m

. Pending on IARC revision of formaldehyde

carcinogenecity, a guideline should be as low as reasonably achievable. Management options are to restrict and avoid

the use of formaldehyde containing materials and products in buildings:

Nitrogen Dioxide: Proposed guideline values are 40 µg/m

(1-week) and 200 µg/m

(1-hour). Management options are

to connect each indoor combustion device/appliance to chimney or vented hood, and to ensure sufficient local extract

ventilation in kitchens with gas stove.

Carbon Monoxide: Proposed guideline values are 10 mg/m

(8-hour) and 30 mg/m

(1-hour). Management options are

to connect each combustion equipment/appliance to chimney or vented hood, to ensure sufficient local extract

ventilation in kitchens with gas stove, mandatory inspection and maintenance of indoor combustion devices, and CO

alarms.

Benzene: Benzene is a carcinogen, its indoor air concentration should be kept as low as reasonably achievable, and not

exceed outdoor concentrations. Management options are to ban benzene sources indoors, and lower the permissible

benzene content in any building material and consumer product.

The INDEX project Final report

Naphtalene: Proposed long term guideline value is 10 µg/m

. Management option is to restrict the use of naphthalene

containing household products.

Second priority chemicals,

which are not considered to urgently require regulatory risk management actions

specifically in indoor air are Acetaldehyde, o-, p- and m-Xylene, Toluene and Styrene.

Additional chemicals of interest

for indoor air risk management, which are considered to require further research with

regard to human exposure or dose response before recommendations can be made, are Ammonia, delta-Limonene, and

alpha-Pinene.

The INDEX project Final report

1. Introduction

Human exposure to air pollution occurs over 90% indoors, but it depends on both indoor and outdoor air pollution.

Outdoor air pollution is important mainly because indoor air is linked to outdoor air via ventilation. The strength of the

personal exposure-outdoor air association varies considerably between the individuals, their activities,

microenvironments and pollutants. On one hand leisure time exposures of active individuals to various pollutants are

mainly associated with outdoor air pollution levels. On the other hand workday exposures, exposures of individuals

leading sedentary lifestyles in closed or air conditioned spaces and exposures to pollutants like formaldehyde and radon

are essentially independent of outdoor air pollution levels.

Past European Air Quality Directives have not taken population exposure sufficiently into account. They have up to

now been mainly oriented on outdoor air pollution. EU legilslation addresses inter alia air quality standards, national

emission ceilings and emissions from vehicles and industries. However, the Sixt Environmental Action Plan and the

new launched Environmental and Health Strategy are oriented towards the impact of environmental risk factors on

human health, and DG SANCO and other relevant DGs are developing proposals for a public health policy. So, in

addition to ambient air pollution, the pollution inside confined environments as well as the extent of personal mobility

and specific activities all play a significant role in exposure to air pollutants. The relative importance of these pollutants

varies greatly depending on the sources, pollutants, and individuals or populations of concern. Concentrations of e.g.

volatile organic compounds (VOC), in general, are higher indoors than outdoors, yet for some VOCs outdoor air levels

may significantly affect respective indoor levels. For many chemicals occurring indoors the risk for human health and

comfort is almost unknown and difficult to predict, in particular, the risk associated with chronic low dose exposures to

these compounds because of a quite limited toxicological data and information on dose-response characteristics in

human or animal models. Only very recently some few data have been become available, which partly make possible to

carry out reliable exposure and risk assessments. Due to the missing risk assessments, it has been difficult to regulate

the presence of these chemicals in indoor environments up to now.

It is, therefore, recommended to develop a strategy for indoor air quality assessment and management in Europe and

that future clean air policies take into account the total air exposure of European citizens, which will necessarily include

exposures to pollutants from both outdoor and indoor sources. This report offers background information for this

strategy planning.

The INDEX project Final report

2. Objectives

The aim of the project was to create a network of European leading scientists in the area of indoor air pollution and the

herewith-associated health impacts in order to identify priorities for a Community strategy and action plan. The key

issues that were addressed within the project were:

- to provide information on links with existing knowledge, ongoing studies, legislation etc.,

- the setting up of a list of priority substances to be regulated in indoor environments on the basis of health

impact criteria,

- to provide suggestions and recommendations on potential exposure limits or other risk management

procedures for these substances, and

- to assess essential research needs for pollutants with high risk potential, but insufficient information for

carrying out risk assessment and setting regulatory objectives or selecting regulatory options.

The INDEX project Final report

3. Methodology

3.1 Risk Assessment

Risk assessment was based on indoor air quality and exposure data from the scientific literature, databases and directly

from the researchers. Similarly, dose/response data was reviewed from scientific literature. Two working groups (WG)

of experts from several countries were established to assess these data. Finally, risk characterisation, recommendations

for risk management and conclusions were drawn by the scientific steering committee based on the work of the working

groups for exposure assessment and dose-response assessment. The main steps of the project are presented in Figure

3.1.1. The risk assessment approach applied to this project is presented in more detail in the following paragraphs.

Step 1:

Literature

review

Step 2: Setting up

criteria to select

compounds

Step 3: Review of

exposure and

dose/response data

Step 4: Risk

characterisation of the

selected compounds

Step 5: Prioritisation of

the selected compounds

Step 6:

Recommendations and

risk management

options on potential

exposure limits

Figure 3.1.1: The main steps of the INDEX project.

3.1.1 Selection of Indoor Air Chemicals for Consideration (Hazard

Identification)

The hazard identification of the indoor air pollutants was assessed combining the information of the prevalence of

pollutants in European homes with the available knowledge of adverse health effects that these compounds had been

linked to in toxicological or epidemiological studies. If a compound was present in the indoor air and it had had adverse

health effects, it was considered as a potential hazard to European populations and was thus included in the risk

assessment process. The process and the criteria to include or exclude each compound to the risk assessment process in

this project are presented in the following paragraphs.

The INDEX project Final report

The selection criteria of the compounds to be included in the risk analysis

Considering the time limit of this exercise, and the fact that no new data could be generated in this project, the steering

group defined the following criteria to be able to select the chemical compounds to the risk analyses:

1. Only single compounds will be considered

2. The compound should have common indoor sources, which dominate the exposures of at least significant

fraction of the population

3. The compound should have known health effects.

It was also decided that compounds, which have been regulated by specific guidelines or regulations would be excluded

from this analysis. For example, radon and tobacco smoke were excluded from the risk assessment process due to these

criteria.

Phase 1: Literature review

In the first phase, a literature review was carried out to collect information about candidate pollutants to be assessed in

the later stages of the project. The scientific literature of the indoor air pollutants was reviewed by using several search

engines in the Internet and by searching from the web pages of the relevant journals. In addition to electronic search,

numerous study reports concerning indoor air pollutants were reviewed.

The main focus of the review was on recent population-based studies to be able to evaluate current population

exposures to selected pollutants in Europe. Only compounds with known indoor sources and known health effects were

taken into consideration.

Based on the literature review, a summary table of the concentrations in residential indoor, workplace indoor,

residential outdoor and personal exposures was created for the compounds meeting the first criteria (see Annex 1).

Simultaneously, some dose/response data were collected for the selected compounds. Also the existing international

(i.e. EU and WHO) and national guidelines for those compounds in indoor air were reviewed. Finally, the output of the

literature review was used as an input for the next steps of the risk assessment.

Establishment of the working groups

For the evaluation and development of the reviewed literature, two working groups (WG) were established, one for

exposure assessment (WG

) and one for dose/response assessment (WG

). The following experts were nominated to

the working groups:

on exposure assessment

Members: M. Jantunen (co-ordinator, KTL), Finland, C. Cochet, France, S. Kirchner, France, K. Koistinen, Finland, J.

Mc Laughlin, Ireland, S. Kephalopoulos, EU/JRC-Ispra, E. Oliveira-Fernandez, Portugal, B. Seifert, Germany, and

on dose/response assessment

Members: P. Carrer (co-ordinator, UniMi), Italy, T. Lindvall, Sweden, M. Maroni, Italy, L. Mølhave, Denmark and C.

Schlitt, Italy.

Phase 2: The selection of compounds to the further analysis

In the second phase of the selection process, the working groups assessed the reviewed data and collected more detailed

information for the previously selected compounds. The aim of the work was to select about 20 compounds for further

analyses. In this phase the steering group excluded more compounds using the following criteria:

- no expressed concerns for health at present levels (for example acetone, decane, ethylbenzene, phenol ,

propylbenzene, trimethylbenzene)

- compound already regulated by use restrictions for indoor materials (pentachlorophenol)

- incomplete or no dose-response data available at present levels (methyl-ethyl-ketone, propionaldehyde)

- the main route/media for the exposure to the compound is other than indoor air (lead, mercury).

After detailed review and discussion of the available information, the working groups selected the 25 compounds

presented in Annex 2 to the more detailed analysis.

The INDEX project Final report

Phase 3: Compounds selected for a detailed risk assessment

In addition to exposure and dose/response data, also data about odour threshold values were considered important and

thus, these data were added to the background information. The standardised human odour threshold values were taken

from Devos et al. (1990).

On the basis of the available information and after an extensive discussion on the chemical substances, the steering

group decided to conduct detailed assessment of the 14 compounds presented in Figure 3.1.1.1.(phase 3).

Based on the potential/estimated population risk caused by concentrations from indoor sources, toxicological properties

including hypersensitivity for allergy and asthma, other known health effects and comfort, the steering group decided to

separate the 14 compounds into two groups: 1) compounds to be considered with high priority and 2) compounds that

could be included in the final prioritisation after further examination and new published findings of their hazardous

potential and the other criteria set within the project.

After a first evaluation the compounds were initially classified as follows:

Group 1 (high priority) Benzene, Acetaldehyde, Formaldehyde, CO and NO

Group 2 (further information needed) m&p-Xylenes, o-Xylene, Naphthalene, Styrene, Toluene, a-Pinene, d-Limonene,

Flame-retardants were regarded as an emerging issue, which will require further consideration in the future. The

compound Tris-(2-chloroethyl) phosphate belongs to this group, but because reliable data on its sources and occurrence

in indoor environments, exposure routes and on toxicological properties were lacking, the compound was not included

in the evaluation procedure in this project.

Figure 3.1.1.1: The indoor originated compounds that were assessed and considered the most hazardous in the three

phases of the hazard identification process.

Phase 1

Phase 2

Phase 3

1. priority

Formaldehyde

Carbon monoxide

Nitrogen dioxide

Benzene

Naphtalene

Acetaldehyde

Ammonia

a-Pinene

Benzene

Carbon monoxide

d-Limonene

Formaldehyde

m&p-Xylene

Naphtalene

Nitrogen dioxide

o-Xylene

Styrene

Toluene

1-Butanol

2-Ethyl-1-hexanol

3-Carene

Acetaldehyde

Ammonia

a-Pinene

Benzaldehyde

Benzene

Cadmiuim

Carbon monoxide

Dichloromethane

Diisocyanate

d-Limonene

Formaldehyde

Hexaldehyde

m&p-Xylene

Naphtalene

Nitrogen dioxide

o-Xylene

Styrene

Tetrachloroethylene

Toluene

Trichloroethylene

Tris-(2-chloroethyl) phosphate

1-Butanol

2-Buthoxyethanol

2-Ethyl-1-hexanol

2-Methyl-1-propanol

3-Carene

Acetaldehyde

Acetone

Ammonia

a-Pinene

Benzaldehyde

Benzene

Benzo[a]pyrene

Cadmiuim

Carbon monoxide

Decane

Dichloromethane

Diisocyanate

d-Limonene

Ethylbenzene

Formaldehyde

Hexaldehyde

Lead

m&p-Xylene

Mercury

Methyl-ethyl-ketone

Naphtalene

Nitrogen dioxide

Nonane

o-Xylene

Pentachlorophenol

Phenol

Propionaldehyde

Propylbenzene

Styrene

Tetrachloroethylene

Toluene

Trichloroethylene

Trimethylbenzenes

Tris-(2-chloroethyl) phosphate

Undecane

The INDEX project Final report

3.1.2 Exposure Assessment

Exposure to selected indoor air pollutants was estimated by collecting exposure data from scientific literature, from

available databases, and by personal communications. The aim of this work was to summarise prevailing indoor air and

personal exposure concentrations of these compounds in Europe and also worldwide. These reviews are mainly focused

on indoor air and exposure concentrations measured recently in European population based studies such as EXPOLIS

(Jantunen et al 1998), German Environmental Surveys, GerES, (Seifert et al 2000), the German study on Indoor Factors

and Genetics in Asthma, INGA (Schneider et al 1999 and 2001), and a national survey of air pollutants in English

homes (Raw et al 2002). Also some preliminary results of the French National Survey (Golliot et al 2003, Kirchner

2004) were available during this project. Some comparisons have been also done to the TEAM (Wallace et al 1991) and

the NHEXAS (Sexton et al 1995, Pellizzari et al 1995) studies carried out in the USA. Results from population-based

studies have been used to be able to generalise the results from studied individuals to larger populations, targeting to

assess exposures of all the Europeans.

Population exposures are typically reported in literature using parameters such as arithmetic or geometric mean and

standard deviation. Mean concentrations give us a general picture of the concentration levels, but due to presence of

subpopulations that are exposed to much higher concentrations, the whole exposure distribution is needed when linking

these exposures to toxicological or epidemiological dose-response data. The distributions presented in this report are

drawn using arithmetic or geometric means and respective standard deviations, reported in the literature or extracted

from the databases, assuming that the measured data is log-normally distributed, which is a typical shape for the

distributions of the naturally occurring pollutants. These parameters taken from the large scale European population

based studies are presented in Annexes 4, 5, and 6.

Concentrations have been linked to the main emission sources if possible. Short time concentration peak values are

presented in tables and graphs to be used in the assessment of acute health effects.

European Union has recently published risk assessment reports (RAR) for some of the chemicals that are reviewed in

this report. Final RARs are available for naphthalene (EU 2003a), styrene (EU 2002a), toluene (EU 2003b) and a draft

report for benzene (EU 2002b). These reports were used as a data source in this project, and therefore the contents

presented in those reports have not been repeated in this report.

3.1.3 Dose/Response Assessment

In preparing the dose-response assessment fact sheets of 12 chemicals (m,p, and o-xylenes handled together),

information were retrieved from scientific literature (mainly by electronic search), comprehending toxicological reviews

of leading health organizations, risk evaluation documents and available databases, as outlined in Table 3.1.3.1. In

addition, Toxline and Medline were searched for relevant scientific communications published up to September 2004.

Nearly all key-studies referred to in the present assessment, establishing effect levels for appropriate toxicological

endpoints, are those selected by health organizations for the derivation of health based limits of exposure (WHO/GV,

IRIS/REL, OEHHA/RfC, ATSDR/MRL, HC/TC, UBA/GVII&I) or among risk assessment requirements (ECB).

Although not specifically addressing health hazards and risks associated with indoor air exposure, i.e. not being

designed for the expression of effects at lowermost exposure concentrations, nearly all studies were aimed at identifying

the most sensitive endpoint considered to be of relevance to humans. Summary definitions of exposure limits/guidelines

established by these health organisations are given in Table 3.1.3.2. Where relevant, studies conducted on susceptible

sub-populations (e.g. asthmatics, infants, children, pregnant women etc.) were quoted and taken into consideration in

the risk characterization.

The INDEX project Final report

Table 3.1.3.1: Toxicological reviews, risk evaluation documents

and databases consulted and referred to in the dose response assessment

World Health Organization

Air Quality Guidelines for Europe

International Program on Chemical Safety

Environmental Health Criteria, Concise International

Chemical Assessment Documents and Health and Safety

Guides

European Chemicals Bureau - Institute for Health and

Consumer Protection

European Union Risk Assessment Reports

Agency for Toxic Substances and Disease Registry

Toxicological Profiles

U.S.EPA - Integrated Risk Information System

Office of Environmental Health Hazard Assessment of

the Californian EPA

Acute and Chronic Toxicity Summaries

IARC monographs

Summaries and Evaluations

WHO IPCS ECB ATSDR IRIS OEHHA Others when relevant IARC

Formaldehyde

2000

EHC 89

1989

CICADS

40 2002

1999

being

reassessed

acute &

chronic

1999

NIWL 2003

Vol. 88

1995

2004 in

preparation

Acetaldehyde

EHC 167

1995

HSG 90

1995

1991

being

reassessed

chronic

1997

Vol. 71

1999

Ammonia

EHC 54

1986

HSG 37

1990

2002

draft

2003

acute &

chronic

1999

Carbon monoxide

2000

EHC 213

1999

acute

1999

Nitrogen dioxide

2000

EHC 188

1997

1993

acute

1999

UBA 1998

Benzene

2000

EHC 150

1993

2004

draft

1997 2003

acute &

chronic

1999

Vol. 29

Suppl.7

1987

Toluene

2000

EHC 52

1986

2003 2000

1994

being

reassessed

acute &

chronic

1999

UBA 1996

Vol. 71

1999

Xylenes

EHC 190

1997

1995 2003

acute &

chronic

1999

Vol. 71

1999

Styrene

2000

EHC 26

1983

1992

1993

being

reassessed

acute &

chronic

1999

Vol. 82

2002

Naphthalene

2003

draft

1998

being

reassessed

chronic

1999

UBA 2004

Vol. 82

2002

Limonene

CICADS

1998

1993 NICNAS 2002

Vol. 73

1999

α-Pinene

UBA 2003

NIEHS 2002

including: Evaluation of current California Air Quality Standards with respect to protection of children (2000)

including: EPA-Toxicological review of benzene/noncancer effects (2002) and EPA-Carcinogenic Effects of Benzene: An Update (1998)

National Institute for Working Life - The Nordic Expert Group for Criteria Documentation of Health Risks from Chemicals and The Dutch Expert

Committee on Occupational Standards: 132. Formaldehyde - Anton Wibowo

Umweltbundesamt: Bundesgesundheitsblatt 11/96: Richtwerte für die Innenraumluft: Toluol. H.Sagunski,

Umweltbundesamt: Bundesgesundheitsblatt 7 · 2004: Richtwerte für die Innenraumluft: Naphthalin. H.Sagunski, W.Heger

The INDEX project Final report

Umweltbundesamt: Bundesgesundheitsblatt 4 · 2003: Richtwerte für die Innenraumluft: Bicyclische Terpene (Leitsubstanz α-Pinen). H.Sagunski,

B.Heinzow

National Institute of Environmental Health Sciences, Toxicological Summary For Turpentine, Review of Toxicological Literature, February 2002

Table 3.1.3.2: Exposure limits/guidelines established by health organisations and their summary definitions

World Health Organization - Air Quality Guidelines for Europe

Guidelines: The term “guidelines”, in the context of the WHO - Air Quality Guidelines for Europe, implies not only numerical values (guideline

values), but also any kind of guidance given. Accordingly, for some substances the guidelines encompass recommendations of a more general nature

that will help to reduce human exposure to harmful levels of air pollutants. For some pollutants no guideline values are recommended, but risk

estimates are indicated instead.

The starting point for the derivation of guideline values is to define the lowest concentration at which adverse effects are observed. On the basis of the

body of scientific evidence and judgements of uncertainty factors, numerical guideline values were established to the extent possible. Compliance

with the guideline values does not, however, guarantee the absolute exclusion of undesired effects at levels below the guideline values. It means only

that guideline values have been established in the light of current knowledge and that uncertainty factors based on the best scientific judgements have

been incorporated, though some uncertainty cannot be avoided. The numerical values for the various air pollutants should be considered in the context

of the accompanying scientific documentation giving the derivation and scientific considerations. Any isolated interpretation of numerical data should

therefore be avoided, and guideline values should be used and interpreted in conjunction with the information contained in the appropriate sections.

Guidelines based on carcinogenic effects are indicated in terms of incremental unit risks in respect of those carcinogens that are considered to be

genotoxic. To allow risk managers to judge the acceptability of risks, this edition of the guidelines has provided concentrations of carcinogenic air

pollutants associated with an excess lifetime cancer risk of 1 per 10 000, 1 per 100 000 and 1 per 1 000 000.

Agency for Toxic Substances and Disease Registry (ATSDR)

Minimal Risk Levels (MRLs): During the development of toxicological profiles, Minimal Risk Levels (MRLs) are derived when reliable and sufficient

data exist to identify the target organ(s) of effect or the most sensitive health effect(s) for a specific duration for a given route of exposure. An MRL is

an estimate of the daily human exposure to a hazardous substance that is likely to be without appreciable risk of adverse noncancer health effects over

a specified duration of exposure. MRLs are based on noncancer health effects only and are not based on a consideration of cancer effects. These

substance-specific estimates, which are intended to serve as screening levels, are used by ATSDR health assessors to identify contaminants and

potential health effects that may be of concern at hazardous waste sites. It is important to note that MRLs are not intended to define clean-up or action

levels. MRLs are derived for hazardous substances using the no-observed-adverse-effect level/uncertainty factor approach. They are below levels that

might cause adverse health effects in the people most sensitive to such chemical-induced effects. MRLs are derived for acute (1–14 days),

intermediate (15–364 days), and chronic (365 days and longer) durations and for the oral and inhalation routes of exposure.

U.S.EPA - Integrated Risk Information System (IRIS)

The inhalation Reference Concentration (RfC) is analogous to the oral RfD and is likewise based on the assumption that thresholds exist for certain

toxic effects such as cellular necrosis. The inhalation RfC considers toxic effects for both the respiratory system (portal-of-entry) and for effects

peripheral to the respiratory system (extrarespiratory effects). It is expressed in units of mg/cu.m. In general, the RfC is an estimate (with uncertainty

spanning perhaps an order of magnitude) of a daily inhalation exposure of the human population (including sensitive subgroups) that is likely to be

without an appreciable risk of deleterious effects during a lifetime. Inhalation RfCs were derived according to the Interim Methods for Development

of Inhalation Reference Doses (EPA/600/8-88/066F August 1989) and subsequently, according to Methods for Derivation of Inhalation Reference

Concentrations and Application of Inhalation Dosimetry (EPA/600/8-90/066F October 1994). RfCs can also be derived for the noncarcinogenic

health effects of substances that are carcinogens. Therefore, it is essential to refer to other sources of information concerning the carcinogenicity of

this substance. If the U.S. EPA has evaluated this substance for potential human carcinogenicity, a summary of that evaluation will be contained in

Section II of this file.

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency (OEHHA)

The concentration, at or below which no adverse health effects are anticipated in the general human population, is termed the reference exposure level

(REL). RELs are based on the most sensitive relevant adverse health effect reported in the medical and toxicological literature. RELs are designed to

protect the most sensitive individuals in the population by the inclusion of margins of safety. Protection against carcinogenicity and against adverse

health effects of short-term exposures are not considered in these guidelines. For this reason, chemicals should be evaluated separately for their

carcinogenic potential and additional acute health effects that may occur.

Methods for the evaluation of acute and chronic health effects and for the carcinogenic potential of chemicals are provided in the OEHHA documents

entitled Air Toxics Hot Spots Program Risk Assessment Guidelines:

Acute Reference Exposure Levels (RELs): three categories of acute severity levels are developed in accordance with criteria established by NRC

(1993): the level protective against mild adverse effects, the level protective against severe adverse effects, and the level protective against life

threatening effects. Each of these three acute exposure levels is determined for one-hour exposure duration. However, the major focus of this

document is in developing acute RELs for the preparation of risk assessments for non-emergency routine releases. Thus, the RELs used in the risk

assessment are generally levels protective against mild adverse effects; a few are based on severe effects (e.g., reproductive/developmental).

Chronic Reference Exposure Levels (RELs): Chronic reference exposure levels are concentrations or doses at or below which adverse health effects

are not likely to occur. A central assumption is that a population threshold exists below which adverse effects will not occur in a population; however,

such a threshold is not observable and can only be estimated. Areas of uncertainty in estimating effects among a diverse human population exposed

continuously over a lifetime are addressed using extrapolation and uncertainty factors.

Federal Environmental Agency of Germany (Umweltbundesamt – UBA)

In Germany, an important framework for the setting of indoor-related guideline values is given by the building codes (which are under the jurisdiction

of the German States). The building codes demand that there be no health hazard to occupants from the building. Hence the work of the ad-hoc group

focused on defining concentration levels at which such hazard would probably occur. The introduction of a safety margin would then allow the

definition of a concentration where there would be no more concern for adverse health effects. The following two concentration levels were defined:

- Guideline Value II (GV II): GV II is a health-related value based on current toxicological and epidemiological knowledge. If the concentration

corresponding to GV II is reached or exceeded immediate action must be taken because permanent stay in a room at this concentration level is

likely to represent a threat to health, especially for sensitive people. In this context, taking action means an immediate examination of the situation

with regard to a need for control measures. It may include the evacuation of the room in question. If by measurement GV II has been found to be

The INDEX project Final report

exceeded, the results should be checked by repetitive measurements carried out immediately under normal conditions of occupancy. If possible and

deemed meaningful, biological monitoring of the occupants should be carried out in addition.

- Guideline Value I (GV I): GV I is the concentration level at which a substance – taken individually – does not give rise to adverse health effects

even at life-long exposure. An exceedance of GV I is linked with an exposure beyond normal which is undesirable from a hygienic viewpoint.

Thus, there is also a need for action at concentrations between GV II and GV I. GV I is obtained by dividing GV II by a factor of 10. This factor is

a convention. However, for odorous substances GV I must be defined based on the odour threshold (“odour detection”) if the odour threshold has a

lower numerical value than the concentration derived according to the general scheme. GV I can be used as the level to be reached after control

measures have been applied. The level should not be “filled up”; rather, the final concentration should fall below.

Environment Canada - Health Canada (EC-HC)

Different approaches were adopted for assessments of Priority Substances for chemicals for which the critical effect is believed to have a threshold

and those for which it is considered not to have a threshold. For substances for which the critical effect is considered to have a threshold (i.e., non-

neoplastic effects), Tolerable Daily Intakes (TDIs) or Tolerable Concentrations (TCs) have been developed by dividing effect levels observed in

studies in exposed populations or animal species, by uncertainty factors.

Priority Substances are classified into one of 6 categories, based on the weight of evidence of carcinogenicity (see Section 3). For genotoxic

carcinogens (i.e., primarily compounds which are considered “carcinogenic to humans” or “probably carcinogenic to humans”, Groups I or II of the

scheme for classification of carcinogenicity under CEPA), quantitative estimates of the carcinogenic potency within or close to the experimental

range have been developed. This approach was adopted for several reasons, one of the most important of which was to avoid expressing risk in

precise absolute terms (i.e., predicted excess numbers of cancers per unit of the population) based on uncertain low-dose extrapolation procedures.

Tolerable Concentration (TC): Tolerable concentrations (Section 3a) (often expressed in mg/m3) are generally airborne concentrations to which it is

believed that a person can be exposed continuously over a lifetime without deleterious effect. They are based on non-carcinogenic effects.

Tumorigenic Concentration 05 (TC05): The Tumorigenic Concentration 05(TC05) is the concentration generally in air (expressed, for example, in

mg/m3) associated with a 5% increase in incidence of, or deaths due to, tumours considered to be associated with exposure, observed in

epidemiological studies in human populations or bioassays in experimental animals Values derived based on division of the TC05s by a suitable

margin (e.g., 5.000 to 50.000)* can provide a benchmark against which the adequacy of indoor or ambient air can be judged, with respect to potential

carcinogenicity.

It should be noted that Health Canada does not necessarily deem as “acceptable” from a societal viewpoint health risks associated with these values

and that the Health Protection Branch continues to subscribe to the position that exposure to substances for which the critical effect has no threshold

be reduced to the extent possible.

* Since Tumorigenic Concentration05s were computed directly from the curve within or close to the experimental region, division by an additional

factor of 2 would equate approximately to the lower 95% confidence limit.

Key-studies were summarised among each chapter treating effects of short- and long-term exposure and itemised in

tables at the end of the chapter. One-page fact sheets resuming the most relevant toxicological properties are given at

the end of each D/R assessment chapter. In the key-study Tables, subscripts were assigned to effect levels (NOAELs

and LOAELs), stating on whether occupational average levels or experimental concentrations are quoted or identifying

the extrapolation process applied for the given value. Details on these subscripts are given in Table 3.1.3.3.

Table 3.1.3.3: Subscripts used in summary tables and fact sheets quoting key-studies

and the exposure limit derivation process accounted for by health organisation

Subscript Description Details

Study

average concentration time weighted average concentration with no information on

maximum concentrations, in chronic studies generally estimated

from numerous intermittent measurements in occupational settings,

even over years

EXP

experimental concentration concentrations artificially generated in inhalation/exposure chambers

in animal or volunteer studies

ADJ

concentration adjusted from an intermittent to a

continuous exposure

When extrapolating from occupational to population based

exposures on a [hour/day and days/7 days] basis (generally: division

by 4.2)

1h-ADJ

concentration adjusted to 1-hour exposure duration For the extrapolation from sub-acute to acute

HEC

human equivalent concentration Generally a blood-to-air partition coefficient of the chemical for the

experimental animal species was used in the HEC derivation of an

RfC

MLE

maximum likelihood estimate for 5% response A statistical best estimate of the value of a parameter from a given

data set.

BC05

BC05 is the 95% lower confidence limit of the

concentration expected to produce a response rate

of 5%

Following a Benchmark (BM) approach, alternative to the traditional

NOAEL/LOAEL approach. A Benchmark Concentration (BMC) is a

statistical lower confidence limit on the dose producing a

predetermined, altered response for an effect.

STAT

lowest statistically significant effect concentration

The INDEX project Final report

3.1.4 Risk Characterization and prioritisation of chemicals

In this final step of the general risk assessment process, the incidence of health hazards and risks in the European

population, associated with indoor exposure to individual chemicals, was estimated. It is pointed out that the assessment

of risk is a scientific one, which has been kept separate from any consideration regarding the risk management process,

including the setting or the proposal of Indoor Exposure Limits. Namely, an important uncertainty, not accounted for in

the assessment, is the possibility of antagonistic and synergistic effects arising from the exposure to mixtures of

chemicals, little scientific information existing in this area. Multiple contaminants are typically occurring in indoor

environments (although at low concentrations) and the resulting uncertainty (uncontrolled factor) should be taken into

consideration in the risk management.

Nevertheless, Limits of Exposure (ELs) had to be established in order to perform the risk characterization associated

with individual chemicals, following inhalatory exposure in indoor environments, for both short-term (indoor-activity

related) and long-term exposures (background indoors).

An EL was derived for each chemical after the identification of key-studies (critical-study) describing the appropriate

toxicological endpoints (among those selected by health organizations for the derivation of health based reference

concentrations).

Uncertainty factors (here named assessment factors, AF) applied in the present assessment to account for the adequacy

of the critical study are the product of the individual factors outlined in Table 3.1.4.1.

Table 3.1.4.1: Elements considered for the derivation of Uncertainty (Assessment) Factors (AS)

Description Detail Factor

Extrapolation from a LOAEL to a NOAEL When in the critical study no NOAEL could be observed 10

Interspecies extrapolation

Critical study = experimental animal study

(no human study available or appropriate)

Inter-interindividual (intraspecies) variability

in humans

Always, unless the critical study was performed on individuals of the sub-

population considered susceptible

Susceptible population

asthmatic individuals, infants, children, individuals with heart diseases,

individuals with (hereditary) enzyme deficiencies, pregnant women

10; 3; 2

Adequacy or quality of toxicological data Old study 2

Extrapolation from sub-acute to chronic Deficiencies in toxicological database 10

Extrapolation from sub-acute to acute Deficiencies in toxicological database 10

For all chemicals a threshold-level of action could be identified, enabling a “no-observed-adverse-effect level

(NOAEL)/assessment factor (AF)” approach, i.e. EL derived by dividing the critical effect level by the AF, with the AF

based on the available scientific evidence. Where no NOAEL observation was documented, a lowest-observed-adverse-

effect level (LOAEL) was taken into consideration and an additional assessment factor of 10 used for EL derivation.

For one compound only (benzene) the characterization has been based on population cancer risk estimation rather than

on an EL.

In those cases where large differences in sensitivity for different susceptible groups were documented, a bimodal

distribution of population responses was supposed to exist and a tenfold difference in sensitivity, usually accepted as

higher than the encountered range, taken in account in the AF derivation. Susceptible subpopulations considered in the

present characterization are: asthmatic individuals, infants, children, individuals with heart diseases, individuals with

(hereditary) enzyme deficiencies, pregnant women.

Information from the exposure assessment (Chapters 1-3) and toxicity assessment (Chapters 4-5) were integrated and a

risk characterization performed for each chemical (Chapter 6). Based on the conclusions of the assessments and on the

completeness of individual databases a priority ranking of the 14 chemicals was established with the chemicals assigned

to one of three groups, and the results of the assessment handed over to Risk Management.

Group 1: High priority chemicals

Group 2: Low priority chemicals

Group 3: Chemicals requiring further research with regard to human exposure or dose response

The INDEX project Final report

3.2 Risk Management

Buildings are built for shelters from the cold (or heat), wind and rain of the outdoor air. Typically indoor air contains

higher levels of most air pollutants than outdoor air. As long as people feel that they are in control of their own indoor

environment, however, there is usually a higher tolerance to poor indoor air quality compared to outdoor air quality -

after all one can always go outdoors or open a door or window. This tolerance is greatly decreased if self-control is

lacking.

Risk assessment disentangles an identified risk complex (e.g. intoxication from a high indoor CO concentration peak)

into its details (e.g. formation of CO in a faulty combustion device, release of CO into room instead of chimney,

exposure to CO in the room, absorption to bloodstream in the lung, subsequent reduction of oxygen delivery, and

hypoxia), and analyses these details one by one and in relation to each other. Risk management must entangle these

details back again into an efficient and enforceable policy to reduce the risk (e.g. building codes requiring combustion

devices to be vented to chimneys, regular maintenance & inspections, CO-alarms, etc.).

While risk assessment is a science based exercise, risk management by necessity involves also technical, economic,

social and legal realities, as well as need for actions in the face of incomplete information. Risk management decisions

concerning a specific risk are based on one hand on risk characterisation, produced by risk assessment, and on the other

hand on the political and administrative processes that generate, evaluate, select and implement policies in the society

(EU Commission directive 93/67/EEC). The indoor air risk management options include, in principle:

• IAQ Standards and guidelines

• Building codes and ventilation standards

• Equipment standards and permits

• Mandatory maintenance and inspections

• Limits, labelling and reporting of the contents of or releases from building products, furnishing materials,

equipment and consumer products

• Public awareness raising and information

These alternatives will be discussed separately for each of the final shortlist components in Chpt. 6.1. The following

discussion looks at indoor air risk management in general.

The core of ambient air risk management is in the ambient air quality guidelines and standards. The ambient air

concentration limits have questionable applicability for indoor environments, because they are derived for different

pollution mixtures and exposure patterns, because comparable monitoring and implementation alternatives do not exist

indoors, and because indoor air quality for an individual may depend strongly on the behaviour of the same individual.

When ambient air quality standards are applied for indoor air quality (as determined by indoor sources and activities),

short term (24 h) measurements of indoor pollutants should in most cases be related to long term (annual) outdoor air

guidelines.

The occupational exposure limits (OEL, TLV, MAK, etc.), although derived for indoor environments, are not directly

applicable either, because they have been developed for healthy adult populations and more or less controlled 40 h

exposure in a 168 h week.

Risk managers can intervene in many points (that lead from the primary source of risk to its materialisation). They can

prevent the hazardous process, reduce exposures, modify effects, alter perceptions or valuations through education and

public relations or compensate for damage after the fact. (Morgan 1993).

The issue, which is most specific for indoor air risk management, becomes obvious by visualising the millions of indoor

environments in just one city, and the daily multitudes of potential hazards in them. A vast majority of both indoor air

risk observations (assessments) as well as decisions (management) will remain to be made by occupants themselves -

and this is particularly true for the acute effects from short term hazardous exposures. No imaginable public service or

professional expert resources can ever cover but a tip of the iceberg of all hazardous IAQ situations. At best there are

enforceable governmental regulations, respected industrial/professional standards/practices and/or authoritative

examples.

In contrast to ambient air quality management, which relies on air quality standards and monitoring networks, the

capabilities of public services and experts to manage most indoor air risks lie crucially on their abilities to establish and

maintain public confidence, to communicate the risks in a way that the public is able to comprehend, to relate to their

own living conditions and willing to accept, and to present practical solutions which are technically, financially and

legally feasible, and which sufficiently acknowledge the diversity of the buildings, the occupants, expectations and

risks.

The largest investments in the life of most individuals are their homes, and buildings represent in the order of 2/3 of

national property in most European countries. This fact sets very concrete requirements for any interventions into the

The INDEX project Final report

residential building stock, as well as for the allocation of financial burdens and benefits. The fact that most indoor

environments are private property and protected by privacy of the individuals significantly limits the options of public

indoor air risk management.

The public services can, however, rely more on standards and regulations in managing risks in rented apartments,

hotels, public buildings, workplaces, schools, hospitals, shopping malls etc. These buildings should be ensured to be

safe for a great majority of the population (excluding only exceptionally sensitive individuals) by public health,

occupational health and building inspection authorities as the last resort. When credible concerns arise they should be

addressed promptly and properly for the sake of public health, confidence and liability. One of the most important

special tasks for the public services and experts is prediction, identification, assessment and management of the

situations where susceptible population groups and risky indoor environments coincide.

Risk - benefit: In managing risks from indoor air pollutants, there is a case for erring on the side of caution and acting

to reduce risks if there is any reasonable suspicion that a hazard exists. However:

- any chemical can be hazardous at sufficient concentration, and yet attempt to reduce all measurable concentrations

to zero is not feasible;

- blind application of precaution may divert resources into irrelevant actions, and thus lead to sub-optimal protection

of health and/or the environment. It may also unjustifiably deny the benefits of using products or techniques; and

- consideration should be given to the fact that limiting the use of one chemical may entail the application of a

substitute that may not have lower, but only less well-characterised risks.

Hence, legislation in Europe has moved towards approaches based on overall risk rather than mere hazard

identification. Article 10 of the Council Regulation (EEC) 793/93 (Existing Substances Regulation) requires that:

"Where such control measures include recommendations for restrictions on the marketing and use of the substance in

question the rapporteur" [on behalf of the body carrying out the risk assessment] "shall submit an analysis of the

advantages and drawbacks and the availability of replacement substances".

Risk Communication: Risk communication is by far the most important tool for the society to manage indoor air

health risks. Successful communication will not raise undue concerns, but suggests behaviour, which reduces avoidable

risks, yet restricts nonessential interference into the lives of the individuals or to the alternatives of the enterprises. Poor

risk communication may not only fail to produce the wanted risk reduction, it may also unnecessarily limit individuals'

options, add public and private costs, and generate new risks due to non-productive public concerns and potentially

harmful behavioural changes.

The essence of good risk communication is simple: Learn what people believe and the decisions that they are facing and

tailor the language and message to this knowledge. Increased (decreased) trust in the managing organisation of the

activity under concern - including indoor air risk management - will lead to lower (higher) levels of perceived risk and

thus increase public support (opposition). Understanding the dimensions of trust is essential for developing policy

strategies that will gain public acceptance. (Flynn et al. 1992)

Economically and technically sound [indoor air risk management] policies may be doomed if people believe that they

distribute benefits and burdens unfairly. Three concepts emerge from the philosophical and cultural basis of risk

sharing: parity, in which each individual, group or country is treated equally; priority, or giving the burden to those

most deserving of it; and proportionality, or the sharing of burdens according to need or contribution. (IIASA 1993)

The INDEX project Final report

4. Risk Assessment of the selected chemicals

The exposure and dose/response data reviewed during the project are presented in this chapter. The summary of the

typical concentrations in Europe for the selected compounds are presented in Table 4.0.1. To be able to compare

recently published results reviewed in this report to the ‘normal’ indoor concentrations defined by the Working Group

of WHO (WHO 1989), these concentrations are summarised in Table 4.0.2.

Table 4.0.1. Typical European microenvironmental and exposure concentrations (µg/m

, except CO in mg/m

)

summarised from population based studies reviewed in this report.

Out

Aromatics

Benzene 2-13 4-14 1-21 3-23

Naphthalene 1-90 2-8 1-4 2-46

Styrene 1-6 3-7 1-2 1-5

Toluene 15-74 25-69 3-43 25-130

m&p-Xylenes 4-37 25-121 2-23 25-55

o-Xylene 2-12 7-29 1-8 8-15

Aldehydes

Acetaldehyde 10-18 3 1-2 8

Formaldehyde 7-79 12 2-4 21-31

Terpenes

a- Pinene 11-23 1-17 1-7 7-18

Limonene 6-83 11-23 5-9 19-56

Classical pollutants

CO 0.5-1 1 2 0.8-1.7

13-62 27-36 24-61 25-43

In, W, Out, P = Indoor, workplace, outdoor and personal exposure concentrations

Table 4.0.2. ‘Normal ‘residential indoor air concentrations of selected organic pollutants reviewed by a Working Group

of WHO in 1987 (adopted from WHO 1989 and Maroni et al 1995).

Pollutant 10% median 90%

M Sources

Aromatics

Benzene 2 10 20 10 fuel component, tobacco

Naphtalene 2 5 solvent moth balls

Styrene < 1 1 5 fuel component

Toluene 30 65 150 80 fuel component, solvent

m,p-xylene 10 20 40 20 fuel component, solvent

o-xylene 3 5 10 10 fuel component, solvent

Aldehydes

acetaldehyde 10 30 cigarette smoke

formaldehyde 25 60 40 chipboard, urea-formaldehyde insulation

Terpenes

a-pinene 2 10 20 10 wax, wood product

limonene 2 15 70 30 odorant, detergent

10% and 90% = ith percentile, AM = arithmetic mean

Concentration (µg/m

)

The INDEX project Final report

Formaldehyde

Synonyms: Formalin, methanal, oxoymethane, oxomethylene, methylene oxide, formic

aldehyde, methyl aldehyde

CAS Registry Numbers: 50-00-0

Molecular Formula: CH

1. Compound Identification

Formaldehyde is a colourless, odorous, highly reactive, flammable and easily polymerised gas at room temperature and

pressure. It is formed by oxidation or incomplete combustion of hydrocarbons. Formaldehyde is soluble in water,

ethanol and diethyl ether and is used in solution or in polymerised form (paraformaldehyde). In atmosphere

formaldehyde is photo-oxidised in sunlight to carbon dioxide. Its half-life in urban air, under the influence of sunlight,

is short. In the absence of nitrogen dioxide, the half-life of formaldehyde is approximately 50 minutes during the

daytime, but in the presence of nitrogen dioxide, it decreases down to 35 minutes (WHO 1989).

In solution formaldehyde has a wide range of uses: in the manufacture of resins and textiles, as a disinfectant, and as a

laboratory fixative or preservative in consumer products, such as food, cosmetics, and household cleaning agents (SIS

2003, EPA/Cal 2003).

Formaldehyde is used commonly in homes as an adhesive resin in pressed wood products. There are two types of

formaldehyde resins: urea formaldehyde (UF) and phenol formaldehyde (PF). Products made of urea formaldehyde

readily release formaldehyde. Urea-formaldehyde foam is used to insulate buildings. It can continue to emit

formaldehyde after installation or constituting a source of persistent emission. Formaldehyde is used also in phenolic

and polyacetal plastics, but they are not expected to release formaldehyde (WHO 1989, EPA 2002).

2. Physical and Chemical Properties

Molecular weight (g/mol) 30.03

Melting point (°C) -118

Boiling point (°C) -19.2

Density (g/l at 20 °C, 1 atm) 0.815

Relative density (air =1) 1.04

Explosivity range in air (vol %) 7-73

Solubility: Soluble in water, ethanol, ether, and acetone

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 1.248 µg/m

1 µg/m

= 0.815 ppb

Sources: (CRC, 1994, WHO 1989, Verschueren 2001, HSDB 2003,)

3. Indoor Air Exposure Assessment

Emission sources

There are several indoor sources that can result in increased human exposure to formaldehyde including cigarette

smoke, insulating materials, particle board or plywood furniture containing formaldehyde-based resins, water based

paints, fabrics, household cleaning agents, disinfectants, pesticide formulations, paper products and adhesives

containing formaldehyde used for plastic surfaces, parquet, carpets and other building materials containing urea-

formaldehyde resins. Also gas cookers and open fireplaces emit formaldehyde to indoor air (WHO 1989, COMEAP

1997, Jurvelin et al 2001, EPA/Cal 2003).

The INDEX project Final report

There might be also formaldehyde sinks in indoor environment. Zwiener et al (1999) reported reduction of

formaldehyde concentration between 80% and 87% in chamber air in two hours after the test had begun. This reduction

is due to chemical reaction of formaldehyde and the wool proteins.

Building age was identified a factor for indoor air formaldehyde concentration in UK (Brown VM et al 2002). Indoor

air concentration increased with decreasing age, having highest concentrations in buildings built in 1990’s. A source

identified in this study was particleboard flooring. These findings show that formaldehyde in indoor air is still an issue

in European homes and should not be neglected in the forthcoming indoor air surveys.

Emission rates of 55 diverse materials and consumer products were determined in a chamber study in California (Kelly

et al 1999). The tests showed that among dry products highest formaldehyde emissions were found from bare pressed-

wood materials made with urea-formaldehyde (UF) resins, and from new permanent press fabrics. An acid-cured floor

finish showed the highest emissions from wet products, clearly exceeding those of any dry product.

Puhakka and Karkkainen (1993) studied factors that increased formaldehyde concentrations in indoor air. Their results

showed the amount of chipboard and the air exchange efficiency in a room being the dominant factors. Formaldehyde

emissions from chipboard could be reduced by painting the surfaces. On the other hand elevated relative humidity

increased formaldehyde emissions significantly.

Increased formaldehyde concentrations may be present in hospitals and scientific facilities where formaldehyde is used

as a sterilising and preserving agent, and in living spaces, such as schools, kindergartens, and mobile homes or

apartments where there may be uncontrolled emissions of formaldehyde from tobacco smoking, building materials, and

furniture (WHO 1989).

A French study carried out in 61 recently refurbished ﬂats with no previous history of complaint for odour or speciﬁc

symptoms, showed that aldehyde levels of formaldehyde, acetaldehyde, pentanal, and hexanal were depend on the age

of wall or ﬂoor coverings (renovations less than 1 year old), smoking, and ambient parameters such as carbon dioxide

levels and temperature. Geometric mean and geometric standard deviation of formaldehyde concentrations in bedrooms

were 24.5 µg/m

and 2.0 respectively (Clarisse et al 2003).

Formaldehyde is formed naturally by photochemical reactions in the atmosphere. It is also emitted by fuel combustion

such as motor vehicles and by industrial fuel combustion such as power generators, oil refining processes, copper

plating, and incinerators (EPA/Cal 2003).

Indoor air and exposure concentrations

A large-scale population based indoor air survey has been carried out in UK (Brown VM et al 2002). Geometric mean

and the maximum value of 3-day samples (n=833) were 22.2 µg/m

and 171 µg/m

, respectively. In Helsinki, average

indoor air concentration of formaldehyde, 41.4 µg/m

, was more than tenfold compared to the respective ambient

concentration. Average personal exposure, 26.8 µg/m

, was slightly lower than the respective residential indoor

concentration, but anyway clearly higher than the workplace concentration, 15 µg/m

. In Sweden (Gustafson et al 2004)

average 24-hour personal exposure (n=24, AM=47 µg/m

, std=111) was higher than the respective indoor air (bedroom)

concentration, (n= 24, AM= 26 µg/m

, std = 14). In another campaign of the same study, average 6-day exposure (n=40,

AM=31 µg/m

, std=18) was slightly higher than the respective indoor air concentration (n=40, AM=35 µg/m

, std=22).

Both of these levels were about an order of magnitude higher than respective outdoor air concentration (n=20, AM=3.8

µg/m

, std=1.2). These results show the dominating role of residential indoor emissions for population exposure to

formaldehyde.

In a German study (GerES IV) weekly mean formaldehyde concentration in residential indoor air in a randomly

selected population of children and teenagers was 36 µg/m

(n=58) (Ullrich et al 2002). This level was only slightly

lower than the respective concentrations in last few years demonstrating the limitations of the source control actions.

A large-scale population based formaldehyde study was also carried out in Austria (Hutter et al 2002). Formaldehyde

concentrations in Austrian homes ranged from 8.8 µg/m

to 115 µg/m

(n= 160, AM= 31.3 µg/m

and median =25

µg/m

Sakai et al (2004) compared indoor air concentrations of formaldehyde in Swedish and Japanese dwellings, showing

higher geometric mean values in Japan than in Sweden being, 17.6 µg/m

(n=37) and 8.3 µg/m

(n=27) respectively.

The highest concentration was measured in a detached house with an unvented kerosene heater. In general,

The INDEX project Final report

formaldehyde concentrations were higher in newer (< 10 years) dwellings than in older ones, possibly due to less

natural ventilation and more emissions from building materials.

In Paris, geometric mean and geometric standard deviation of formaldehyde concentrations in recently refurbished ﬂats

in bedrooms were 24.5 µg/m

and 2.0 respectively (Clarisse et al 2003). In a study of 174 homes in UK, the annual

mean indoor concentration ranged from 7 µg/m

to 76 µg/m

, all of them being clearly higher than the mean annual

outdoor concentration of 2 µg/m

(COMEAP 1997). In an Italian study (Cavallo et al 1993), formaldehyde

concentrations in schools and office buildings ranged 8 – 210 µg/m

and 1.5 – 71 µg/m

respectively. VOC emissions

were related to cleaning materials in schools and to carpeting and furniture in offices. Based on the results of a

European survey indoor concentration of formaldehyde in homes and schools ranged from 10 µg/m

to 100 µg/m

(COMEAP 1997). Similar concentration levels of formaldehyde were reported by WHO in conventional homes with no

urea-formaldehyde foam insulation ranging from 25 to 60 µg/m

(WHO 2000).

In 1970’s and 1980’s residential indoor formaldehyde concentrations were high, and therefore formaldehyde in indoor

air became an issue. It has been suggested to be due to off-gassing from new materials and decreased ventilation in

order to decrease energy consumption of house heating. Maximum formaldehyde concentrations were measured in new

mobile homes in 1980’s showing formaldehyde concentration of several mg/m

(WHO 2000). In the NHEXAS study in

Arizona, USA (Gordon et al 1999) median and 90

percentile indoor concentrations, 21 µg/m

and 46 µg/m

, were

about the same levels than those measured in EXPOLIS in European cities. In a Taiwanese study (Li et al 2002), mean

8-hour formaldehyde concentrations in offices ranged 75 µg/m

– 300 µg/m

. Maximum short time concentrations were

as high as 1000 µg/m

based on the continuous monitoring. Building materials were identified as the main contributor

to these high formaldehyde concentrations. Lee and Wang (2004) studied emissions of incense burning in chamber (18

) tests. They found that six tested incense types caused concentrations exceeding the Recommended Indoor Air

Quality Objectives for Office Buildings and Public Places in Hong Kong (HKIAQO) of 100 µg/m

for formaldehyde.

Maximum concentration exceeded 250 µg/m

during incense burning.

Brown et al (2002) studied emissions and concentrations caused by unflued gas heaters in a room chamber of 32.4 m

In normal conditions formaldehyde concentrations ranged from <10 µg/m

to 180 µg/m

. When the gas supply was

restricted due to the backpressure, as high as 2100 µg/m

of formaldehyde was measured.

Elevated formaldehyde concentrations, up to 375 µg/m

, in manufactured houses in US were related to engineered

wood products such as particleboard and plywood bonded with urea-formaldehyde resin in 1970’s (Sexton et al 1989).

Recently measured formaldehyde concentrations in site-built houses were clearly lower, ranging from 8 – 66 µg/m

and

33 – 81 µg/m

prior to occupancy and five months after occupancy respectively (Lindstrom et al 1995). Similar results

were reported by Hodgson et al (2000) showing indoor concentrations in site-built houses in eastern United States,

ranging from 17 µg/m

to 73 µg/m

. Cabinetry materials, passage doors and plywood subfloor were the predominant

sources of formaldehyde in a new manufactured house studied by Hodgson et al (2002).

Studies in Denmark (Granby and Kristensen 1997), Finland (Viskari et al 2000), and Italy (Possanzini et al 1996)

showed average (3-8 months periods) ambient concentrations of formaldehyde being 3.3 µg/m

, 1.3 – 2.8 µg/m

, and

13.7 µg/m

, respectively. In the USA ambient concentrations ranged from 3.3 to 3.7 µg/m

(Grosejan et al 1993, CARB

1999). In Chinese studies ambient levels ranged from 4.1 µg/m

to 13.3 µg/m

, while the mean indoor concentration of

formaldehyde in hotels was more than twofold (Ho et al 2002, Feng et al 2004). Relatively high ambient concentrations

were found in Japan, (Satsumabayashi et al 1995), Brazil (Grosejan et al 2002), Greece (Bakeas et al 2003) and in

Mexico (Baez et al 1995) 3.1–14 µg/m

, 10.8 µg/m

, 17.2 µg/m

and 43.5 µg/m

, respectively.

Cumulative distributions of the indoor and personal exposure concentrations in Europe are presented in Figure 3.1 and

Figure 3.2. Indoor air concentrations of formaldehyde in European countries before 1990 are summarised in Table 3.1

and short time peak concentrations in Table 3.2.

The INDEX project Final report

Table 3.1. Indoor air concentrations of formaldehyde measured before 1990 (adopted from ECA 1990).

Country Environment

Expected source

Mmax

Denmark apartments 9.9

Germany dwellings 48-hour 56 279

Greece dwellings 30-minute 6 - 9 22

France nursery school 30-minute 29

dwellings 24-hour 22 < 70

apartments with complaints 600 2800 Particleboard

collective sites with complaints 3000 UFFI

Norway dwellings < 60 110 Particleboard

Switzerland dwellings 480 2760 UFFI

new buildings 840 Particleboard

The Netherlands 50% of the dwellings with complaints > 120 1000

66% of the schools with complaints > 120 2500

UK buildings without UFFI 57

buildings with UFFI 114 UFFI

AM = aritmethic mean, max = maximum value, UFFI = urea formaldehyde foam insulation

Concentration (µg/m

)Averaging time

Table 3.2. Short time formaldehyde concentrations related to specific microenvironments or emission sources.

Averaging time Reference

MGMmax

Unflued gas heater Room chamber tests Brown et al 2002

- normal burning <10 - 180

- gas supply restricted 2100

Offices Taiwanese study 8-hour 75 - 300 Li et al 2002

peak 2000

Incense burning chamber tests 250 Lee and Wang 2004

Public buildings Schools 5-hour 8 - 210 Cavallo et al 1993

Offices 5-hour 1.5 - 71

Homes in UK with UF foam insulation 30 - 60 min 130 Brown et al 1993

without UF foam insulation 31 - 60 min 70

Houses in US manufactured 30-min 26 - 59 45 Hodgson et al 2000

site-built 30-min 17 - 73 43

Urban dwellings Sweden 24-hour 8.3 19 Sakai et al 2004

Japan 24-hou

18 73

AM = arithmetic mean, GM = geometric mean

Environment or

emission source

Concentration (µg/m

)

The INDEX project Final report

100

0 102030405060708090

Indoor concentration (µg/m

)

Cumulative frequency (%)

Hel

Fre

Swe

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of formaldehyde in Helsinki (Hel, n=15,

Jurvelin et al 2001), in England (UK, n=833, Brown VM et al 2002), in the French Survey (Fre, n=201, Kirchner 2004),

and in Sweden (Swe, n=40, Gustafson et al 2004).

100

-20 0 20 40 60 80 100 120 140 160 180 200

Exposure (µg/m

)

Cumulative frequency (%)

Hel

Swe-24h

Swe-6day

Figure 3.2. Cumulative frequency distribution of personal exposure concentrations of formaldehyde in Helsinki (Hel,

48-hour average, n= 15, Jurvelin et al 2001) and Sweden (Swe-24h = 24-hour average, n=24, and Swe-6day = 6-day

average, n=40, Gustafson et al 2004).

The INDEX project Final report

4. Toxicokinetics

Absorption

Owing to its solubility in water, formaldehyde is rapidly absorbed in the respiratory and gastrointestinal tracts and

metabolized. Over 90% of inhaled formaldehyde gas is absorbed in the upper respiratory tract of rats and monkeys. In

rats, it is absorbed in the nasal passages; in monkeys, it is also absorbed in the nasopharynx, trachea and proximal

regions of the major bronchi. Although formaldehyde or its metabolites can penetrate human skin – it induces allergic

contact dermatitis in humans – dermal absorption appears to be very slight (WHO, 1989; IARC, 1995).

Distribution

Owing to its rapid metabolism, exposure of humans, monkeys or rats to formaldehyde by inhalation does not alter the

concentration of endogenous formaldehyde in the blood, which is about 2–3 mg/litre for each of the three species.

Intravenous administration of formaldehyde to dogs, cats and monkeys does not result in accumulation of formaldehyde

in the blood, because of its rapid conversion to formate. In dogs, orally administered formaldehyde results in a rapid

increase in formate levels in the blood. Following a 6-hour inhalation exposure of rats to

C-formaldehyde,

radioactivity was extensively distributed in other tissues, the highest concentration occurring in the oesophagus,

followed by the kidneys, liver, intestine and lung, indicating that absorbed

C-formaldehyde and its metabolites are

rapidly removed by the mucosal blood supply and distributed throughout the body (WHO, 1989).

Metabolism and elimination

Formaldehyde reacts virtually instantaneously with primary and secondary amines, thiols, hydroxyls and amides to

form methylol derivatives. Formaldehyde acts as an electrophile and can react with macromolecules such as DNA,

RNA and protein to form reversible adducts or irreversible cross-links. Absorbed formaldehyde can be oxidized to

formate along three different pathways, and can be exhaled as carbon dioxide or incorporated into biological

macromolecules via tetrahydrofolate-dependent one-carbon biosynthetic pathways. In the body, formaldehyde is

produced in small quantities as a normal metabolite and also in the oxidative demethylation of xenobiotics; it may

therefore be found in the liver (IARC, 1995). Formaldehyde disappears from the plasma with a half-time of about 1–1.5

minutes, most of it being converted to carbon dioxide and exhaled via the lungs. Smaller amounts are excreted in the

urine as formate salts and several other metabolites (WHO, 1987).

5. Health effects

Effects of short-term exposure

Predominant signs of short-term exposure to formaldehyde in humans are irritation of the eyes, nose and throat, together

with concentration-dependent discomfort, lachrymation, sneezing, coughing, nausea, dyspnoea and finally death (Table

5.1). Symptoms are often more severe at the start of exposure than after minutes or hours, when they gradually

diminish.

Table 5.1: Effects of formaldehyde in humans after short-term exposure

Concentration range

or average (mg/m

)

Time range or average Health effects in general population

0.03 Repeated exposure Odour detection threshold (10th percentile)

0.18 Repeated exposure Odour detection threshold (50th percentile)

0.6 Repeated exposure Odour detection threshold (90th percentile)

0.1 – 3.1 Single and repeated exposure Throat and nose irritation threshold

0.6–1.2 Single and repeated exposure Eye irritation threshold

0.5 – 2 3–5 hours Decreased nasal mucus flow rate

2.4 40 minutes on 2 successive days with

10 minutes of moderate exercise on

second day

Post-exposure (up to 24 hours) headache

The INDEX project Final report

2.5 – 3.7

– b

Biting sensation in eyes and nose

3.7 Single and repeated exposure Decreased pulmonary function only at heavy

exercise

5 – 6.2 30 minutes Tolerable for 30 minutes with lachrymation

12 – 25

– b

Strong lachrymation, lasting for 1 hour

37 – 60

– b

Pulmonary oedema, pneumonia, danger to life

60 – 125

– b

Death

Frequency of effect in population.

Time range or average unspecified.

Sources: WHO, 1989; IARC, 1995; WHO, 1987

Numerous acute controlled and occupational human exposure studies have been conducted with both asthmatic and

normal subjects to investigate formaldehyde’s irritant effects on the eyes and the upper respiratory tract. Feinman

(1988) states that most people cannot tolerate exposures to more than 6 mg/m

(5 ppm) formaldehyde in air

A series of pulmonary function studies has been conducted in healthy nonsmokers and asthmatics exposed to

formaldehyde vapour; generally, lung function was unaltered. Fifteen healthy nonsmokers and 15 asthmatics were

exposed to 2.4 mg/m

formaldehyde for 40 minutes to determine whether acute exposures could induce asthmatic

symptoms (Schachter et al. 1986; Witek. et al. 1987). No significant airway obstruction, changes in pulmonary function

or bronchial hyperreactivity were noted. Similar observations were made on a group of 15 hospital laboratory workers

who had been exposed to formaldehyde (Schachter et al. 1987).

Healthy nonsmokers and asthmatics were exposed to 3.7 mg/m

formaldehyde for 1 or 3 hours, either at rest or when

engaged in intermittent heavy exercise. No significant changes in pulmonary function and nonspecific airway reactivity

were observed among asthmatic subjects. Small decreases (< 5%) in pulmonary function were observed in healthy non-

smokers exposed to formaldehyde while engaged in heavy exercise. Two normal and two asthmatic subjects had

decrements greater than 10% (Sauder et al. 1986; Green et al. 1987).

When 15 asthmatic subjects were exposed for 90 minutes to concentrations of 0.008–0.85 mg/m

formaldehyde, no

change in pulmonary function was seen, and there was no evidence of an increase in bronchial reactivity (Harving, et

al., 1990).

In a study of controlled exposure to formaldehyde, 18 subjects, 9 of whom had complained of adverse effects from

urea–formaldehyde foam insulation installed in their homes, were exposed to 1.2 mg/m

formaldehyde, or to off-gas

products of urea–formaldehyde foam insulation containing 1.5 mg/m

formaldehyde, for 90 minutes (Day et al., 1984).

No statistically or clinically significant change in pulmonary function was seen either during or 8 hours after exposure,

and no evidence was obtained that urea–formaldehyde foam insulation off-gas acts as a lower airway allergen.

Kulle et al. (1987; 1993) exposed 19 healthy subjects to 0, 1.2, and 2.5 mg/m

(0, 1.0, and 2.0 ppm) for 3-hour periods

and asked them to note symptoms of eye and nose/throat irritation and to rate severity on a 0-3 scale: 0=none; l=mild

(present but not annoying); 2=moderate (annoying); and 3=severe debilitating). Ten of the subjects were also exposed

to 0.6 mg/m

(0.5 ppm) and nine were exposed to 4 mg/m

(3 ppm) for 3-hour periods. The frequencies of subjects

reporting eye irritation or nose/throat irritation increased with increasing exposure concentration, especially at

concentrations ≥ 1.2 mg/m

. Under nonexposed conditions, 3/19 subjects noted mild nose/throat irritation and l/19

noted mild eye irritation. At 0.6 mg/m

, l/10 subjects noted mild nose/throat irritation, but none reported eye irritation.

Frequencies for subjects with mild or moderate eye irritation were 4/19 at 1.2 mg/m

(1 moderate), 10/19 at 2.5 mg/m

(4 moderate), and 9/9 at 4 mg/m

(4 moderate). The increased frequency for eye irritation (compared with controls) was

statistically significant at 2.5 mg/m

. Frequencies for mild nose/throat irritation were l/19 at 1.2 mg/m

, 7/19 at 2.5

mg/m

, and 2/9 at 4 mg/m

. Compared with control frequency for nose/throat irritation, only the response at 2.5 mg/m

was significantly elevated.

Weber-Tschopp et al. (1977) exposed a group of 33 healthy subjects for 35 minutes to concentrations of formaldehyde

that increased during the period from 0.04 to 3.9 mg/m

(0.03 to 3.2 ppm); another group of 48 healthy subjects was

exposed to 0.03, 1.2, 2.1, 2.8, and 4.0 ppm for 1.5 minute intervals. Eye and nose irritation were reported on an l-4 scale

(l=none to 4=strong) in both experiments, and eye blinking rate was measured in the second experiment. Average

indices of eye and nose irritation were increased in both experiments to a small, but statistically significant at 1.2 ppm

compared with indices for nonexposed controlled conditions. The published report of this study graphically showed

average severity scores of about 1.3-l.4 for both indices at 1.2 ppm compared with 1.0-l.1 for non exposed conditions.

The average severity score was increased to a greater degree at higher concentrations, but was less than about 2.5 at the

The INDEX project Final report

highest exposure concentration, 4 ppm. Average rates of eye blinking were not significantly affected at 1.2 ppm, but

were statistically significantly increased at 2.1 ppm (about 35 blinks/minute at 2.1 ppm versus about 22 blinks/minutes

under nonexposed conditions).

The relation of chronic respiratory symptoms and pulmonary function to formaldehyde (HCHO) in homes was studied

in a sample of 298 children (6-15 years of age) and 613 adults. HCHO measurements were made with passive samplers

during two 1-week periods. Data on chronic cough and phlegm, wheeze, attacks of breathlessness, and doctor diagnoses

of chronic bronchitis and asthma were collected with self-completed questionnaires. Peak expiratory flow rates (PEFR)

were obtained during the evenings and mornings for up to 14 consecutive days for each individual. Significantly greater

prevalence rates of asthma and chronic bronchitis were found in children from houses with HCHO levels 74-147 µg/m³

(60-120 ppb) than in those less exposed, especially in children also exposed to environmental tobacco smoke. In

children, levels of PEFR decreased linearly with HCHO exposure, with the estimated decrease due to 74 µg/m³ of

HCHO equivalent to 22% of PEFR level in nonexposed children. The effects in asthmatic children exposed to HCHO

below 61 µg/m³ were greater than in healthy ones. The effects in adults were less evident: decrements in PEFR due to

HCHO over 49 µg/m³ were seen only in the morning, and mainly in smokers (Krzyzanowski et al., 1990).

Studies in children, including the Krzyzanowski study above, indicate adverse health impacts in children at

concentrations as low as 37 µg/m

3 (

30 ppb). Wantke et al. (1996) reported that formaldehyde-specific IgE and

respiratory symptoms were reduced when children transferred from schools with formaldehyde concentrations of 53 to

92 µg/m

3 (

43 to 75 ppb) to schools with concentrations of 28 to 36 µg/m

(23 to 29 ppb). Garrett et al. (1999) reported

increased sensitization associated with the formaldehyde level in children’s homes which had a median value of 15.8

µg/m

(12.6 ppb). And Franklin et al. (2000) reported significantly higher exhaled nitric oxide, an indicator of airway

inflammation, in the breath of children living in homes with formaldehyde concentrations greater than 61 µg/m

(50

ppb) than in the breath of those children living in homes with formaldehyde levels below 61 µg/m

. These human

studies are not entirely consistent with each other, and there is potential for confounding in each. Nevertheless, taken

together, they suggest that children may be more sensitive to formaldehyde toxicity than adults.

Pulmonary function has been assessed in residents of mobile and conventional homes (Broder et al., 1988) and mobile

offices (Main & Hogan, 1983)) exposed to concentrations of 0.007–2.0 mg/m

. No changes were seen in pulmonary

function or airway resistance.

Eleven healthy subjects and nine patients with formalin skin sensitization were exposed to 0.5 mg/m³ formaldehyde for

2 hours (Pazdrak et al., 1993). Nasal lavage was performed prior to and 5 to 10 minutes, 4 hours, and 18 hours after

exposure. Rhinitis was reported and increases in the number and proportion of eosinophils, elevated albumin and

increased protein levels were noted in nasal lavage fluid 4 and 18 hours after exposure. No differences were found

between patients with skin sensitization and healthy subjects.

Summary of short-term exposure effect levels

Studies of humans under controlled conditions clearly indicate that acute exposures to air concentrations ranging from

0.5 to 3.7 mg/m

(0.4 to 3 ppm):

- induce reversible eye, nose, and throat irritation (Andersen and Mølhave 1983; Bender et al. 1983; Day et al.

1984; Gorski et al. 1992; Krakowiak et al. 1998; Kulle 1993; Kulle et al. 1987; Pazdrak et al. 1993; Schachter et

al. 1986; Weber-Tschopp et al. 1977; Witek et al. 1986);

- produce changes in nasal lavage fluid contents, indicative of irritation of the nasal epithelium (Gorski et al. 1992;

Krakowiak et al. 1998; Pazdrak et al. 1993); and

- do not consistently or markedly affect pulmonary function variables in most individuals (Andersen and Mølhave

1983; Day et al. 1984; Gorski et al. 1992; Green et al. 1987; Harving et al. 1986, 1990; Kulle et al. 1987;

Nordman et al. 1985; Sauder et al. 1986; Schachter et al. 1986; Witek et al. 1986).

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

0.6

EXP

1.2

EXP

0.53

BC05

0.94

1h-ADJ-BC05

Mild and moderate eye irritation Volunteers, 3h Kulle et al. (1987) OEHHA 1999;

REL: 0.094

0.5

EXP

Nasal and eye irritation

Volunteers, 2h

(potentially sensitive

population)

Pazdrak et al. (1993) ATSDR 1999;

MRL: 0.05

0.24

STAT

Respiratory tract irritation cited in Sloof et al.

(1992)

RIVM 1992

0.1 Nose and throat irritation Human, short-term cited in IPCS/WHO, WHO 2000

The INDEX project Final report

1989

(0.037) Respiratory symptoms (depending on study) Children; additional

investigation requested

Krzyzanowski et al.,

(1990); Wantke et al.

(1996); Garrett et al.

(1999); Franklin et al.

(2000)

UCLA 2001

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of Acute Reference Exposure Level (for a 1-hour exposure)

Study Kulle et al. (1987)

Study population 19 nonasthmatic, nonsmoking human subjects

Exposure method 0.6- 3.6 mg/m³ (0.5-3.0 ppm)

Critical effects mild and moderate eye irritation

LOAEL 1.2 mg/m³ (1 ppm)

NOAEL 0.6 mg/m³ (0.5 ppm)

Benchmark concentration 0.44 ppm (BC

05)

Exposure duration 3 hours

Extrapolated 1 hour concentration 0.94 mg/m³ [0.76 ppm (0.44

ppm* 3 h = C

* 1 h )]

LOAEL uncertainty factor not required in BC approach

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 10

Reference Exposure Level 0.076 ppm (0.094 mg/m³; 94 µg/m³)

The recommended REL was estimated by a benchmark concentration (BC

05) approach, using logprobit analysis

(Crump, 1984; Crump and Howe, 1983). The BC

05 is defined as the 95% lower confidence limit of the concentration

expected to produce a response rate of 5%. The resulting BC05 from this analysis was 0.53 mg/m³ formaldehyde. This

value was adjusted to a 1-hour duration using the formula Cn * T = K, where n = 2 (AICE, 1989), resulting in a value of

0.94 mg/m³. An uncertainty factor (UF) of 10 was used to account for individual variation.

Generally an uncertainty factor of 3 would be used with the BC

05 for intraindividual variability, since the BC05 accounts

for some degree of individual variation. However, information from the literature indicates a wide variability in

response to formaldehyde irritancy including reports of irritation (NIOSH HHE reports 1981-1996; Liu et al. 1991;

Horvath et al, 1985) or cellular changes associated with irritation and an immune response at levels below the one-hour

extrapolated BC

05 (Pazdrak et al, 1993; Gorski et al, 1992). For these reasons, we used an uncertainty factor of 10 to

account for intraindividual variability in the human population.

ATSDR (1999)

Agency for Toxic Substances and Disease Registry

Derivation of the Minimal Risk Level (MRL):

ATSDR has derived an acute inhalation MRL of 0.05 mg/m³ (0.04 ppm) on the basis of clinical symptoms (increased

itching, sneezing, mucosal congestion, transient burning sensation of the eyes and of the nasal passages) and nasal

alterations (elevated eosinophil counts and a transient increase in albumin content of nasal lavage fluid) in humans

(Pazdrak et al. 1993). This MRL is based on a minimal LOAEL of 0.5 mg/m³ and an uncertainty factor of nine (three

for use of a minimal LOAEL and three for human variability).

The Anderson and Mølhave (1983) study identified an apparent effect level at 0.3 mg/m³ (0.2 ppm), based on subjective

reports of irritation that is lower than the effect levels at 0.43-0.5 mg/m³ (0.35-0.4 ppm) in the studies by Pazdrak et al.

(1993), Krakowiak et al. (1998), and Bender et al. (1983), which used more objective measures of acute irritation

(eosinophil counts and protein concentrations in nasal lavage fluid or time to first reporting of irritation). Because of the

use of objective measures of toxicity and the general weight of the available data indicating that some people will not

experience eye or upper respiratory tract irritation from formaldehyde even at 1.2 mg/m³ (see Day et al. 1984; Kulle et

al. 1987, Weber Tschopp et al. 1977, and Witek et al. 1986), the Pazdrak et al. (1993) LOAEL of 0.5 mg/m³ was

The INDEX project Final report

considered a minimal LOAEL in a group of potentially sensitive individuals (some subjects had dermal hypersensitivity

to formaldehyde) and selected as the basis of the acute MRL.

RIVM (1992)

National Institute of Public Health and the Environment - The Netherlands

Sensory reactions are apparently the most typical effects in the non-industrial indoor environment (Sloof et al., 1992). A

large number of observations of people in various settings support a conclusion that the generally observed range over

which more than 95 % of people respond is 125-3750 µg/m

(0.10-3.1 ppm). According to the Dutch Health Council

(1984) and WHO (1989) the lowest statistically significant effect concentration for respiratory tract irritation is 240

µg/m

(LOAEL) (0.20 ppm) (Sloof et al., 1992).

The Dutch Health Council (1984) recommended the following three limit values (Sloof et al., 1992):

120 µg/m

as ceiling value (based on 30 min. means);

40 µg/m

as 98-percentile value (based on 24 h means);

30 µg/m

as 95-percentile value (based on 24 h means).

WHO (2001)

Air Quality Guidelines for Europe 2000

The lowest concentration that has been associated with nose and throat irritation in humans after short-term exposure is

0.1 mg/m

, although some individuals can sense the presence of formaldehyde at lower concentrations.

To prevent significant sensory irritation in the general population, an air quality guideline value of 0.1 mg/m

as a 30-

minute average is recommended. Since this is over one order of magnitude lower than a presumed threshold for

cytotoxic damage to the nasal mucosa, this guideline value represents an exposure level at which there is a negligible

risk of upper respiratory tract cancer in humans.

Effects of long-term exposure (noncancer)

Long term exposure to elevated levels of formaldehyde in the occupational setting has been shown to result in upper

and lower airway irritation and eye irritation in humans; degenerative, inflammatory and hyperplastic changes of the

nasal mucosa in humans and animals. The voluminous body of data describing these effects has been briefly

summarized below. For the sake of brevity, only the studies that best represent the given effects are presented.

Ritchie and Lehnen (1987) reported a dose-dependent increase in health complaints (eye and throat irritation, and

headaches) in 2000 residents living in 397 mobile and 494 conventional homes, that was demonstrated by logistic

regression. Complaints of symptoms of irritation were noted at concentrations of 0.12 mg/m³ formaldehyde or above.

Similarly, Liu et al. (1991) found that exposure to 0.135 mg/m

formaldehyde exacerbated chronic respiratory and

allergy problems in residents living in mobile homes.

Employees of mobile day-care centers (66 subjects) reported increased incidence of eye, nose and throat irritation,

unnatural thirst, headaches, abnormal tiredness, menstrual disorders, and increased use of analgesics as compared to

control workers (Olsen and Dossing, 1982). The mean formaldehyde concentration in these mobile units was 0.43

mg/m

(range = 0.24 - 0.55 mg/m3). The exposed workers were exposed in these units a minimum of 3 months. A

control group of 26 subjects in different institutions was exposed to a mean concentration of 0.06 mg/m

formaldehyde.

Occupants of houses insulated with urea-formaldehyde foam insulation (UFFI) (1726 subjects) were compared with

control subjects (720 subjects) for subjective measures of irritation, pulmonary function (FVC, FEV

1, FEF25-75, FEF50),

nasal airway resistance, odour threshold for pyridine, nasal cytology, and hypersensitivity skin-patch testing (Broder et

al., 1988). The mean length of time of exposure to UFFI was 4.6 years. The mean concentration of formaldehyde in the

UFFI-exposed group was 0.053 mg/m³, compared with 0.043 mg/m³ for the controls. A significant increase in

symptoms of eye, nose and throat irritation was observed in subjects from UFFI homes, compared with controls. No

other differences from control measurements were observed.

An increase in severity of nasal epithelial histological lesions, including loss of cilia and goblet cell hyperplasia (11%),

squamous metaplasia (78%), and mild dysplasia (8%), was observed in 75 wood products workers exposed to between

0.1 and 1.1 mg/m

3 formaldehyde for a mean duration of 10.5 years (range = 1 - 39 years), compared to an equal number

of control subjects (Edling et al., 1988). Only three exposed men had normal mucosa. A high frequency of symptoms

relating to the eyes and upper airways was reported in exposed workers. Nasal symptoms included mostly a runny nose

The INDEX project Final report

and crusting. The histological grading showed a significantly higher score for nasal lesions when compared with the

referents (2.9 versus 1.8). Exposed smokers had a higher, but non-significant, score than ex-smokers and non-smokers.

When relating the histological score to duration of exposure, the mean histological score was about the same regardless

of years of employment. In addition, no difference in the histological scores was found between workers exposed only

to formaldehyde and those exposed to formaldehyde and wood dust.

Alexandersson and Hedenstierna (1989) evaluated symptoms of irritation, spirometry, and immunoglobulin levels in 34

wood workers exposed to formaldehyde over a 4-year period. Exposure to 0.5 – 0.6 mg/m³ (0.4 - 0.5 ppm)

formaldehyde resulted in significant decreases in FVC, FEV

1, and FEF25-75. Removal from exposure for 4 weeks

allowed for normalization of lung function in the non-smokers.

Symptoms of irritation were reported by 66 workers exposed for 1 - 36 years (mean = 10 years) to a mean concentration

of 0.26 mg/m

formaldehyde (Wilhelmsson and Holmstrom, 1992). Controls (36 subjects) consisted of office workers in

a government office and were exposed to a mean concentration of 0.09 mg/m

formaldehyde. The significant increase in

symptoms of irritation in exposed workers did not correlate with total serum IgE antibody levels. However, 2 exposed

workers, who complained of nasal discomfort, had elevated IgE levels. In another occupational health study, 37

workers, who were exposed for an unspecified duration to formaldehyde concentrations in the range of 0.004 to 0.090

mg/m³, reported ocular irritation; however, no significant serum levels of IgE or IgG antibodies to formaldehyde-human

serum albumin were detected (Grammer et al., 1990). An epidemiological study of the effects of formaldehyde on 367

textile and shoe manufacturing workers employed for a mean duration of 12 years showed no significant association

between formaldehyde exposure, pulmonary function (FVC, FEV

1, and PEF) in normal or asthmatic workers, and

occurrence of specific IgE antibodies to formaldehyde (Gorski and Krakowiak, 1991). The concentrations of

formaldehyde did not exceed 0.75 mg/m

. Workers (38 total) exposed for a mean duration of 7.8 years to 0.14 – 2.60

mg/m³ (mean = 0.41 mg/m³) formaldehyde were studied for their symptomatology, lung function, and total IgG and IgE

levels in the serum (Alexandersson and Hedenstierna, 1988). The control group consisted of 18 unexposed individuals.

Significant decrements in pulmonary function (FVC and FEV

1) were observed, compared with the controls. Eye, nose,

and throat irritation was also reported more frequently by the exposed group, compared with the control group. No

correlation was found between duration of exposure, or formaldehyde concentration, and the presence of IgE and IgG

antibodies.

Holmstrom et al. (1989) examined histological changes in nasal tissue specimens from a group of 70 workers in a

chemical plant that produced formaldehyde and formaldehyde resins for impregnation of paper, a group of 100 furniture

factory workers working with particle board and glue components, and a nonexposed control group of 36 office workers

in the same village as the furniture factories. Mean durations of employment in the groups were 10.4 years (SD 7.3,

range 1–36 years) for the chemical workers and 9.0 years (SD 6.3, range 1–30 years) for the furniture workers.

Estimates of personal breathing zone air concentrations ranged from 0.05 to 0.5 mg/m³ (median 0.29±0.16 ppm) for the

chemical workers, from 0.20 to 5 mg/m

(median 0.25±0.05 ppm) for the furniture workers, and from 0.09 to 0.16

mg/m³ in the late summer for the office workers with a year-round office worker median reported as 0.09 mg/m

with

no standard deviation. The mean wood dust concentration in the furniture factory was reported to have been between 1

and 2 mg/m

. Nasal mucosa specimens were taken from the medial or inferior aspect of the middle turbinate. Histology

scores were assigned to each specimen based on a 0–8 scale, identical to the scale used by Edling et al. (1988; described

previously). Nasal mucosal biopsy sections for each subject were examined and assigned scores as follows: 0 - normal

respiratory epithelium; 1- loss of ciliated epithelium cells; 2 - mixed cuboid/squamous epithelium, metaplasia; 3 -

stratified squamous epithelium; 4 - keratosis; 5 - budding of epithelium; 6 - mild or moderate dysplasia; 7 - severe

dysplasia; and 8 - carcinoma.

Nasal histology scores ranged from 0 to 4 (mean 2.16; n=62) for the chemical workers, from 0 to 6 (mean 2.07; n=89)

for the furniture workers, and from 0 to 4 (mean 1.46; n=32) for the office workers. The mean histological score for the

chemical workers, but not the furniture workers, was significantly different from the control score, thus supporting the

hypothesis that the development of the nasal lesions is formaldehyde-related and not obligatorily related to possible

interaction between formaldehyde and wood dust. The most severe epithelial change found (light or moderate epithelial

dysplasia) was found in two furniture workers. Among the chemical workers (not exposed to wood dust), loss of cilia,

goblet cell hyperplasia and cuboidal and squamous cell metaplasia replacing the columnar epithelium occurred more

frequently than in the control group of office workers. Within both groups of formaldehyde-exposed workers, no

evidence was found for associations between histological score and duration of exposure, index of accumulated dose, or

smoking habit.

The INDEX project Final report

Summary of long-term exposure effect levels (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

0.09

Study

0.032

ADJ

0.26

Study

0.093

ADJ

Nasal and eye irritation, nasal obstruction, and

lower airway discomfort; histopathological nasal

lesions including rhinitis, squamous metaplasia,

and dysplasia

Occupational, 10y

average;

Wilhelmsson and

Holmstrom, 1992;

supported by Edling et

al., 1988

OEHHA 1999;

REL: 0.003

0.3

Study

Mild irritation of the eyes and upper respiratory

tract and mild damage to the nasal epithelium

Occupational, 10.4y Holmstrom et al., 1989

ATSDR 1999;

MRL: 0.01

0.31

STAT

Sensory irritation

for low but significant

percentage of exposed

workers

Weighting the total

body of data

NIWL 2003

0.12 Symptoms of irritation

LOAEL may be lower

only for a very small

proportion of the

population

well conducted studies Health Canada

1999; (noTC)

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment

Derivation of Chronic Reference Exposure Level (REL):

Studies Wilhelmsson and Holmstrom, 1992; supported by Edling et al.,

1988

Study population Human chemical plant workers (66 subjects)

Exposure method Discontinuous occupational exposure

Critical effects Nasal and eye irritation, nasal obstruction, and lower airway

discomfort; histopathological nasal lesions including rhinitis,

squamous metaplasia, and dysplasia

LOAEL Mean of 0.26 mg/m

(range = 0.05 to 0.6 mg/m

) (described as

exposed group)

NOAEL Mean of 0.09 mg/m

(described for control group of office workers)

Exposure continuity 8 hours/day, 5 days/week (assumed)

Exposure duration 10 years (average); range = 1-36 years

Average occupational concentration 0.032 mg/m

for NOAEL group (0.09 x 10/20 x 5/7)

Human equivalent concentration 0.032 mg/m

LOAEL uncertainty factor 1

Subchronic uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 10

Inhalation reference level 0.003 mg/m

(3 µg/m

; 0.002 ppm; 2 ppb)

The Wilhelmsson and Holmstrom study was selected by the Office of Environmental Health Hazard Assessment

(OEHHA, reference) for the derivation of the Chronic Reference Exposure Level (REL), because it was a human

occupational study that contained a LOAEL and a NOAEL and contained a reasonable number of subjects. The

supporting occupational study by Edling et al. (1988) noted similar sensory irritation results due to long-term

formaldehyde exposure. In addition, nasal biopsies from exposed workers in the Edling et al. (1988) study exhibited

nasal epithelial lesions similar to those found in subchronic and chronic animal studies.

For comparison with the proposed REL of 3 µg/m

, we estimated a REL from Edling et al. (1988). A median

concentration of 0.6 mg/m

was determined for the LOAEL from the TWA range of 0.1- 1.1 mg/m

. A NOAEL was not

reported. The average continuous occupational concentration was 0.2 mg/m

(0.6 x 10/20 x 5/7) and the exposure

duration was 10.5 years (range = 1 – 39 years). Application of a UF of 10 for intraspecies variability and a UF of 10 for

estimation of a NOAEL from the LOAEL would result in a REL of 2 µg/m

(2 ppb).

Data Strengths and Limitations for Development of the REL: The strengths of the inhalation REL include the use of

human exposure data from workers exposed over a period of years and the observation of a NOAEL. In addition, a

The INDEX project Final report

number of well conducted animal studies supported the derivation of the REL. The major areas of uncertainty are the

uncertainty in estimating exposure in the occupational studies and the potential variability in exposure concentration.

Table 5.2 presents a summary of potential RELs based on chronic and subchronic animal studies. The toxicological

endpoint was nasal lesions, consisting principally of rhinitis, squamous metaplasia, and dyplasia of the respiratory

epithelium.

Table 5.2: Summary of chronic and subchronic formaldehyde studies in experimental animals

Study Animal Exposure

Duration

LOAEL/

NOAEL

(mg/m

)

HEC

adj.

(mg/m

)

Cumulativ

REL

(mg/m3)

Woutersen et al.,

1989

rat 28 m 9.8 / 1.0 0.06 30 2

Kerns et al., 1983 rat 24 m 2.0 / NA 0.1 300 0.3

Monticello et al.,

1996

rat 24 m 6.01 / 2.05 0.1 30 4

Kamata et al., 1997 rat 24-28 m 0.30 / NA 0.02 100 0.2

Appelman et al.,

1988

rat 52 w 9.4 / 1.0 0.06 30 2

Rusch et al., 1983 rat 26 w 2.95 / 0.98 0.2 30 7

Kimbell et al., 1997 rat 26 w 6 / 2 0.1 30 3

Wilmer et al., 1989 rat 13 w 4 / 2 0.2 300 0.7

Woutersen et al.,

1987

rat 13 w 9.7 / 1.0 0.03 100 0.3

Zwart et al., 1988 rat 13 w 2.98 / 1.01 0.2 300 0.7

Kerns et al., 1983 mouse 24 m 2.0 / NA 0.05 100 0.5

Maronpot et al., 1986 mouse 13 w 10.1 / 4.08 0.09 100 0.9

Rusch et al., 1983 monkey 26 w 2.95 / 0.98 none 300 4

The most striking observation is the similarity of potential RELs among the rat chronic studies (exposures > 26 weeks)

that contain a NOAEL. The range of RELs from these animal studies, 2 – 7 mg/m

, is comparable to the proposed REL

based on a human study. Another related observation is that the NOAEL and LOAEL are similar among all the studies,

regardless of exposure duration. The NOAEL and LOAEL are generally in the range of 1-2 mg/m3 and 2-10 mg/m3,

respectively, with the exception of the study by Kamata et al. (1997). These results indicate that the formation of

formaldehyde-related nasal lesions are more concentration dependent than time, or dose, dependent. A limitation of a

majority of the occupational studies is their high reliance on surveys and other methods that focus on sensory irritation.

Such sensory irritant results, as exhibited in the Wilhelmsson and Holmstrom (1992) study, may be more related to

recurrent acute injury rather than a true chronic injury. The concentration dependent nature of the nasal lesions in the

supporting animal studies, and suggested in the supporting human nasal biopsy study, would also imply that the nasal

cavity endpoint may be a recurrent acute effect. However, Kerns et al. (1983) and Kamata et al. (1997) clearly

demonstrated that near the LOAEL, increasing exposure durations would result in nasal lesions at lower formaldehyde

concentrations. Also, the rat study by Woutersen et al. (1989) demonstrated that subchronic exposure to formaldehyde

concentrations that produce nasal lesions could result in lifelong changes of the nasal epithelium. These findings

substantiate the chronic nature of the nasal/upper airway injury that results from long-term formaldehyde exposure.

ATSDR (1999)

Agency for Toxic Substances and Disease Registry

Derivation of the chronic Minimal Risk Level (MRL).

The chronic inhalation MRL of 0.01 mg/m

was derived from a LOAEL of 0.3 mg/m

in humans for nasal effects

(Holmstrom et al., 1989). This minmal LOAEL was divided by a total uncertainty factor of 30 (3 for use of a LOAEL

and 10 for human variability). The exposure concentration was not adjusted to a continuous exposure basis based on

evidence that concentration is more important than the product of concentration and duration of exposure in

determining the severity of formaldehyde-induced epithelial damage in the upper respiratory tract (Wilmer et al. 1987).

The INDEX project Final report

National Institute for Working Life (NIWL) - Sweden (2003)

Nordic Expert Group for Criteria Documentation of Health Risks from Chemicals (NEG) and Dutch Expert Committee

on Occupational Standards (DECOS)

Summary of the report n°132.Formaldehyde (Scientific basis for an occupational exposure limit):

Inhaled formaldehyde is almost completely absorbed in the upper respiratory tract in rodents. Formaldehyde is a normal

metabolite in mammalians and is rapidly metabolised to formate, which may be further oxidised to carbon dioxide.

Final elimination occurs via exhalation and via the kidneys. The target organs of formaldehyde vapour are the eyes,

nose, and throat. The effects of concern are sensory irritation and cytotoxicity-induced regenerative hyperplasia and

metaplasia of the nasal respiratory epithelium accompanied by nasal carcinomas in rats. Weighting the total body of

data 0.31 mg/m

(0.25 ppm) formaldehyde is the LOAEL at which sensory irritation may occur in a low but significant

percentage of exposed workers. The majority of short- and long-term inhalation animal studies reveal a NOAEL of 1.2-

2.5 mg/m

(1-2 ppm). However, in a few studies slight histopathological changes of the nasal respiratory epithelium

were observed at 0.4-2.5 mg/m

(0.3-2 ppm).

Formaldehyde is genotoxic in a variety of experimental systems, including rodents in vivo. There is overwhelming

evidence that high, cytotoxic formaldehyde vapour concentrations (≥10 ppm) can induce nasal cancer in rats. A large

body of data suggests an association between the cytotoxic, genotoxic, and carcinogenic effects of formaldehyde. The

crucial role of tissue damage followed by hyperplasia and metaplasia of the nasal respiratory epithelium in

formaldehyde carcinogenesis has been demonstrated in a convincing way. Thus, formaldehyde in non-cytotxic

concentrations most probably cannot act as a complete carcinogen. However, if human exposure to formaldehyde is

accompanied by recurrent tissue damage at the site of contact, formaldehyde may be assumed to have carcinogenic

potential in man.

Formaldehyde-induced allergic contact dermatitis has been estimated to occur in 3-6% of the population, and skin

sensitisation by direct skin contact has been induced with formaldehyde solutions. There is no consistent evidence of

formaldehyde being a respiratory sensitiser.

Health Canada (1999)

Canadian Environmental Protection Act (CEPA).

Health Canada did not derive a Tolerable Concentration (TC) for formaldehyde. However, there are considered to be

sufficient data from clinical studies and cross-sectional surveys of human populations, as well as supporting

observations from experimental studies conducted with laboratory animals, to indicate that the irritant effects of

formaldehyde on the eyes, nose and throat occur at lowest concentration. Although individual sensitivity and exposure

conditions such as temperature, humidity, duration and co-exposure to other irritants are likely to influence response

levels, in well-conducted studies, only a very small proportion of the population experiences symptoms of irritation

following exposure to less than or equal to 0.12 mg/m

formaldehyde. This is less than the levels that reduce

mucociliary clearance in the anterior portion of the nasal cavity in available clinical studies in human volunteers (0.3

mg/m

) and induce histopathological effects in the nasal epithelium in cross-sectional studies of formaldehyde-exposed

workers (Andersen and Mølhave, 1983). Additional investigation of preliminary indication of effects on pulmonary

function in children in the residential environment associated with lower concentrations of formaldehyde (0.048-0.072

mg/m

) (Krzyzanowski et al., 1990) is warranted.

Carcinogenic effects

On June 15, 2004, IARC announced there was sufficient evidence that formaldehyde causes nasopharyngeal cancer in

humans and re-classified it as a Group 1, known human carcinogen (previous classification: Group 2A). IARC also

reported there was limited evidence that formaldehyde exposure causes nasal cavity and paranasal cavity cancer and

“strong but not sufficient“ evidence linking formaldehyde exposure to leukemia.

EPA (1991) classified formaldehyde in Group B1 - probable human carcinogen, based on an evaluation of limited

human evidence and sufficient laboratory animal evidence. More recently, in a collaborative review and evaluation of

the formaldehyde epidemiology studies, EPA and CIIT (CIIT 1998) concluded, “It appears that a weak association

between nasopharyngeal cancer and formaldehyde exposure cannot be completely ruled out.” Adopting another view,

McLaughlin (1994) concluded, “Clearly, the causal criteria used by epidemiologists to evaluate an association, such as

strength of the association, consistency, specificity, dose-response, plausibility, and coherence, are not satisfied by the

epidemiologic studies in the formaldehyde-cancer research domain”. ECETOC (1995) similarly concluded, “After a

The INDEX project Final report

careful review of the cytologic, cytogenic and epidemiological studies there is an absence of evidence to support the

judgement of an etiologic relationship between formaldehyde and human cancer risk.”

Human carcinogenicity data

A series of papers on the cytogenetic effects of formaldehyde in humans has been published (IARC, 1995). The studies

mostly concern occupational exposures (plywood makers, makers of wood splinter materials, morticians, paper makers

and hospital autopsy service workers) and measure chromosomal anomalies, micronuclei in peripheral lymphocytes,

buccal epithelium (Norppa, et al., 1993) or nasal epithelium and sperm abnormalities. Both positive and negative

results were obtained, but their interpretation is difficult because of the small number of subjects studied,

inconsistencies in the findings, inadequate reporting of the data, and concomitant exposure to other chemicals. No

increase in urinary mutagenicity was observed in autopsy workers as compared to controls (Connor, et al. 1985). No

data are available on DNA–protein cross-links in humans. The overall conclusion is that adequate data on genetic

effects of formaldehyde in humans are not available (IARC, 1995).

The relationship between exposure to formaldehyde and cancer has been investigated in over 25 cohort studies of

professionals (pathologists, anatomists and embalmers) and industrial groups (formaldehyde producers, formaldehyde

resin makers, plywood and particle board manufacturers, garment workers and workers in the abrasives industry).

Relative risks have been estimated as standardized mortality ratios (SMRs), proportionate mortality ratios (PMRs),

proportionate cancer mortality ratios (PCMRs) and standardized incidence ratios (SIRs). In some studies, exposure was

not assessed but was assumed on the basis of the subject’s occupation or industry; in others, it was based on duration of

exposure and quantitative estimates of historical exposure levels. Mortality in several of the cohorts was followed

beyond the period covered by the original report; only the latest results are reviewed below, unless there were important

differences in the analyses performed or changes in the cohort definition (IARC, 1995,).

Recent meta-analyses by Blair et al. (Blair et al. 1990) and Partanen (Partanen, 1993) contain most of the available data.

The International Agency for Research on Cancer (IARC) summarized the results of the two meta-analyses in tabular

form. IARC (IARC, 1995) also systematically reviewed case-control studies of cancer of the oral cavity, pharynx and

respiratory tract. Also, McLaughlin (McLaughlin, 1994) recently published a critical review on formaldehyde and

cancer.

Excess numbers of nasopharyngeal cancers were associated with occupational exposure to formaldehyde in two of six

cohort studies of industrial or professional groups, in three of four case-control studies and in meta-analyses. In one

cohort study performed in 10 plants in the United States, the risk increased with category of increasing cumulative

exposure. In the cohort studies that found no excess risk, no deaths were observed from nasopharyngeal cancer. In three

of the case-control studies, the risk was highest in people in the highest category of exposure and among people

exposed 20–25 years before death. The meta-analyses found a significantly higher risk among people estimated to have

had substantial exposure than among those with low/medium or no exposure. The observed associations between

exposure to formaldehyde and risk of cancer cannot reasonably be attributed to other occupational agents, including

wood dust, or to tobacco smoking. Limitations of the studies include misclassification of exposure and disease and loss

to follow-up, but these would tend to diminish the estimated relative risks and dilute exposure–response gradient. Taken

together, the epidemiological studies suggest a causal relationship between exposure to formaldehyde and

nasopharyngeal cancer, although the conclusion is tempered by the small numbers of observed and expected cases in

the cohort studies (IARC, 1995).

Of six case-control studies in which the risk for cancer of the nasal cavities and paranasal sinuses in relation to

occupational exposure to formaldehyde was evaluated, three provided data on squamous cell tumours and three on

unspecified cell types. Of the three studies of squamous cell carcinomas, two (from Denmark and the Netherlands)

showed a positive association, after adjustment for exposure to wood dust, and one (from France) showed no

association. Of the three studies of unspecified cell types, one (from Connecticut, United States) gave weakly positive

results and two (also from the United States) reported no excess risk. The two case-control studies that considered

squamous cell tumours and gave positive results involved more exposed cases than the other case-control studies

combined. In the studies of occupational cohorts overall, however, fewer cases of cancer of the nasal cavities and

paranasal sinuses were observed than were expected. Because of the lack of consistency between the cohort and case-

control studies, the epidemiological studies can do no more than suggest a causal role of occupational exposure to

formaldehyde in squamous cell carcinoma of the nasal cavities and paranasal sinuses (IARC, 1995).

The cohort studies of embalmers, anatomists and other professionals who use formaldehyde tended to show low or no

risk for lymphatic or haematopoietic cancers, and excess risks for cancers of the brain, although they were based on

small numbers. These findings are countered by consistent lack of excess risk for brain cancer in the studies of the

industrial cohorts, which generally included more direct and quantitative estimates of exposure to formaldehyde than

did the cohort studies of embalmers and anatomists (IARC, 1995).

The INDEX project Final report

IARC (IARC, 1995) interpreted the above data as limited evidence for the carcinogenicity of formaldehyde in humans,

and classified formaldehyde as “probably carcinogenic to humans” (Group 2A).

In a published cohort study (Hansen & Olsen, 1995), a significantly increased relative risk of 3.0 (1.4–5.7; 95%

confidence limits) for sinonasal cancer was found among blue-collar workers potentially exposed to formaldehyde, but

with no probable exposure to wood dust, the major confounder.

Animal carcinogenicity data

Formaldehyde was tested for carcinogenicity by inhalation in mice, rats and hamsters, by oral administration in

drinking-water in rats, by skin application in mice, and by subcutaneous injection in rats. In additional studies in mice,

rats and hamsters, modification of the carcinogenicity of known carcinogens was tested by administration of

formaldehyde in drinking-water, by application to the skin or by inhalation (IARC, 1995).

Several carcinogenicity studies in which formaldehyde was administered to rats by inhalation produced evidence of

carcinogenicity, particularly induction of squamous cell carcinomas of the nasal epithelium, usually only at the highest

exposure level which ranged from about 18 to 25 mg/m

. In one long-term inhalation study, a high incidence of nasal

squamous cell carcinomas (15/58) was found in rats exposed to 12.3 mg/m

, the nasal mucosa of which had been

severely damaged by electrocoagulation before the exposure to formaldehyde started. In a comparable group of rats

exposed to the same formaldehyde concentration but with an undamaged nasal mucosa, no formaldehyde-related nasal

tumours were found (IARC, 1995). In one study in mice, no statistically significantly increase in the incidence of nasal

tumours was found after chronic (2-year) exposure to formaldehyde concentrations up to 17.6 mg/m

. Similar long-term

studies in hamsters showed no evidence of carcinogenicity (IARC, 1995).

In rats administered formaldehyde in the drinking-water, increased incidences were seen of forestomach papillomas in

one study and of leukaemias and gastrointestinal tract tumours in another; two other studies in which rats were treated

through the drinking-water gave negative results. Studies in which formaldehyde was applied to the skin or injected

subcutaneously were inadequate for evaluation (IARC, 1995).

In experiments to test the effect of formaldehyde on the carcinogenicity of known carcinogens, oral administration of

formaldehyde concomitantly with N-nitrosodimethylamine to mice increased the incidence of tumours at various sites;

skin application in addition to 7,12-dimethylbenz[a]anthracene reduced the latency of skin tumours. In rats, concomitant

administration of formaldehyde and N-methyl-N´-nitro-N-nitrosoguanidine in the drinking water increased the

incidence of adenocarcinoma of the glandular stomach. Exposure of hamsters by inhalation to formaldehyde increased

the multiplicity of tracheal tumours induced by subcutaneous injections of N-nitrosodiethylamine (IARC, 1995).

Genotoxic effects

Formaldehyde induces DNA–protein cross-links in mammalian cells in vitro and in vivo (2). The precise nature of these

cross-links is unknown. They are removed from normal cells with a half-time of 2–3 hours; the removal rates were

similar at nontoxic and toxic concentrations of formaldehyde (Grafström et al., 1984).

DNA–protein cross-links were measured in the mucosa of several regions of the respiratory tract of rats exposed by

inhalation to 0.4, 0.9, 2.4, 7.3 or 12.2 mg/m

C-formaldehyde, and of rhesus monkeys exposed by inhalation (mouth-

only) to 0.9, 2.4 or 7.3 mg/m

C-formaldehyde for 6 hours. The concentration of cross-links increased non-linearly

with the airborne concentration in both species. The concentrations of cross-links in the turbinates and anterior nasal

mucosa were significantly lower in monkeys than in rats. Cross-links were also formed in the nasopharynx and trachea

of monkeys, but they were not detected in the sinus, proximal lung or bone marrow. The investigators suggested that the

differences between rats and monkeys with respect to DNA–protein cross-link formation may be due to differences in

nasal cavity deposition and in the elimination of absorbed formaldehyde (Casanova et al., 1991; Heck et al., 1989).

In addition to DNA–protein cross-links, formaldehyde induced DNA single-strand breaks, chromosomal aberrations,

sister chromatid exchanges and gene mutations in human cells in vitro. It induced cell transformation, chromosomal

aberrations, sister chromatid exchanges, DNA strand breaks, DNA–protein cross-links and gene mutations in rodent

cells in vitro. Administration of formaldehyde in the diet to Drosophila melanogaster induced lethal and visible

mutations, deficiencies, duplications, inversions, translocations and crossing-over in spermatogonia. Formaldehyde

induced mutations, gene conversion, DNA strand breaks and DNA–protein cross-links in fungi, and mutations and

DNA damage in bacteria. Assays for dominant lethal mutations in rodents in vivo gave inconclusive results. In single

studies, formaldehyde induced sperm head abnormalities in rats but not in mice (IARC, 1995).

While there is conflicting evidence that formaldehyde can induce chromosomal anomalies in the bone marrow of

rodents exposed by inhalation in vivo, recent studies have shown that it induced cytogenetic damage in the cells of

The INDEX project Final report

tissues that are more locally exposed, either by gavage or by inhalation. Rats given 200 mg per kg body weight

formaldehyde orally were killed 16, 24 or 30 hours after treatment and examined for the induction of micronuclei and

nuclear anomalies in cells of the gastrointestinal epithelium. The frequency of mitotic figures was used as an index of

cell proliferation. Treated rats had significant (greater than five-fold) increases in the frequency of micronucleated cells

in the stomach, duodenum, ileum and colon; the stomach was the most sensitive, with a 20-fold increase in the

frequency of micronucleated cells 30 hours after treatment, and the colon the least sensitive. The frequency of nuclear

anomalies was also significantly increased at these sites. These effects were observed in conjunction with signs of

severe local irritation (Migliore et al., 1989).

In another experiment, rats were exposed to 0, 0.62, 3.7 or 18.5 mg/m

formaldehyde for 6 hours per day on five days

per week, for one and eight weeks. There was no significant increase in chromosomal abnormalities in the bone-marrow

cells of formaldehyde-exposed rats relative to controls, but there was a significant increase in the frequency of

chromosomal aberrations in pulmonary lavage cells (lung alveolar macrophages) from rats that inhaled 18.5 mg/m

formaldehyde. Aberrations, which were predominantly chromatid breaks, were seen in 7.6% and 9.2% of the scored

pulmonary lavage cells from treated animals, and in 3.5% and 4.8% of the cells from controls, after one and eight

weeks, respectively (Dallas et al., 1992).

The spectrum of mutations induced by formaldehyde has been studied in human lymphoblasts in vitro, in Escherichia

coli, and in naked pSV2gpt plasmid DNA (IARC, 1995; Crosby et al. 1988). About 50% of formaldehyde-induced

tumours of the nasal epithelium of rats appeared to have a point mutation in the p53 tumour suppressor gene (Recio et

al., 1992).

Overall, formaldehyde was genotoxic in a variety of experimental systems, ranging from bacteria to rodents, in vivo.

Formaldehyde given by inhalation or gavage to rats in vivo induced chromosomal anomalies in lung cells and

micronuclei, respectively, in the gastrointestinal tract (IARC, 1995).

Interactions with other chemicals

A group of 24 healthy nonsmokers were exposed while engaged in intermittent heavy exercise for 2 hours to

formaldehyde at 3.7 mg/m

or to a mixture of formaldehyde and 0.5 mg/m

of respirable carbon aerosol, in order to

determine whether adsorption of formaldehyde on respirable particles elicits a pulmonary response. Small (<5%)

decreases were seen in forced vital capacity and forced expiratory volume, but these effects were not considered to be

clinically significant (Green et al. 1989), Risby et al. (1990) and Rothenberg et al. (1989) estimated that the amount of

formaldehyde adsorbed on to carbon black or dust particles and delivered to the deep lung by particle inhalation is

minuscule in relation to the amount that remains in the vapour phase and is adsorbed in the upper respiratory tract.

The industrial manufacture of furniture, plywood and particle board may entail simultaneous exposure to formaldehyde

and wood dust, both being nasal carcinogens, the former in rats and the latter in humans (IARC, 1995). While the

epidemiological studies in woodworkers revealed the cancer excess to be attributable to wood dust per se rather than to

other exposures in the workplace such as formaldehyde, it is also true that exposure to wood dust is often an important

confounder in epidemiological studies on formaldehyde in industrial groups (IARC, 1995), indicating that exposure to

formaldehyde may enhance the nasal cancer risk associated with wood dust exposure.

Lam et al. (1985) studied the effects of inhalation co-exposure to acrolein and formaldehyde in male Fischer 344 rats.

Rats were exposed for 6 hours to room air (controls), 2 ppm acrolein, 6 ppm formaldehyde, or a combination of 2 ppm

acrolein and 6 ppm formaldehyde. The animals were sacrificed immediately after completion of exposure and their

nasal tissues were harvested. Exposure to formaldehyde significantly increased the percentage of interfacial DNA (a

measure of DNA-protein cross linking) compared to rats exposed to room air only (12.5 versus 8.1%, p<0.05). Co-

exposure to acrolein resulted in further increases in the percentage of interfacial DNA (18.6%) which were significantly

greater than the effect of formaldehyde alone (p<0.05). The authors concluded that simultaneous exposure to acrolein

enhanced formaldehyde-induced DNA-protein cross linking and that depletion of glutathione by acrolein inhibited the

metabolism of formaldehyde, thereby increasing formaldehyde-induced DNA-protein cross link formation.

To investigate the possibility of additive or potentiating interactions between inhaled aldehydes, Cassee et al. (1996)

compared responses in nasal epithelial histopathology and cell proliferation in groups of male Wistar rats exposed for 3

days (6 hours/day) to 1.0, 3.2, or 6.4 ppm formaldehyde alone; to 0.25, 0.67, or 1.40 ppm acrolein alone; to 750 or

1,500 ppm acetaldehyde alone; or to several mixtures of these aldehydes. At the concentrations tested, the histological

and cell proliferation responses measured in the nasal epithelium of rats exposed to the mixture which produced effects

(3.2 ppm formaldehyde; 1,500 ppm acetaldehyde; 0.67 ppm acrolein) were attributed by the investigators to the acrolein

alone with no additional effects from the formaldehyde or acetaldehyde. The investigators concluded that combined

The INDEX project Final report

exposures to these aldehydes at exposure levels in the vicinity of individual no-effect-levels was not associated with a

greater hazard than that associated with exposure to the individual chemicals.

Odour perception

Source: WHO (2001)

Formaldehyde has a pungent odour with odour detection thresholds as specified in Table 5.1 (0.03 – 0.6 mg/m

); the

odour recognition threshold is not known. Formaldehyde poses nuisance problems in indoor environments owing to its

release from building materials or furnishings. Indoor air usually contains other organic compounds which, in

combination with formaldehyde or by themselves, may have odorous and irritating properties causing discomfort. It has

been reported that some sensitive individuals can sense formaldehyde concentrations of 0.01 mg/m

and perhaps even

lower as a “warm” feeling on the face (Ahlström et al., 1986).

Source: Devos et al. (1990)

Odour threshold: 35 µg/m

Source: IPCS/WHO (1989)

The absolute odour threshold for formaldehyde is between 0.06 and 0.22 mg/m

(a group of observers detected the

odour in 50% of the presentations, 10% of an untrained population detected a level of 0.03 mg/m

). There is a low

probability that human beings will be able to detect formaldehyde in air at concentrations below 0.01 mg/m

Source: Moncrieff (1955)

The difference between odour and irritation concentration may be noticeable, but there is no evidence that there is a

threshold at which odour is superceded by irritation. However, for most inhaled odorous compounds, the trigeminal

nerve has a higher threshold than the olfactory nerve. When the formaldehyde concentration is increased and affects

both the eyes and the nostrils, sensory irritation is first experienced in the eyes, then the odour is perceived, and

finally nasal irritation occurs.

Summary of Formaldehyde Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Endogeneous formaldehyde levels in blood: 2-3 mg/litre

Toxicokinetics: Almost completely absorbed (>90%) in the upper respiratory tract; endogenous formaldehyde levels in

blood not altered after exposure; does not accumulate, rapid conversion to formate

Health effect levels of short- and long-term exposure (noncancer):

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

0.6

EXP

1.2

EXP

0.53

BC05

0.94

1h-ADJ-BC05

Mild and moderate eye irritation Volunteers, 3h Kulle et al. (1987) OEHHA 1999;

REL: 0.094

0.5

EXP

Nasal and eye irritation

Volunteers, 2h

(potentially sensitive

population)

Pazdrak et al. (1993) ATSDR 1999;

MRL: 0.05

0.24

STAT

Respiratory tract irritation cited in Sloof et al.

(1992)

RIVM 1992

0.1 Nose and throat irritation Human, short-term cited in IPCS/WHO,

1989

WHO 2000

(0.037) Respiratory symptoms (depending on study) Children; additional

investigation requested

Krzyzanowski et al.,

(1990); Wantke et al.

(1996); Garrett et al.

(1999); Franklin et al.

(2000)

UCLA 2001

Long-term exposure

0.09

Study

0.032

ADJ

0.26

Study

0.093

ADJ

Nasal and eye irritation, nasal obstruction,

and lower airway discomfort;

histopathological nasal lesions including

rhinitis, squamous metaplasia, and dysplasia

Occupational, 10y

average;

Wilhelmsson and

Holmstrom, 1992;

supported by Edling et

al., 1988

OEHHA 1999;

REL: 0.003

0.3

Study

Mild irritation of the eyes and upper Occupational, 10.4y Holmstrom et al., 1989 ATSDR 1999;

The INDEX project Final report

respiratory tract and mild damage to the

nasal epithelium

MRL: 0.01

0.31

STAT

Sensory irritation

for low but significant

percentage of exposed

workers

Weighting the total

body of data

NIWL 2003

0.12 Symptoms of irritation

LOAEL may be lower

only for a very small

proportion of the

population

well conducted studies Health Canada

1999; (noTC)

Carcinogenicity: IARC: 1 (15.06.04); U.S.EPA: B1; Unit risk (EPA-IRIS): 1.3E-5 (µg/m

)

-1

; An association exists

between cytotoxic, genotoxic, and carcinogenic effects. Most probably HCHO cannot act as a complete carcinogen at

non-cytotoxic concentrations. If exposure is accompanied by recurrent tissue damage at the site of contact, it may be

assumed to have carcinogenic potential in humans. Evidence of carcinogenicity in rats: induction of squamous cell

carcinoma at > 18 mg/m

In June 2004, IARC announced there was sufficient evidence that formaldehyde causes nasopharyngeal cancer in

humans and re-classified it as a Group 1, known human carcinogen (previously classified as Group 2A). IARC also

reported there was limited evidence that formaldehyde exposure causes nasal cavity and paranasal cavity cancer and

strong but not sufficient evidence linking formaldehyde exposure to leukemia.

Genotoxicity: Mutagenic in vitro, in vivo only in tissue of contact at doses adapt to give cytotoxic effects (no primary

mutagen)

Odour threshold: range: 0.03 - 0.6 mg/m

(WHO); 0.035 mg/m

(Devos); > 6 mg/m

untolerable to most people

Susceptible population: Wide variability in response to HCHO irritancy; Children seem to bee more sensitive than

adults (still additional investigation requested; results would be expected from recent German environmental survey

GerES IV); Asthmatics: no significant changes in pulmonary function and non-specific airway reactivity at 3.7 mg/m

;

population with allergic contact dermatitis (3-6%) seem not to react differently at 0.5 mg/m

(Pazdrak, 1993: 9

volunteers with formalin skin sensitization).

Remarks: RD

: 4 mg/m

; HCHO is a reaction product of ozone and unsaturated VOCs (e.g. limonene, pinene,

styrene); neither dose addition nor potentiating interactions occured upon exposure to aldehyde mixtures

(formaldehyde, acetaldehyde, acrolein) at exposure levels around the MOEL.

________________________________________________________________________________________________

The INDEX project Final report

6. Risk Characterization

Health hazard evaluation

Predominant symptoms of formaldehyde exposure in humans are irritation of the eyes, nose and throat, together with

concentration-dependent discomfort, lachrymation, sneezing, coughing, nausea and dyspnoea. Although there is

substantial variation in individual responses to formaldehyde in humans, there are sufficient data to indicate that the

irritant effects on the nose and throat after short-term exposure occur at concentrations as low as 0.1 mg/m

(Health

Canada; WHO), set as the LOAEL. A NOAEL of 0.03 mg/m

is defined according to the lowest concentration supposed

not to give effects in the normal population (OEHHA). This level has been further devided by an assessment factor of

30, taking into account the evidence indicating that children may be more sensitive to formaldehyde respiratory toxicity

than adults (3) and intraspecies variation (10) resulting in an Exposure Limit (EL) of 1 µg/m

. For this chemical the

observed acute effects and endpoints are consistent with the pathology seen in long-term studies. Consequently, no

distinction has been done between short-term and long-term ELs.

Percentage of population exposed beyond given threshold levels

Table 6.1

Population-based studies

LOAEL

NOAEL

Description (Study, Year) N 0.1 mg/m

0.03 mg/m

0.01 mg/m

0.001 mg/m

England (BRE) 528 < 35 % 90 % 100 %

England (BRE, 2002) 833 < 35 % 85 % 100 %

Vienna – bedrooms (Hutter, 2001) 160 < 40 % 90 % 100 %

Paris – refurbished flats (Clarisse, 03) 61 < 40 % 90 % 100 %

Helsinki (Expolis, 97) 15 < 65 % 100 % 100 %

French National Survey 7-d TWA (IAQ

observatory, 2003-04)

201 < 20 % 90 % 100 %

< out of the evaluation range (i.e. <5% of the environments investigated)

corresponds with 30-min average WHO guideline value

corresponds with lowest odour threshold reported

Cancer risk evaluation

On June 15, 2004, IARC announced there was sufficient evidence that formaldehyde causes nasopharyngeal cancer in

humans and re-classified it as a Group 1, known human carcinogen (previous classification: Group 2A). IARC also

reported there was limited evidence that formaldehyde exposure causes nasal cavity and paranasal cavity cancer and

“strong but not sufficient“ evidence linking formaldehyde exposure to leukemia.

Pending the outcome of the current IARC revision, the position taken in the present indoor risk evaluation is as follows:

Inhalation of formaldehyde under conditions that induce cytotoxicity and sustained regenerative proliferation within the

respiratory tract is considered to present a carcinogenic risk to humans. At airborne levels for which the prevalence of

sensory irritation is low (i.e., ≤0.1 mg/m

), risks of respiratory-tract cancers for the general population are exceedingly

low.

Comments

Among the available exposure data (see chapter 3), results from a limited number of indoor air surveys enabled the

construction of frequency distributions and hence the evaluation of population percentages exposed at given threshold

levels (see Table 6.1). When taking into account the whole body of published exposure data, the formaldehyde levels

reported in these surveys could be considered as indicative for “indoor air quality non-problem” buildings in the EU.

As summarised in Table 6.1, the results of these measurements indicate that at least more than 20% of the population is

exposed at formaldehyde levels higher than the NOAEL. A concentration slightly higher than the established NOAEL

has been recently proposed as a LOAEL (37 µg/m

) for effects on pulmonary functions in children, although additional

investigation is required to confirm this level. Concerning childrens’ susceptibility, important results are expected from

the German environmental survey GerES IV.

An exposure limit of 1 µg/m

has been derived, which corresponds more or less with the background outdoor level in

rural areas. It could be assessed that indoors almost the entire population in the EU is exposed at levels which are higher

than this limit.

The INDEX project Final report

For safety check purposes indoor formaldehyde levels are habitually refered to the WHO-air quality guideline value *

(100 µg/m

as a 30-minute average), with the assumption that transient sensory effects would be avoided for most

people. It is here asserted that this level should be regarded as hazardous for susceptible individuals (children) when

evidence exists that concentrations are maintained over prolonged periods.

* WHO-air quality guideline value:

“The lowest concentration that has been associated with nose and throat irritation in humans after short-term exposure is 0.1 mg/m

although some individuals can sense the presence of formaldehyde at lower concentrations. To prevent significant sensory irritation

in the general population, an air quality guideline value of 0.1 mg/m

as a 30-minute average is recommended. Since this is over one

order of magnitude lower than a presumed threshold for cytotoxic damage to the nasal mucosa, this guideline value represents an

exposure level at which there is a negligible risk of upper respiratory tract cancer in humans.“

Result

Because of its high chemical reactivity, formaldehyde is the most important sensory irritant among the chemicals

assessed in the present report. Due to its ubiquitousness in indoor environments and to the increasing evidence

indicating that children may be more sensitive to formaldehyde respiratory toxicity than adults, it is considered a

chemical of concern at levels exceeding 1 µg/m

, a concentration more or less corresponding with background levels in

rural areas. Results from available exposure data, although limited, confirm that almost the entire population is exposed

indoors at levels (Median level±sd: 26±6 µg/m

; 90th±sd: 59±7 µg/m

; N = 6) higher than this background level, here

established as the limit of exposure, with more than 20% of the European population exposed at levels exceeding the

no-observed-effect-level (NOAEL: 30 µg/m

). Within the reported concentration range mild irritation of the eyes could

be experienced by the general population as well as the odour perceived starting from about 30 µg/m

Reported formaldehyde concentrations are lower (99th < 150 µg/m

) than a presumed threshold for cytotoxic damage to

the nasal mucosa and hence considered low enough to avoid any significant risk of upper respiratory tract cancer in

humans. The last statement could be subjected to changes due to the current IARC revision of the carcinogenicity of

formaldehyde.

The INDEX project Final report

Carbon monoxide

Synonyms: Carbon oxide, Carbonic oxide

CAS Registry Number: 630-08-0

Molecular Formula: CO

1. Compound identification

Carbon monoxide is a colourless, practically odourless and tasteless gas that is poorly soluble in water, but it is soluble

in alcohol and benzene. It is a product of incomplete combustion of carbon-containing fuels. Carbon monoxide burns

with a violet flame and it is classified as an inorganic compound. It has a slightly lower density than air. Anthropogenic

emissions are responsible about two-thirds of the CO in the atmosphere and natural emissions account for the remaining

one-third Small amounts of carbon monoxide are also produced endogenously (EPA 2000, Alm 1999).

Accidental high exposures to CO can lead to acute health effects that can be fatal. Carbon monoxide exposure causes

unintentional and suicidal poisonings, and a large number of deaths annually both in Europe and in the United States. It

is estimated that more than half of the 6000 annual deaths from fires in the Unites States is caused by CO poisoning

(Raub et al 2000). It is obvious that such homes exist where CO concentrations are high enough to increase chronic

health effects, especially among sensitive populations such as pregnant women, the fetus, children, the elderly, and

individuals suffering from anemia or other diseases that restrict oxygen transport between blood and cells (IEH 1998).

In the human body, CO reacts with haemoglobin to form carboxyhaemoglobin (HbCO). The relationship of CO

exposure and the HbCO concentration in blood is presented in Table 1.1.

The scientific literature on carbon monoxide sources, concentrations, and human exposures and health effects in

outdoor and indoor environments has been comprehensively reviewed by the US Environmental Protection Agency

(EPA) and the health risk assessment has been carried out by WHO (EPA 2000, WHO 1979 and 2000). Raub et al

(2000) have also recently reviewed literature about carbon monoxide poisoning from a public health perspective.

Table 1.1. The relationship between carbon monoxide exposure and carboxyhaemoglobin levels in blood (adopted from

WHO 1979)

% HbCO

mg/m

ppm % in air

0.87 5.7 5 0.0005

1.73 11.5 10 0.001

3.45 23 20 0.002

5.05 34.5 30 0.003

6.63 46 40 0.004

8.16 57.5 50 0.005

9.63 69 60 0.006

11.08 80 70 0.007

12.46 92 80 0.008

13.8 103 90 0.009

15.11 114.5 100 0.01

16.37 126 110 0.011

17.6 130 120 0.012

18.78 149 130 0.013

19.95 160 140 0.014

21.05 172 150 0.015

22.15 183 160 0.016

23.23 195 170 0.017

24.26 206 180 0.018

25.25 218 190 0.019

26.22 229 200 0.02

The INDEX project Final report

2. Physical and Chemical properties

Molecular weight (g/mol) 28.01

Melting point (oC ) -205.1

Boiling point (oC) -191.5

Density, at 0°C, 1 atm (kg/m

) 1.250

at 25°C, 1 atm (kg/m

) 1.145

Relative density (air =1) 0.967

Solubility in water (at 0°C, 1 atm, ml/100 ml) 3.54

(at 25°C, 1 atm, ml/100 ml) 2.14

(at 37°C, 1 atm, ml/100 ml) 1.83

Conversion factors at 20 °C and 760 mm Hg:

1 ppm = 1.164 mg/m

1 mg/m

= 0.859 ppm

Sources: WHO 1979, 2000a, Verschueren 2001

3. Indoor Air Exposure assessment

Indoor air and exposure concentrations

Residential indoor concentrations of CO were typically lower in Northern Europe than in Central Europe, where they

were again lower than in Southern Europe based on the European exposure study, EXPOLIS (Georgoulis et al 2002).

Average residential indoor concentration in Helsinki was 1.2 mg/m

for non-ETS population. Average 48-hour

exposure to CO, being 1.4 mg/m

, was slightly higher than the respective indoor concentration. In Basle and Prague

average exposures to CO were higher than in Helsinki, but lower than in Milan and Athens. The highest geometric

mean exposure concentrations were found in a subpopulation of smokers in Athens, being 4.0 mg/m

(Georgoulis et al

2002). Average residential indoor CO concentrations in Milan vary from 2.1 to 3.9 mg/m

. In another Italian study

carried out by Maroni et al (1996) an average personal exposure of 2.2 mg/m

was reported for office workers. Weekly

average CO concentrations in 14 homes in UK ranged from 0.2 to 2.7 mg/m

in a study published by Ross (1996). Alm

et al (2000) reported mean exposures of 15-min, 1-hour and 8-hour periods for preschool children in Finland, being 8.4

mg/m

, 6.0 mg/m

and 3.3 mg/m

, respectively.

In UK typical ambient background CO levels range between 0.01 and 0.23 mg/m

, but in traffic environments, the 8-

hour mean concentrations are much higher, but typically less than 20 mg/m

(IEH 1998).

The cumulative distributions of the indoor, 48-hour and 1-hour personal exposure concentrations of carbon monoxide in

European studies are presented in Figure 3.1, Figure 3.2, Figure 3.3 and Figure 3.4.

The INDEX project Final report

100

012345

Indoor concentration (mg/m

)

Cumulative frequency (%)

Hel

Mil

Fre

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of carbon monoxide in Helsinki (Hel, n=

258785) for non-ETS population (Scotto di Marco et al 2003), in the French Survey (Fre n= 550666) (Kirchner 2004)

and in Milan (Mil, n= 4822) (Bruinen de Bruin et al 2004) for the whole population (sampling periods were 1-min, 15-

min, and 5-min in Helsinki, Milan and France, respectively).

100

0.00.51.01.52.0

Indoor concentration (mg/m

)

Cumulative frequency (%)

CO 14-day

Figure 3.2. Cumulative frequency distribution of 14-day indoor air concentrations of carbon monoxide in the kitchens in

England (GM= 0.47 mg/m3, max 4.45 mg/m3, n=830, Raw et al 2002).

The INDEX project Final report

100

0123456

Exposure (mg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Pra

Figure 3.3. Cumulative frequency distributions of 48-hour exposures to carbon monoxide in Athens (Ath), Basel (Bas),

and Prague (Pra) (Georgoulis et al 2002), Helsinki (Hel) (EXPOLIS 2002), and Milan (Mil) (Bruinen de Bruin et al

2004).

100

0 1020304050

Exposure (mg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.4. Cumulative frequency distributions of 1-hour exposure concentrations of carbon monoxide in Athens (Ath,

n=44), Basel (Bas, n=50), Helsinki (Hel, n=195), Milan (Mil, n=45) Oxford (Oxf, n=26) and Prague (Pra, n=23)

(Hanninen et al 2004).

Emission sources

The most common cause of high carboxyhaemoglobin concentrations in man is the smoking of tobacco and the

inhalation of the products by the smoker. Faulty domestic cooking and heating appliances, inadequately vented to

outside air, may cause high indoor concentrations of CO (WHO 1979). Also gas stoves, water heaters, and exhaust from

vehicles in attached garages might be important indoor sources (Alm 1999).

The most important source of carbon monoxide in ambient air is the exhaust of gasoline-powered motor vehicles. The

emission rate depends on the type of vehicle, its speed, and its mode of operation. Other common ambient sources

include heat and power generators, especially when using coal, industrial processes such as the carbonisation of fuel,

and the incineration of refuse (WHO 2000).

The INDEX project Final report

Based on the EXPOLIS results average residential indoor CO concentrations in Milan were the lowest when no special

indoor sources were present and the highest if gas cooking and environmental tobacco smoke (ETS) were present. The

highest short-time peak concentration was found during gas cooking, 7.4 mg/m

. Average 15-min ambient CO

concentration, 2.6 mg/m

was higher than the respective indoor concentration when no indoor sources were present, but

lower if gas cooking or ETS was present. The highest in-transit concentration was measured in cars/taxis being 6.5

mg/m

(Bruinen de Bruin et al 2003). Average 1-hour exposures to CO were higher, 8.4 mg/m

, than the respective

ambient concentrations, 5.7 mg/m

. Instead, average 8-hour and 48-hour ambient concentrations were at the same level

than the respective exposures, being 3.8 and 2.4 mg/m

, respectively.

Short time carbon monoxide concentrations related to some typical indoor sources such as tobacco smoke, gas cooking

and commuting in five European cities are presented in Figure 3.5. Much higher concentrations were found in a Finnish

study determining personal exposures of preschool children in Helsinki. The highest exposures to carbon monoxide

were as high as 80 mg/m

, 69 mg/m

and 28 mg/m

for 15-min, 1-hour and 8-hour averages, respectively. Elevated

exposures were related to gas stoves, mothers’ smoking and living in high rise buildings (Alm et al 2000).

Brown et al (2002) studied emission rates and concentrations caused by unflued gas heaters in a room chamber with a

volume of 32.4 m

. In normal conditions carbon monoxide concentrations ranged from <1.15 mg/m

to 4.6 mg/m

. But

when the gas supply was restricted due to the backpressure of 0.02 kPa, concentration of 19.5 mg/m

was measured.

Even higher concentration, 34.4 mg/m

, was determined when the researchers caused a slight misalignment of the

burner. Because fatal concentrations may occur during malfunction of burning appliances, estimated concentrations

caused by burning different fuels in a stove are presented in Table 3.1. Despite of the fact that the assumptions are valid

for houses common in developing countries, these estimates may help to assess the concentrations that may be caused

by burning these fuels in stoves also in developed countries.

Typical sources and microenvironments where high concentrations of carbon monoxide may occur are reviewed in

Table 3.2 based on the recently published studies. The highest maximum values ranging 121 – 182 mg/m

were

measured in homes when using a gas grill attached to the gas stove (IEH 1998), in the underground parking facilities (El

Fadel et al 2001), and in a home with a faulty boiler (Ross 1996). Elevated concentrations were also found in other

microenvironments such as motor vehicles, indoor ice arenas, bars and restaurants. Carbon monoxide concentrations in

public buildings studied before 1985 are summarised in Table 3.3.

Lately attention is paid to incense burning in homes and other public buildings including stores and shopping malls.

Jetter et al (2002) reported emission rates of 23 different types of incense such as incense rope, cones, sticks, rocks,

powder etc. that are typically used indoors. The measured emission rates of carbon monoxide ranged 144 - 531 mg/h.

The authors estimated a peak concentration of 9.6 mg/m

caused by incense burning and, therefore concluded that

carbon monoxide concentrations could exceed the US EPA’s National Ambient Air Quality Standard (NAAQS) 10

mg/m

for an 8-hour average depending on the room volume, ventilation rate and the amount of incense burned. Lee

and Wang (2004) reported similar results when studying emissions of incense burning in chamber (18 m

) tests. They

measured maximum carbon monoxide concentrations up to 44 mg/m

during burning and concluded that incense

burning is an important indoor air pollution source in addition to carbon monoxide also to fine particles and VOCs.

Especially, incense burning might be a significant contributor to population exposure in such cultures, where incense is

burned frequently, for example in religious rituals.

The INDEX project Final report

ETS non-ETS Gas

cooking

No gas

cooking

Car Bus-

tram

Metro-

train

Concentration (mg/m

)

Ath

Bas

Hel

Mil

Pra

Figure 3.5. Short time (geometric means of the 15-min periods) exposures to carbon monoxide stratified by

environmental tobacco smoke (ETS) and gas cooking, and in-vehicle concentrations in European cities Athens (Ath),

Basel (Bas), Helsinki (Hel), Milan (Mil) and Prague (Pra) (Georgoulis et al 2002).

Table 3.1. Estimated indoor CO concentrations resulting from burning 1 h of fuel in stove without flues using

assumptions relevant in developing countries i.e. fuel consumption equal to the energy used for burning 1.7 kg of fuel

wood, ventilation rate of 15 h-1, and room volume of 40 m3 (adopted from Zhang et al 1999).

Fuel

Avgerage during burning maximum

Charbiquette 562 603

Charcoal 528 566

Brush wood 511 548

Coal 178 191

Fuel wood 150 161

Kerosene 9.4 10

LPG 4.5 4.8

Biogas 1.6 1.7

Natural gas 0.083 0.089

CO concentration (mg/m

)

The INDEX project Final report

Table 3.2. Short time CO concentrations related to specific microenvironments or emission sources.

Environment/

emission source

Location Averaging time Reference

MGMmax

Gas stove with pilot light UK peak when using a grill 10-182 IEH 1998

Underground parking Beirut, Lebanon 30-min 80 26 - 140 El Fadel et al 2001

facilities

Home UK 1-min faulty boiler 121 Ross 1996

gas cooking 6 - 49

all-electric homes 3.5 - 4

Undergrounds, car parks, several hours >115 60 - 115 WHO 2000

enclosed ice arenas, etc.

homes with gas appliances

Gas appliances The Netherlands max 1-min kitchen 5-108 Lebret et al 1987

max 1-hour kitchen 3-56

Exposure Helsinki, Finland max 15-min pre-school 80 Alm et al 2000

max 1-hour children 69

max 8-hour 28

Incense burning peak chamber tests 44 Lee and Wang 2004

Bars, restaurants Italy 12 - 23 35 Maroni et al 1995

Indoor ice arenas Finland max 1-hour 20-33 Pennanen et al 1997

Exposure Milan, Italy max 15-min 12 11 25 Bruinen de Bruin 2002

max 1-hour 8.4 8.4 20

max 8-hour 3.8 3.8 9.3

Unflued gas heaters room chamber test 1-min normal cond. <1.2 - 4.6 Brown et al 2002

gas supply restricted 20

misaligned burner 34

Shops and bars Genoa, Italy 8-hour shops 15 Valerio et al 1997

bars 18

Indoor ice skating rink Hong Kong, China 15-min gasoline-fueled 8-16 Guo et al 2004

15-min propane-fueled 5.7

Exposure Helsinki, Finland max 15-min ETS 9.0 5.7 Scotto di Marco et al 2003

non-ETS 7.1 5.3

max 1-hour ETS 5.7 3.7

non-ETS 4.3 3.3

max 8-hour ETS 2.6 2.1

non-ETS 2.0 1.8

Concert hall Switzerland event (5.5 hours) 3.5 5.2 Junker et al 2000

Gas appliances NY, USA RTI 1990

3-day gas stove 2.6

kerosene heater 3.9

Public buildings Athens, Greece 1-hour Office 3.7 Chaloulakou et al 2003

AM = arithmetic mean, GM = geometric mean, max = maximum value

Concentration (mg/m

)

The INDEX project Final report

Table 3.3. Indoor air concentrations of carbon monoxide in selected indoor microenvironments (adopted from Akland et

al 1985).

Indoor environment

n ppm std

mg/m

Public garage 116 13.46 18.14 15.41

Service station/motor vechile repair facility 125 9.17 9.33 10.50

Repair shop 55 5.64 7.67 6.46

Shopping mall 58 4.90 6.50 5.61

Residential garage 66 4.35 7.06 4.98

Restaurant 524 3.71 4.35 4.25

Office 2287 3.59 4.18 4.11

Sport arena, concert hall 100 3.37 4.76 3.86

Store 734 3.23 5.56 3.70

Health care facility 351 2.22 4.25 2.54

Other public buildings 115 2.15 3.26 2.46

Residence 21543 2.04 4.06 2.34

School 426 1.64 2.76 1.88

Church 179 1.56 3.35 1.79

Mean concentration

The INDEX project Final report

4. Toxicokinetics

Endogenous sources of carbon monoxide:

In addition to inhaled carbon monoxide, humans are also exposed to small amounts of carbon monoxide produced

endogenously, conducing in the healthy population to a baseline carboxyhaemoglobin concentration range of 0.4-0.7%.

Sources of carbon monoxide in the body include the endogeneous oxidation of halomethanes (e.g. dichloromethane),

phenols, flavonoids, as well as the lipid peroxidation of membrane lipids (Rodgers et al., 1994).

Carbon monoxide is a natural degradation product of haemoproteins, mainly haemoglobin* but also myoglobin,

cytochromes, peroxidases and catalase. Approximately 0.4 ml/h CO is formed by haemoglobin catabolism, and about

0.1 ml/h originates from non-haemoglobin sources (Coburn et al., 1964). In both males and females, week-to-week

variations in carbon monoxide production are greater than day-to-day or within-day variations. In females,

carboxyhaemoglobin levels fluctuate with the menstrual cycle; the mean rate of carbon monoxide production in the

premenstrual, progesterone phase almost doubles (Delivoria-Papadopoulos et al., 1970; Lynch & Moede, 1972).

Neonates and pregnants also showed a significant increase in endogenous carbon monoxide production related to

increased breakdown of red blood cells.

Any disturbance leading to increased destruction of red blood cells and accelerated breakdown of other haemoproteins

would lead to increased production of carbon monoxide. Haematomas, intravascular haemolysis of red blood cells,

blood transfusion and ineffective erythropoiesis will all elevate the carbon monoxide concentration in the blood.

Degradation of red blood cells under pathological conditions such as anaemias (haemolytic, sideroblastic, sickle cell),

thalassaemia, Gilbert’s syndrome with haemolysis and other haematological diseases will also accelerate carbon

monoxide production (Berk et al., 1974; Solanki et al., 1988). In patients with haemolytic anaemia, the carbon

monoxide production rate was 2–8 times higher and blood carboxyhaemoglobin concentration was 2–3 times higher

than in normals (Coburn et al., 1966). As a result, baseline carboxyhaemoglobin levels can be as high as 4%. In subjects

with haemoglobin Zürich, where affinity for carbon monoxide is 65 times that of normal haemoglobin,

carboxyhaemoglobin levels range from 4% to 7%.

Increased carbon monoxide production rates have been reported after administration of phenobarbital, progesterone

(Delivoria-Papadopoulos et al., 1970) and diphenylhydantoin (Coburn, 1970).

* In the process of natural degradation of haemoglobin to bile pigments, in concert with the microsomal reduced nicotinamide

adenine dinucleotide phosphate (NADPH) cytochrome P-450 reductase, two haem oxygenase isoenzymes, HO-1 and HO-2, catalyse

the oxidative breakdown of the alpha-methene bridge of the tetrapyrrol ring of haem, leading to the formation of biliverdin and

carbon monoxide. The production of haem oxygenase isoenzyme HO-1 is increased by any kind of stress, including heat, light,

sound, odors, infection, physical trauma and mental or psychological stress. Chronic stress in any of these pathways thus results in

chronic destruction of haem and chronic low-level CO poisoning. Stress-induced HO-1 activity and the relatively constant activity of

isozyme HO-2, that does not respond to stress, together account for about 75% of the human body's CO production.

Absorption

After reaching the lungs, inhaled carbon monoxide diffuses rapidly across the alveolar and capillary membranes. It also

readily crosses the placental membranes. Carbon monoxide binds reversibly to one of the haem proteins. Approximately

80–90% of the absorbed carbon monoxide binds with haemoglobin, which causes a reduction in the oxygen-carrying

capacity of the blood. The affinity of haemoglobin for carbon monoxide is 200–250 times that for oxygen, while the

relative affinities of other haem proteins (e.g. myoglobin), cytochrome oxidase and cytochrome P-450 for carbon

monoxide are much lower (U.S.EPA, 1991; ACGIH, 1991ab).

When in equilibrium with ambient air, the carboxyhaemoglobin (COHb) content of the blood will depend mainly on the

concentrations of inspired carbon monoxide and oxygen. If equilibrium has not been achieved, the COHb concentration

will also depend on the duration of exposure, pulmonary ventilation and the COHb originally present before inhalation

of the contaminated air.

The absorption and elimination of carbon monoxide have been described in various mathematical models (U.S.EPA,

1991; WHO, 1979). The most important of these models is the Coburn-Foster-Kane exponential equation, which takes

into account all known physiological variables affecting carbon monoxide uptake (Coburn et al., 1965). The most

important variables determining the COHb level are: baseline COHb level, blood haemoglobin level, minute ventilation,

diffusion capacity of the lungs, and mean oxygen tension in pulmonary capillaries. During an exposure to a fixed

concentration of carbon monoxide, the COHb concentration increases rapidly at the onset of exposure, starts to level off

The INDEX project Final report

after 3 hours, and reaches a steady state after 6–8 hours (see Figure 4.1). At steady state, the carbon monoxide

concentrations in alveolar breath and ambient air become practically equal (Peterson, 1970).

Figure 4.1: Log-log plot of carbon monoxide uptake by humans from

low ambient CO concentrations as computed from the Coburn-Forster-Kane equation.

Abbreviations: Pco2 = mean partial pressure of O2 in lung capillaries, VA = alveolar ventilation rate, Vb = blood volume, M = equilibrium constant,

DL = diffusing capacity of lungs, [COHb]0 = value prior to CO exposure, Vco = rate of endogenous CO production. (From: Peterson and Stewart,

1975).

Using an adaptation of the Coburn equation the average 1 and 8 hr ambient concentration required to achieve a COHb

level of 2.5% (the target COHb concentration served as a basis for the U.S. air quality standard for CO) for children

with different baseline levels of COHb was calculated. As shown in Figure 4.2, as baseline COHb concentration

increased, the amount of inhaled CO required to raise the blood level to 2.5% was decreased.

The INDEX project Final report

Figure 4.2: Increasing baseline COHb will reduce the time-weighted average

CO concentration required to reach 2.5% COHb after a given exposure. (from OEHHA, 2000)

Elimination

Carbon monoxide is eliminated unchanged via the lungs. The decline in COHb concentration depends on the rate of

carbon monoxide release from haem proteins, alveolar ventilation, oxygen concentration in inhaled air, duration of

carbon monoxide exposure, and the level of COHb saturation. The formation of COHb is a reversible process, but

because of the tight binding of carbon monoxide to haemoglobin, the elimination half-life while breathing room air is

2–6.5 hours depending on the initial COHb level. The elimination half-life of COHb is much longer in the fetus than in

the pregnant mother (U.S.EPA, 1991; ACGIH, 1991ab).

5. Health effects

Effects of short-term exposure

CO affects health by interfering with the systemic transport of oxygen to tissues (especially the heart and other muscles

and brain tissue) (Costa and Amdur, 1996). The resulting impairment of O

delivery cause tissue hypoxia and interferes

with cellular respiration. Direct intracellular uptake of CO could permit interactions with haemoproteins such as

myoglobin, cytochrome oxidase and cytochrome P-450, and therefore interfere with electron transport processes and

energy production at the cellular level (Brown and Piantidosi, 1992). Thus, in addition to observed physiological effects

and cardiovascular effects, CO can modify electron transport in nerve cells resulting in behavioral, neurological and

developmental toxicological consequences, and may itself play a role in neurotransmission.

The health effects associated with inhaled CO vary with its concentration and duration of exposure. Effects range from

subtle cardiovascular and neurobehavioral effects at low concentrations to unconsciousness and death after prolonged

exposures or after acute exposures to high concentrations of CO. First signs and symptoms on healthy individuals, such

as decreases in work capacity and decrements of neurobehavioral functions start at [COHb] of 5% (WHO, 1999;

U.S.EPA, 2000), whereas first signs of CO poisoning appear at [COHb] concentrations of 10% (see Table 5.1).

However, the variability within the human population must be considered high. A [COHb] of about 15 % only leads to

slight symptoms, such as headache, in healthy adults (Stewart et al., 1970; WHO, 1999). In contrast, the same [COHb]

can cause long-lasting defects in the cognitive development and behavioral alterations in children (Klees et al., 1985),

or even contribute to death from myocardial infarction in individuals with coronary artery disease (Grace and Platt,

1981; Balraj, 1984).

The INDEX project Final report

Table 5.1: Carboxyhaemoglobin levels resulting from steady-state exposure to increasing concentrations of

CO in ambient air and associated symptoms in healthy adult humans and susceptible (adapted from

U.S.EPA, 2001; Winter and Miller, 1976; Ellenhorn and Barceloux, 1988)

[CO] in atmosphere [COHb] Signs and symptoms

ppm mg/m

% Healthy adults Susceptible subpopulations

0 0 0.4 – 0.7 Physiologic background concentration

10 11.5 2 Asymptomatic

17 19.5 2.9

during physical exertion reduced

time to onset of angina and

electrocardiogram signs of

myocardial ischaemia in

sunjects with coronary artery

disease

5 - 6

Decreases in work capacity and

decrements of neurobehavioral function

Increase in cardiac arrythmias in

subjects with coronary artery

disease

42 48 7 Headache, nausea in children

- - 3 - 8 Background concentration in smokers

70 80 10

No appreciable effect, except shortness

of breath on vigorous exertion; possible

tightness across the forehead; dilation of

cutaneous blood vessels.

Cognitive development deficits

in children

Myocardial infarction in subjects

with coronary artery disease

120 137 20

Shortness of breath on moderate

exertion; occasional headache with

throbbing in temples

25 syncopes in children - stillbirths

220 252 30

Decided headache; irritable; easily

fatigued; judgment disturbed; possible

dizziness; dimness of vision

350 - 520 401 – 595 40 – 50

Headache, confusion; collapse; fainting

on exertion

800 - 1220 916 - 1400 60 – 70

Unconsciousness; intermittent

convulsion; respiratory failure, death if

exposure is long continued

1950 2230 80 Rapidly fatal

At CO levels typically encountered in indoor and outdoor environments, health effects are most likely to occur in

individuals who are physiologically stressed, either by exercise or by medical conditions that can make them more

susceptible to low levels of CO. Subpopulations at increased risk of adverse effects are :

1. Individuals with cardiovascular diseases: COHb levels of 2-6% may impair the delivery of oxygen to the

myocardium causing hypoxia and increasing coronary blood flow demand by nearly 30%. When myocardial

oxygen demands are increased, as in exercise, the hypoxic effects of CO may exceed the limited coronary

reserve producing adverse health effects including earlier onset of myocardial ischaemia, reduced exercise

tolerance in persons with stable angina pectoris, increased number and complexity of arrhythmias, and

increased hospital admissions for congestive heart failure.

2. Fetuses are more susceptible to CO exposure for several reasons: CO crosses the placenta; fetal Hb has greater

affinity for CO than maternal Hb; the half-life of COHb in fetal blood is three times longer than that of

maternal blood, and the fetus has high rate of oxygen consumption and lower oxygen tension in the blood than

adults. Also, maternal smoking during pregnancy exposes the fetus to greater than normal concentrations of

CO leading to a decrease in birth weight.

3. Children develop acute neurotoxic effects (e.g. headaches, nausea), long-lasting neurotoxic effects (e.g.

memory deficits) and impaired ability to escape (i.e. syncopes) at lower [COHb] than adults. Children have

greater activity levels and smaller body masses than adults and should therefore experience higher levels of

CO uptake than will adults for the same average exposure concentration.

4. Pregnant women have increased alveolar ventilation, increasing the rate of CO uptake from inspired air. Also,

a pregnant woman produces nearly twice as much endogenous CO.

The INDEX project Final report

5. Individuals with chronic obstructive pulmonary disease such as chronic bronchitis, emphysema and chronic

obstructive pulmonary disease are more susceptible to CO effects, since their lungs are less efficient at

oxygenating the blood.

6. Individuals with reduced blood haemoglobin concentrations, or with abnormal haemoglobin, will have reduced

carrying capacity in blood. In addition, disease processes that result in increased destruction of red blood

cells (haemolysis) and accelerated breakdown of haemoproteins accelerate endogenous production of CO,

resulting in higher COHb concentrations than in normal individuals. For example, patients with haemolytic

anemia have COHb concentrations 2 to 3 times those seen in normal individuals.

7. Certain occupation groups are at risk from ambient CO exposure including those who work on city streets

(street repairmen, street cleaners, street vendors, deliverymen, and garage attendants, taxi and bus drivers).

Individuals who work in industrial processes including those exposed to other chemical substances (e.g.

methylene chloride) that increase endogenous CO formation.

8. Individuals who have not adapted to high altitude and are exposed to a combination of high altitude and CO.

Cardiovascular effects

Numerous controlled human studies have been conducted in healthy subjects and in patients with ischaemic heart

disease in order to characterize the effects of low-level carbon monoxide exposures on the cardiorespiratory responses

to exercise. In these experiments, the subjects have typically been exposed to clean air and carbon monoxide in a

chamber or through a face mask. After the exposure, which has usually been conducted at rest to achieve a

predetermined COHb concentration in the blood, the subjects have engaged in an exercise test on a treadmill or cycle

ergometer until exhaustion (healthy subjects) or the appearance of angina pectoris or electrocardiographic signs of

cardiac ischaemia.

In apparently healthy subjects, the maximal exercise time and the maximal oxygen consumption have decreased at

COHb levels as low as 5%. The regression between the percentage decrease in maximal oxygen consumption and the

percentage increase in COHb concentration appears to be linear, with approximately a one percentage point fall in

oxygen consumption per one percentage point rise in COHb level above 4% (U.S.EPA, 1991; Bascom et al, 1996).

Patients with cardiovascular disease, especially ischaemic heart disease, are expected to be particularly sensitive to

carbon monoxide. Atherosclerotic narrowing of the coronary arteries and impaired dilatation mechanisms restrict blood

flow to the myocardium and prevent physiological compensation for lowered oxygen delivery caused by elevated levels

of COHb. In exercise, these subjects experience myocardial ischaemia, which can impair myocardial contractility,

affect cardiac rate and rhythm, and cause angina pectoris (U.S.EPA, 1991; Bascom et al, 1996).

The early studies of Aronow et al. (1972), Aronow & Isbell (1973) and Anderson et al. (1973) have suggested that low-

level carbon monoxide exposures resulting in COHb levels of 2.5–3.0% shorten the time to onset of exercise-induced

chest pain in patients with angina pectoris. Although the validity of the studies of Aronow and colleagues has been

questioned by the US Environmental Protection Agency, subsequent studies by other investigators have actually given

similar results (U.S.EPA, 1991; Bascom et al, 1996).

The design and results of the five most important clinical studies (Kleinman et al., 1989; Allred et al.,1989; Sheps et al.,

1987; Anderson et al., 1973; Adams et al., 1988) conducted in patients with ischaemic heart disease are summarized in

Table 5.2. Despite the obvious differences between these studies, they all show a significant shortening in the time to

onset of angina at mean post-exposure COHb levels of 2.9–5.9% (post-exercise COHb levels in Table 5.2 are somewhat

lower), which represent mean incremental increases of 1.5–4.4% COHb from the pre-exposure baseline levels

(U.S.EPA, 1991).

The potential arrhythmogenic effects associated with low-level carbon monoxide exposures have not been fully

resolved at COHb levels of ≤ 5% (U.S.EPA, 1991; ACGIH, 1991b). Hinderliter et al. (1989) reported no effects at 3.5%

and 4.9% COHb levels (post-exercise concentrations) on resting and exercise-induced arrhythmias in ten patients with

coronary artery disease and no baseline ectopia. In contrast, Sheps et al. (1990) showed in 41 nonsmoking patients with

documented coronary artery disease and various levels of baseline ectopia that the frequencies of both single and

multiple ventricular depolarizations increased significantly at a mean post-exercise COHb level of 5.0% but not at

3.5%. Dahms et al. (1993) found no additional effect of either 3% or 5% COHb over the exercise-induced increases in

single or multiple ectopic beats experienced by patients with myocardial ischaemia and baseline ectopia.

The INDEX project Final report

Table 5.2: A summary of results of the five most important double-blind clinical studies on the effects of low-level

carbon monoxide exposures on patients with documented ischaemic heart disease and exercise-induced angina

Exposure Reference

(mg/m

)

COHb

(%)

Exposure duration

and activity

Subject

characteristics

Effects of CO exposure (symptoms,

ECG changes, etc.)

Anderson et

al., 1973

115

1.3

2.9

4.5

4-hour exposure at

rest, post-exposure

exercise on a

treadmill

10 males, mean age

49.9 years (5 smokers, 5

nonsmokers), repro-

ducible angina

Time to onset of angina shortened at

COHb 2.9% and 4.5% (P < 0.005) and

duration of angina prolonged at

COHb 4.5% (P < 0.01).

Deeper ST-segment depressions with CO

in 5 subjects.

Kleinman et

al., 1989

115

1.4

2.8

1-hour exposure at

rest, post-exposure

incre-mental exercise

on a cycle ergometer

24 males, mean age

58.8 years nonsmokers

for at least 6 months),

reproducible angina

Time to onset of angina shortened by

5.9% (P = 0.046), no significant effect on

duration of angina, oxygen uptake

at angina reduced by 2.2% (P = 0.04).

Time to 0.1 mV ST-segment depression

shortened by 19.1% (P = 0.044) in 8

subjects.

Allred et al.,

1989

134

290

0.6

2.0

3.9

50- to 70-minute

exposure at rest, pre-

and post-exposure

incremental exercise

on a treadmill

63 males, mean age 62

years (nonsmokers),

reproducible angina

Time to onset of angina shortened by

4.2% (P = 0.054) at COHb 2.0% and by

7.1% (P = 0.004) at COHb 3.9%.

Time to threshold ischaemic ST-segment

changes shortened by 5.1% (P = 0.02) at

COHb 2.0% and by 12.1% (P < 0.0001) at

COHb 3.9%.

Significant dose relationships in the

changes of both the onset of angina (P =

0.02) and the onset of ST-segment

changes (P < 0.0001).

Sheps et al.,

1987

115

1.5

3.6

1-hour exposure at

rest, post-exposure

incremental exercise

on a cycle ergometer

25 males and 5 females,

mean age 58.2 years

(nonsmokers for at least

2 months), ischaemia in

a screening test

No significant changes in time to onset of

angina, duration of angina, maximal

exercise time, maximal ST-segment

depression, time to significant ST-

segment depression, or maximal left

ventricular ejection fraction.

3 subjects experienced angina only on CO

exposure, actuarial analysis including

these subjects showed shortening in time

to onset of angina in the study group (P <

0.05).

Carbon monoxide, 1 mg/m

= 0.873 ppm.

Carboxyhaemoglobin concentrations are from venous blood samples taken immediately after exercise; in the study of Anderson et al.

(40) samples were taken only immediately after carbon monoxide exposure.

According to some epidemiological and clinical data, carbon monoxide from recent smoking and environmental or

occupational exposures may contribute to cardiovascular mortality and the early course of myocardial infarction

(U.S.EPA, 1991). It is not known whether this contribution is due to arrhythmogenic effects or to some longer-term

effects, as suggested by some authors. In patients with severe ischaemic heart disease, carbon monoxide poisonings

have been lethal at COHb levels of 10–30%, while usual COHb levels in lethal poisonings are around 50–60%

(U.S.EPA, 1991).

A number of recent epidemiological studies reported associations between levels of ambient air pollutants (CO, PM,

O3, NOx, SO2) and hospital admissions for cardiovascular diseases (Atkinson et al., 1999; Burnett et al., 1997; Morris

et al., 1995; Linn et al., 2000; Le Tertre et al., 2002; Mann et al., 2002; Yang et al., 2004). In all the studies cited a

positive association was found between CO ambient concentrations and the daily number of cardiovascular disease

hospitalizations. at the local level (Litovitz et al., 1992; Mathieu et al., 1996).

Often, individuals suffering from CO poisoning are unaware of their exposure because symptoms are similar to those

associated with viral illness or clinical depression (Raub et al., 2000). This may result in a significant number of

misdiagnoses by medical professionals (Grace and Platt, 1981; Fisher and Rubin, 1982; Barret et al., 1985; Dolan et al.,

1987; Kirkpatrick, 1987; Heckerling et al., 1987, 1988, 1990). Although the precise number of individuals who suffer

from CO poisoning is not known, it is certainly much larger than that indicated by mortality figures. Schaplowsky et al.

(1974) estimated that more than 10 000 people per year in the United States required medical attention or missed at

least 1 day of work in the early 1970s because of sublethal exposures to CO. A recent study (Hampson, 1998) estimated

over 40 000 emergency department visits annually for recognized acute CO poisoning in the United States.

The INDEX project Final report

Developmental effects

The pregnant mother, the fetus in utero and the newborn infant are at high risk of adverse health effects from

atmospheric carbon monoxide exposures. During pregnancy, the endogenous production of carbon monoxide can be

elevated as much as 3-fold, the concentration of maternal haemoglobin is often reduced, and the mothers have

physiological hyperventilation. As a result of these changes, maternal COHb levels are usually about 20% higher than

the nonpregnant values. Carbon monoxide diffuses readily across the placental membranes, and the carbon-monoxide-

binding affinity of fetal haemoglobin is higher than that of adult haemoglobin. Moreover, carbon monoxide is cleared

much more slowly from fetal blood than from maternal blood. At steady state, fetal COHb levels are up to 10–15%

higher than maternal COHb levels (U.S.EPA, 1991; Longo, 1977).

There are theoretical reasons and supporting laboratory animal data to suggest that the fetus and the developing organs

are especially vulnerable to carbon monoxide. The developing brain seems to have the highest sensitivity of all organs.

There is a well established and probably causal relationship between maternal smoking and low birth weight at fetal

COHb levels of 2–10%. In addition, maternal smoking seems to be associated with a dose-dependent increase in

perinatal deaths and with behavioural effects in infants and young children. Carbon monoxide is probably one of the

most important etiological factors for these effects, although there are numerous other toxic pollutants in tobacco smoke

(Longo, 1977).

A case-control study of the association between low birthweight infants and maternal CO exposures in approximately

1000 cases in Denver (Alderman et al., 1987) failed to detect a relationship between CO exposure (estimated form

fixed-site outdoor monitoring data) during the last 3 months of pregnancy and lower birth weights. Mean CO levels

ranged from 0.6 to 4.1 mg/m

(0.5 to 3.6 ppm) at 8 monitoring locations in metropolitan Denver. The 5th and 95th

percentile concentrations at the site with the highest (4.1 mg/m

) mean were 1.8 and 5.5 mg/m

(1.6 and 4.8 ppm),

respectively. The odds ratio at the highest concentration site was 1.1 and the 95% confidence interval was 0.8-1.6. This

study did not directly account for unmeasured sources of CO exposure, such as smoking, emissions from gas appliances

and exposures to vehicular exhaust, which are limitations of the study design.

A more extensive study of a cohort of 125573 children born to women living in the Los Angeles area (1989-1993)

found that exposure to ambient concentrations > 6.3 mg/m

(3 mo average) during the last trimester of pregnancy was

associated with a significantly increased risk of low birthweight (odds ratio = 1.22; confidence interval =1.03-1.44)

after adjustment for potential confounders (Ritz and Yu, 1999). Fetotoxicity has been demonstrated in laboratory animal

studies. Altered brain neurochemical development and growth retardation have been demonstrated in rats exposed to

CO in utero (Storm and Fechter, 1985; Leichter, 1993).

Neurological and neurobehavioural effects

Central nervous system (CNS) effects in individuals suffering acute CO poisoning cover a wide range, depending on

severity of exposure: headache, dizziness, weakness, nausea, vomiting, disorientation, confusion, collapse, and coma.

At low concentrations, CNS effects include reduction in visual perception, manual dexterity, learning, driving

performance, and attention level. Earlier work is frequently cited to justify the statement that CO exposure sufficient to

produce COHb levels of ca. 5% would be sufficient to produce visual sensitivity reduction and various neurobehavioral

performance deficits. In a recent literature re-evaluation, however, the best estimate was that [COHb] would have to rise

to 15–20% before a 10% reduction in any behavioral or visual measurement could be observed (Raub and Benignus,

2002). This conclusion was based on (1) critical review of the literature on behavioral and sensory effects, (2) review

and interpretation of the physiological effects of COHb on the CNS, (3) extrapolation from the effects of hypoxic

hypoxia to the effects of CO hypoxia, and (4) extrapolation from rat behavioral effects of CO to humans.

Summary of short-term exposure effect levels

NOAEL

% COHb

LOAEL

% COHb

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

1.1-1.3 2 aggravation of angina and other cardiovascular

diseases

Subjects with angina

pectoris, 1h

Aronow, 1981

OEHHA 1999

REL: 23

2.5

- acute ischaemic heart attacks (in documented

or latent coronary artery disease)

- untoward hypoxic effects in fetuses of

nonsmoking pregnants

- Subjects with

ischaemic heart

disease

- Fetuses

WHO 2000

100 (15m)

60 (30m)

30 (1h)

10 (8h)

2.7 aggravation of angina pectoris, and other

symptoms of myocardial ischaemia

Subjects with angina

pectoris

Aronow and Isbell,

1973; Aronow et al.,

1972; Anderson et al.,

1973

U.S.EPA 1994

10.3 (8h)

35 (1h)

2.9

cardiorespiratory effects in individuals with

ischaemic heart disease.

Subjects with

ischaemic heart

disease performing

Kleinman et al., 1998;

Allred et al.,1989;

Sheps et al., 1987;

Health Canada

2003

MDL(1h/8h)=15/6

The INDEX project Final report

light work Kleinman et al., 1989;

Adams et al., 1988

MAL(1h/8h)=35/1

MTC(8h)=20

Health Canada - Recommended National Ambient Air Quality Objectives - MDL(1h/8h): Maximum desirable level (1and 8 hour time

weighted average); MAL: Maximum acceptable level; MTC: Maximum tolerable level

as measured by CO-Oximeter -

as measured by gas chromatography

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Because angina is a severe effect, there is no level protective against mild adverse effects.

Derivation of Acute Reference Exposure Level (

protective against severe adverse effects, for a 1-hour

exposure)

Study Aronow, 1981

Study population humans

Exposure method inhalation

Critical effects aggravation of angina and other cardiovascular diseases

LOAEL 2% carboxyhaemoglobin in blood

NOAEL 1.1%-1.3% carboxyhaemoglobin in blood (corresponds to 20 ppm

CO,

calculated toxicokinetically)

Exposure duration 1 hour

LOAEL uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 1

Cumulative uncertainty factor 1

Reference Exposure Level 20 ppm (23 mg/m³, 23,000 mg/m³)

WHO Guidelines (2001)

In controlled human studies involving patients with documented coronary artery disease, mean postexposure COHb

levels of 2.9–5.9% (corresponding to postexercise COHb levels of 2.0–5.2%) have been associated with a significant

shortening in the time to onset of angina, with increased electrocardiographic changes and with impaired left ventricular

function during exercise (Anderson et al., 1973; Kleinman et al., 1989; Allred et al., 1989; Sheps et al., 1987; Adams et

al., 1988). In addition, ventricular arrhythmias may be increased significantly at the higher range of mean postexercise

COHb levels (Sheps et al., 1990; Stern et al., 1988). Epidemiological and clinical data indicate that carbon monoxide

from recent smoking and environmental or occupational exposures may contribute to cardiovascular mortality and the

early course of myocardial infarction (U.S.EPA, 1991). According to one study there has been a 35% excess risk of

death from arteriosclerotic heart disease among smoking and nonsmoking tunnel officers, in whom the long-term mean

COHb levels were generally less than 5% (Stern et al., 1988). Current data from epidemiological studies and

experimental animal studies indicate that common environmental exposures to carbon monoxide do not have

atherogenic effects on humans (U.S.EPA, 1991; Smith and Steichen, 1993).

During pregnancy, endogenous production of carbon monoxide is increased so that maternal COHb levels are usually

about 20% higher than the non-pregnant values. At steady state, fetal COHb levels are up to 10–15% higher than

maternal COHb levels U.S.EPA, 1991; Longo, 1977). There is a well established and probably causal relationship

between maternal smoking and low birth weight at fetal COHb levels of 2–10%. In addition, maternal smoking seems

to be associated with a dose-dependent increase in perinatal deaths and with behavioural effects in infants and young

children (Longo, 1977).

In order to protect nonsmoking, middle-aged and elderly population groups with documented or latent coronary artery

disease from acute ischaemic heart attacks, and to protect the fetuses of nonsmoking pregnant women from untoward

hypoxic effects, a COHb level of 2.5% should not be exceeded. The following guidelines are based on the Coburn-

Foster-Kane exponential equation, which takes into account all the known physiological variables affecting carbon

monoxide uptake (Coburn et al., 1965). The following guideline values (ppm values rounded) and periods of time-

weighted average exposures have been determined in such a way that the COHb level of 2.5% is not exceeded, even

when a normal subject engages in light or moderate exercise:

The INDEX project Final report

100 mg/m

(90 ppm) for 15 minutes

60 mg/m

(50 ppm) for 30 minutes

30 mg/m

(25 ppm) for 1 hour

10 mg/m

(10 ppm) for 8 hours.

U.S.EPA (1994)

The National Ambient Air Quality Standards (NAAQS) for CO were promulgated by the Environmental Protection

Agency (EPA) in 1971 at levels of 9 ppm (10 mg/m3) for an 8 h average and 35 ppm (40 mg/m3) for a 1 h average, not

to be exceeded more than once per year. (Primary and secondary standards were established at identical levels). The

1970 CO criteria document (National Air Pollution Control Administration, 1970) cited as the standard's scientific basis

a study which indicated that subjects exposed to low levels of CO, resulting in COHb concentrations of 2 to 3% of

saturation exhibited neurobehavioral effects (Beard and Wertheim, 1967). A revised CO criteria document (U.S.

Environmental Protection Agency, 1979) concluded that it was unlikely that significant, and repeatable,

neurobehavioral effects occurred at COHb concentrations below 5%. However, reports that aggravation of angina

pectoris, and other symptoms of myocardial ischaemia, occurred in men with chronic cardiovascular disease, exposed to

low levels of CO resulting in COHb concentrations of about 2.7% (Aronow and Isbell, 1973; Aronow et al., 1972;

Anderson et al., 1973), lead EPA to retain the 8 h 9 ppm primary standard level and to reduce the 1 h primary standard

from 35 to 25 ppm. (EPA also revoked the secondary CO standards because no adverse welfare effects had been

reported at nearambient levels). Later, concerns regarding the validity of data on which the proposed reduction in the 1

h standard was based caused EPA to decide to retain the 1 h 35 ppm standard.

Health Canada (2003)

The choice of National Ambient Air Quality Objectives for carbon monoxide is based upon the clinical significance of

health effects of concern, the number of people requiring protection, and the relationship between COHb levels,

exposure, and ambient carbon monoxide levels.

COHb levels are a biomarker for the toxicity of ambient-level exposures to carbon monoxide and are used as an

indicator of carbon monoxide exposure. Although more research is needed to evaluate the predictive capabilities of the

CFK model in individuals exposed to low concentrations of carbon monoxide and its applicability to sensitive

subpopulation (U.S. EPA, 1991), it is the best model available at present, and it will be used here to estimate

appropriate National Ambient Air Quality Objectives for carbon monoxide. However, it must be remembered that

models provide estimates based upon small numbers of representative measurements.

Table 5.3: Recommended National Ambient Air Quality Objectives, ppm (mg/m

)

Averaging

Times

Maximum Desirable

Level

Maximum Acceptable

Level

Maximum Tolerable

Level

1 hour 13 (15) 30 (35) n/a

8 hoursb 5 (b) 13 (15) 17.4 (20)

a 1 ppm = 1.146 mg CO/m

b rolling average

The maximum desirable levels are based on the carbon monoxide concentration that will result in a

carboxyhaemoglobin (COHb) blood level of less than 1%, or the upper end of the range of baseline COHb levels

resulting from endogenous production. Based on the Coburn-Foster-Kane (CFK) equation, a 1-hour exposure of 13 ppm

or an 8-hour exposure of 5 ppm would lead to less than 1 % COHB.

The recommended maximum acceptable levels for carbon monoxide are 30 ppm averaged over 1-hour and 13 ppm as

an 8-hour rolling average. Results from five recent studies in 3 laboratories were consistent in finding adverse effects of

COHb levels ranging from 2.9% to 6% (as measured by CO-Oximeter) or as low as 2% (as measured by gas

chromatography) on exercise induced angina and on ECG (electrocardiogram) values. CO levels averaging 13 ppm over

8 hours or 30 ppm over 1 hour resulted in COHb levels at or below 2% for adults performing light work (ventilation

rate of 18L/m).

Therefore, the maximum acceptable levels are based on the maintenance of COHb levels of less than 2%, thereby

providing a small margin of safety. At levels above these concentrations, action would be required to decrease the

probability or severity of effects in sensitive populations. Estimating COHb levels using pNEM indicates that less than

1% of the Toronto study area population will experience COHb levels greater than 2.0% if the ambient air quality is less

than or equal to 13 ppm (15 mg/m3) measured over 8 hours.

The INDEX project Final report

The recommended maximum tolerable level for an 8 hour exposure could be based on the lowest observed adverse

effect level (LOAEL) of 2.9% COHb observed in the same experiments cited previously. This level would be

approximately 21 ppm. However, the current tolerable level of 17 ppm is sufficiently close to this value to be retained

as the objective.

The recommended maximum tolerable level of 17 ppm averaged over 8 hours will result in a COHb level of about 2.5%

as projected by the CFK model. This is still below the COHb levels believed to result in cardiorespiratory effects in the

general population. Moreover, it is considered to be slightly more protective in accounting for non-standard conditions

and people at the high end of the distribution curve for parameters used in the CFK equation. However, owing to a

diminishing margin of safety within the tolerable range, action is recommended without delay when air quality exceeds

the highest concentration of this range to protect the health of sensitive subgroups.

The averaging times chosen for the maximum desirable, acceptable, and tolerable ranges of carbon monoxide in

ambient air are 1 and 8 hours. The latter averaging time approximates the length of time during which people may be

exposed to carbon monoxide continuously in a particular location (e.g., work, sleep). More importantly, most

individuals approach equilibrium levels of COHb in the blood after about 8–12 hours of exposure to (Anderson et al.,

1973). Owing to the possibility of missing some events (high levels of carbon monoxide exposure) using a continuous

averaging time, rolling averages are recommended for the calculation of the 8-hour averages. The 1-hour averaging

period is intended to be protective for effects that might occur following short exposures to high concentrations of

carbon monoxide.

Effects of long-term exposure

There is not enough reliable information on effects of chronic exposures to low concentrations from either controlled

human studies, ambient population-exposure studies, or from occupational studies. For example, current models cannot

confidently predict whether reduction in pollution will decrease monthly rates of hospital admissions or mortality, even

if they imply a reduction of admissions on days with low pollution. This public-health-related issue cannot be addressed

by daily time series analysis, using only admission or mortality counts. In future studies, investigators also could

consider time-averaged health effects over, say, 1 or 3 months, in relation to pollution exposure metrics for the

corresponding periods (Raub and Benignus, 2002).

Chronic exposures to low CO concentrations may not pose as much a problem as high, acute exposure due to

physiological compensation, tolerance, or adaptation. Smokers show an adaptive response to elevated COHb levels, as

evidenced by increased red blood cell volumes (through increased haemopoiesis) or reduced plasma volumes (Raub and

Benignus, 2002). The major source of total exposure to CO for smokers comes from active tobacco smoking. Baseline

COHb concentrations in smokers average 4% (compared to <2% for non-smokers), with a usual range of 3–8%for one

to two pack-per-day smokers, reflecting absorption of CO from inhaled smoke. COHb levels as high as 15% have been

reported for chain smokers.

The only experimental evidence for short- or long-term compensation to increased COHb levels in the blood from

exogenous sources other than smoking is indirect (Raub and Benignus, 2002). Experimental animal data indicate that

incremental increases in COHb produce physiological responses that tend to offset the deleterious effects of CO

exposure on oxygen delivery to the tissues.

Experimental human data indicate that compensatory cardiovascular responses to submaximal upper and lowerbody

exercise (e.g. increased heart rate, cardiac contractility, cardiac output) occur after CO exposures. These changes were

highly significant for exposures attaining 20% COHb. Other compensatory responses are increased coronary blood

flow, cerebral blood flow, Hb, and oxygen consumption in muscle.

Cardiovascular effects

Kristensen (1989) examined the relationship between cardiovascular diseases and chronic occupational exposures, and

concluded that CO exposure increases the acute risk of cardiovascular disease, at least transiently. Stern et al. (1988)

investigated the effects of occupational carbon monoxide exposures on deaths from arteriosclerotic heart disease among

5529 New York City bridge and tunnel officers in the period 1952–1981. Among the more exposed tunnel officers there

was a 35% excess risk compared with the New York City population, whereas among the less exposed bridge officers

the risk was not elevated. The elevated risk among the tunnel officers declined significantly within five years after

cessation of the occupational exposure, and there has also been a significant decline since 1970, when the introduction

of new ventilation systems lowered the carbon monoxide levels in tunnels and tunnel booths. The 24-hour average

carbon monoxide concentrations inside the tunnels were around 57 mg/m

(50 ppm) in 1961 and 46 mg/m

(40 ppm) in

1968. During rush hour traffic in 1968, carbon monoxide concentrations in tunnel toll booths were as high as 74–189

mg/m

(65–165 ppm) and in 1970 the mean concentration over 38 days was 72 mg/m

(63 ppm). However, the mean

The INDEX project Final report

COHb levels measured among smoking and nonsmoking tunnel officers in 1970 and 1981 were generally lower than 5

Hansen (1989) reported the results of a 10-year follow-up study on mortality among 583 Danish male automobile

mechanics between 15 and 74 years of age. The number of deaths expected for the automobile mechanics was

compared with those for a similar group of Danish men employed as carpenters, electricians and other skilled workers

free from occupational exposure to automobile exhaust, petrochemical products, asbestos and paint pigments. The

number of deaths observed among the automobile mechanics exceeded the expected number by 21%. Although the

increased mortality was not confined to any single cause of death, the author reported a remarkable excess of deaths

attributed to ischaemic heart disease where the standardized mortality ratio was 121 and the 95% confidence interval

was 102–145. The only other significant category of death was that due to external causes (SMR = 131; 95% CI = 113–

153). No significant differences were found among the automobile mechanics for other diseases except for an increase

in pancreatic cancer (SMR = 219; 95% CI = 128–351). Exposure to carbon monoxide and polycyclic aromatic

hydrocarbons through the inhalation of automobile exhaust and the handling of solvents and oils may have accounted

for the difference in ischaemic heart disease deaths between the automobile mechanics and the comparison group;

however, other occupational exposures or other lifestyle factors, as indicated above, may also have contributed to the

findings.

The haemodynamic responses to CO have been reviewed by Penney (1988). Chronic CO exposures, at levels sufficient

to raise COHb concentrations to greater than 10% can produce increased numbers of red blood cells (polycythemia),

increased blood volume, and increased heart size (cardiomegaly). In addition, heart rate, stroke volume, and systolic

blood pressure may be increased. Some of these effects have been seen in smokers. Penney and Howley (1991) report

that CO can enhance atherosclerosis in individuals with elevated serum cholesterol.

Current data from epidemiological studies and laboratory animal studies do not suggest that common environmental

exposures to carbon monoxide have atherogenic effects on humans (U.S.EPA, 1991; Smith and Steichen, 1993).

Effects on pregnancy outcomes

A case-control study of the association between low birthweight infants and maternal CO exposures in approximately

1000 cases in Denver (Alderman et al., 1987) failed to detect a relationship between CO exposure (estimated form

fixed-site outdoor monitoring data) during the last 3 months of pregnancy and lower birth weights. Mean CO levels

ranged from 0.5 to 3.6 ppm at 8 monitoring locations in metropolitan Denver. The 5th and 95th percentile

concentrations at the site with the highest (3.6 ppm) mean were 1.6 and 4.8 ppm, respectively. The odds ratio at the

highest concentration site was 1.1 and the 95% confidence interval was 0.8-1.6). This study did not directly account for

unmeasured sources of CO exposure, such as smoking, emissions from gas appliances and exposures to vehicular

exhaust, which are limitations of the study design. A more extensive study of a cohort of 125,573 children born to

women living in the Los Angeles area (1989-1993) found that exposure to ambient concentrations > 5.5 ppm (3 mo

average) during the last trimester of pregnancy was associated with a significantly increased risk of low birthweight

(odds ratio = 1.22; confidence interval =1.03-1.44) after adjustment for potential confounders (Ritz and Yu, 1999).

Fetotoxicity has been demonstrated in laboratory animal studies. Altered brain neurochemical development and growth

retardation have been demonstrated in rats exposed to CO in utero (Storm and Fechter, 1985; Leichter, 1993).

Effects on lung function

Individuals exposed to relatively high concentrations of CO in both indoor and outdoor environments may have lung

function decreases. In most cases, however, causality is difficult to establish because, in addition to CO, these

individuals were also exposed to high concentrations of other combustion products, many of which are respiratory

system irritants.

In the study of tunnel and bridge officers, described earlier, lung functions, forced vital capacity (FVC) and forced

expiration volume in 1 s (FEV1.0), were slightly reduced in tunnel vs. bridge officers (Evans et al., 1988). Exposures of

adults to typical ambient concentrations of CO, both outdoors and indoors, have not been significantly associated with

lung function decrements (Lebowitz et al., 1987).

Pollutants related to automotive traffic, especially CO and nitrogen oxides, were associated with the prevalence of

asthma in middleschool Taiwanese students (Guo et al., 1999; Lin et al., 1999). Physician consultations in London for

lower respiratory diseases were significantly correlated with NO2 and CO levels in children, but not in adults (Hajat et

al., 1999). Exposure of children with mild asthma to environmental tobacco smoke resulted in pulmonary function

decrements, i.e. reduced FEV1.0 (Magnusen et al., 1993).

The INDEX project Final report

Carcinogenic and genotoxic effects

No studies documenting carcinogenic or genotoxic effects of CO in humans were located in the available literature.

Summary of Carbon Monoxide Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Endogenous production of CO due to the catabolism of haemoproteins and

membrane lipids results in 0.4-0.7% COHb (healthy subjects), increased to 0.7-2.5% during pregnancy. Increased

endogenous production of CO are associated with a series of haematological diseases (e.g.haemolytica anemia) and to

the endogenous oxidation of specific pollutants (e.g. halomethanes).

Toxicokinetics: ~80-90% of absorbed dose binds with haemoglobin (COHb; causing a reduction in the oxygen-

carrying capacity of the blood). CO affinity for haemoglobin 200-250 times that for O

. CO is eliminated unchanged via

the lungs, with t

(COHb) of 2-6.5 hours (while breathing clean air), depending on the initial COHb level. t

(COHb)

much higher in the fetus than in the pregnant mother.

Health effect levels of short- and long-term exposure

NOAEL

% COHb

LOAEL

% COHb

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

SHORT-TERM EXPOSURE

1.1-1.3 2 aggravation of angina and other cardiovascular

diseases

Subjects with angina

pectoris, 1h

Aronow, 1981

OEHHA 1999

REL: 23

2.5

- acute ischaemic heart attacks (in documented

or latent coronary artery disease)

- untoward hypoxic effects in fetuses of

nonsmoking pregnants

- Subjects with

ischaemic heart

disease

- Fetuses

WHO 2000

100 (15m-GV)

60 (30m-GV)

30 (1h-GV)

10 (8h-GV)

2.7 aggravation of angina pectoris, and other

symptoms of myocardial ischaemia

Subjects with angina

pectoris

Aronow and Isbell,

1973; Aronow et al.,

1972; Anderson et al.,

1973

U.S.EPA 1994

10.3 (8h)

35 (1h)

2.9 cardiorespiratory effects in individuals with

ischaemic heart disease.

Subjects with

ischaemic heart

disease

Kleinman et al., 1998;

Allred et al.,1989;

Sheps et al., 1987;

Kleinman et al., 1989;

Adams et al., 1988

Health Canada

2003

MDL(1h/8h)=15/6

MAL(1h/8h)=35/1

MTC(8h)=20

LONG-TERM EXPOSURE (not found in available literature)

Carcinogenicity: No evidence

Genotoxicity: No evidence

Odour threshold: practically odorless

Susceptible population: Human characteristics increasing risk:

- Individuals with Ischaemic heart disease or coronary artery disease

- Individuals having reduced blood haemoglobin concentrations (e.g.haemolytic anemia) or with abnormal

haemoglobin

- Individuals with chronic bronchitis, emphysema and chronic obstructive pulmonary disease (COPD), children with

other lung inflammatory problems (e.g.cystic fibrosis)

- Pregnants and fetuses

- Children: physiologically, children have larger metabolic demands and consequently greater oxygen uptake demands

than do adults, on a per unit mass basis.

________________________________________________________________________________________________

The INDEX project Final report

6. Risk Characterization

Health hazard evaluation

Carbon monoxide (CO) is a tasteless, non-irritating, odorless and colorless toxic gas which can cause acute and lethal

poisonings with very few and late occurring warning signs. Until very severe symptoms occur (inability to walk) none

or only nonspecific symptoms are noted in healthy humans. As a consequence, a large number of deaths occur annually

in fires, workplaces and in indoor environments. In the latter instance, poisonings occur because of the incorrect use of

combustion devices or due to faulty unvented gas appliances. The associated risk, the evaluation of its impact on the

European population as well as the related risk management strategies will be outlined in the following chapter (Risk

Management).

The present risk characterization adresses CO exposures experienced by people during their normal daily activities,

while changing from one microenvironment to another, the significance of related health effects, as a function of

individual susceptibility, and the relationship between carboxyhaemoglobin levels (COHb; the biomarker for CO-

exposure and -toxicity), exposure, and indoor carbon monoxide levels.

Currently available evidence suggests that individuals with heart disease (including stable exercise-induced angina,

coronary artery disease, and ischaemic heart disease) represent the population at greatest risk from exposure to indoor

carbon monoxide levels. In addition, population groups with either increased probability or increased severity of health

effects include fetuses, pregnant women, and young infants, individuals with anemia or respiratory disease, the elderly,

children, and persons with peripheral vascular disease and chronic obstructive lung disease.

Results from five clinical studies in 3 laboratories were consistent in finding adverse effects of COHb levels as low as

2.9% (as measured by CO-Oximeter) on exercise induced angina and on electrocardiogram values.

Providing a small margin of safety, a COHb level of 2 % will be here considered as the starting point of the present risk

characterization, and CO exposures considered acceptable

based on the maintenance of COHb levels of less than 2%.

These levels are considered protective for the subpopulation sensitive to cardiovascular effects and are expected to

provide an adequate margin of safety for the other susceptible subpopulations described above and for the general

population with respect to short- and long-term toxicity effects. Additionally, CO exposures are here considered

desirable

when they result in COHb levels of less than 1%, representing the upper end of the range of baseline COHb

levels resulting from endogenous production in normal, non-smoking individuals. This level was chosen taking into

consideration uncertainties in the available data and using the most conservative assumptions.

The CO concentrations in inhaled air associated with these threshold COHb levels (see Table 6.1) were estimated using

the Coburn-Foster-Kane (CFK) model, considering adults performing light work at a ventilation rate of 18L/min. The

CFK model still presents the best model available, although more research is needed to evaluate the predictive

capabilities in individuals exposed to low CO concentrations and its applicability to sensitive subpopulation. The

averaging times chosen for deriving the CO air concentrations are 1 and 8 hours, the latter approximating the length of

time required by most individuals to approach equilibrium levels of COHb in the blood (about 8–12 hours of exposure),

or the time during which people may be continuously exposed to carbon monoxide in a particular location. The 1-hour

averaging period is intended to evidence short-term exposures related to personal activities generating increased CO

levels.

Table 6.1: Carbon monoxide average concentrations in inhaled air derived (CFK-model) from

selected [COHb], considering 1-h and 8-h exposure duration

COHb levels considered [COHb] % [CO] in air mg/m

1 h–avg. 8h–avg.

acceptable ≤ 2 ≤ 35 ≤ 15

desirable ≤ 1 ≤ 15 ≤ 6

The INDEX project Final report

Percentage of population exposed beyond given threshold CO levels

Table 6.1

acceptable desirable

Available exposure studies 1-h averaging period

Description (Study, Year) N 35 mg/m

15 mg/m

1-h personal exposure (indoors)

Athens (Expolis, 96-98) 44 10 % 30 %

Basel (Expolis, 96-98) 50 < < 10 %

Helsinki (Expolis, 96-98) 195 < 5 %

Milan (Expolis, 96-98) 45 < < 10 %

Oxford (Expolis, 98-00) 26 < 5 %

Prague (Expolis, 96-98) 23 < 10 %

8-h averaging period

15 mg/m

6 mg/m

48-h personal through the day

exposure

Athens (Expolis, 96-98) 44 < <

Basel (Expolis, 96-98) 50 < <

Helsinki (Expolis, 96-98) 195 < <

Milan (Expolis, 96-98) 45 < <

Prague (Expolis, 96-98) 26 < <

2-week residential exposure

(kitchen and bedroom)

England (BRE, 97-99) 830 < <

< out of the evaluation range (i.e. <5% of the environments

investigated)

The studies

Expolis study: Within the multi-centre European EXPOLIS study, personal exposure to CO, measured every minute for 48 h, of 401

randomly selected study participants (25–55 yr old; living and working within the metropolitan areas; mainly non-smokers) was

monitored in Athens, Basel, Helsinki, Milan and Prague. 1-h average personal exposures were computed in order to assess the

influence of specifc sources. Each participant also completed a time-microenvironment-activity diary and an extended questionnaire.

In addition, for the same time period, ambient levels of CO from fixed site stations were collected. All measurements were performed

between October 1996 and June 1998.

In Georgoulis et al. (2002) both the 48-h and 15-min averaged personal exposures were calculated and compared, the latter

corresponding to specific 15-min time periods when different activities (i.e. ‘‘under smoke (ETS) exposure’’, ‘‘during use of gas

appliances’’) were taking place.

Selection bias (Oglesby et al., 2000): In Basel, participants of direct monitoring as compared to non-participants were more likely to

live at streets with low traffic volume; although in Helsinki, traffic volume was neither significantly related to participation in direct

nor indirect monitoring, the point estimates indicate a tendency to decreased participation with increasing traffic intensity at home.)

BRE, UK: The Building Research Establishment (BRE) has conducted a national survey of air pollutants in 876 homes in England,

to increase knowledge of pollutant levels and the factors associated with high concentrations. CO levels were measured, once in each

home, using Dräger colorimetric diffusion tubes (passive samplers), as two-week time-weighted averages in the kitchen and main

bedroom of each home. All measurements were performed between October 1997 to February 1999.

Comments

Available measurements from the Expolis study are twofold:

- the 48-h data are ‘through-the-day’ personal average exposures experienced by the working population in European

metropolitan areas and, hence, include the contribution of outdoor exposure (commuting). A typical pattern of CO

levels as measured (1-min resolution) during a 48-h monitoring cycle is shown in Figure 6.1. As examplified by the

Figure, short-term exposures to higher CO levels only marginally affect a 48-h average outcome. The 48-h data are

used in the present risk characterization assuming that even 8-h extrapolations from the overall dataset would not be

affected by short-term peak exposures. As outlined in the study conclusions (Georgoulis et al., 2002), the 48-h

personal CO exposures were highly and consistently associated with urban ambient air concentrations, consistently

but less signifcantly associated with exposure to ETS and inconsistently and mostly insignifcantly associated with

exposure to the use of gas appliances. Time spent in commuting had an insignifcant effect on the 48-h average CO

exposure.

The INDEX project Final report

- the 1-h average personal exposure measurements were exclusively taken indoors. As concluded by the study

members, the presence of smoking or gas cooking in residences signifcantly increased the short-term CO exposures;

the difference with the 48-h averages were explained by the relatively short duration of many activities. The exposure

data are here considered indicative for common indoor-activity related short-term exposures experienced by the

general population living in (sub-)urban European areas.

Significant differences in both personal exposure and ambient levels (Georgoulis et al., 2002) were found within the

five cities, ranging from high values in Milan and Athens to low in Helsinki.

It is noteworthy that, for CO, personal and not microenvironmental samples were drawn in the Expolis study,

considering that people experience various CO exposures due to their daily activities and to unhomogeneous CO levels

within indoor envornments, and that seasonal effects on CO levels were taken into account, supposed to be largely due

to indoor sources.

The BRE study concluded (530 homes; 2-week averaged microenvironmental samples in kitchens and bedrooms) that

CO levels were very low for most of the time in the majority of homes in England. Identified factors influencing long-

term CO levels were: gas cooking, unflued heaters, smoking and outdoor air (in urban locations). Seasonal effects were

largely associated to indoor sources and extract fans were found to have little effect on CO levels in the kitchen.

Figure 6.1: Typical result of a 48-h monitoring cycle (1-min resoltution) obtained during the Expolis study

Doubts could be expressed on whether the selected exposure data could be considered as descriptive of the nowadays

situation, when considering the recent evolution of ambient CO levels. Current ambient air quality regulations, i.e. those

contributing to the reduction of carbon dioxide emissions (the Kyoto Protocol), the EU Limit Value for CO in ambient air

and equivalent national regulations in the EU, have shown to be effective with respect to a reduction of CO ambient levels,

as examplified in Figure 6.2 for a german metropolitan area, the trend expected to continue in the years to come.

The INDEX project Final report

Figure 6.2: CO levels in urban areas - trend in the Ruhr-Rhine region

EU limit value for carbon monoxide in ambient air – Directive 2000/69/EC

Limit value for the protection of human health:

Averaging period: Maximum daily 8-hour mean

Limit value: 10 mg/m

Margin of tolerance: 6 mg/m

on 13 December, reducing on 1 January 2003 and every 12 months thereafter by 2 mg/m

to reach 0 % by 1 January

2005

Date by which limit value is to be met: 1 January 2005

The maximum daily 8-hour mean concentration will be selected by examining 8-hour running averages, calculated from hourly data and updated each

hour. Each 8-hour average so calculated will be assigned to the day on which it ends. i.e. the first calculation period for any one day will be the period

from 17:00 on the previous day to 01:00 on that day; the last calculation period for any one day will be the period from 16:00 to 24:00 on that day.

Result

Available exposure data confim that Carbon Monoxide (CO) sources in EU-residences are contributing to short-term

rather than to long-term exposures. Personal exposure outcomes averaged over 1-hour were considered of moderate

concern even for the most susceptible subpopulations. Nevertheless, uncertainties resulting from the predictive

capabilities of the CFK-model in individuals exposed to low CO concentrations and its applicability to sensitive

subpopulations, suggest that about 10% of the general non-smoking population experience CO levels which could be

hazardous for individuals with heart diseases. Increased exposures could be expected for residences in the vicinity of

busy city streets.

In addidion, there is no evidence that long-term CO exposures in EU residences contribute to carboxyhaemoglobin

levels in blood higher than the baseline levels resulting from endogenous production in normal, non-smoking

individuals.

On the other hand and in contrast with all other chemicals assessed in the present report, carbon monoxide causes a

considerable number of deaths and acute poisonings in the general population (with complications and late sequelae).

Also, individuals suffering from CO poisoning are often unaware of their exposure because symptoms are similar to

those associated with viral illness or clinical depression. In indoor environments, these health risks are nearly

completely associated with the incorrect use of combustion devices or faulty unvented gas appliances.

The INDEX project Final report

Nitrogen dioxide

Synonyms: -

CAS Registry Number: 10102-44-0

Molecular Formula: NO

1. Compound identification

Nitrogen dioxide is a reddish-brown gas in colour, and a strong oxidant with a characteristic pungent odour. It is soluble

in water. A low vapour pressure of 900 mm Hg at 25

C indicates nitrogen dioxide will exist as a gas in the ambient

conditions. It is corrosive and highly oxidising gas. Typically less than 10% by volume of the total emissions of NOx

from combustion sources is in the form of NO

(WHO 1997). NO is rapidly oxidised to NO

by available oxidants,

particularly ozone, in ambient conditions. In indoor air, it is generally a much slower process. Photolysis of NO

sunlight is the most important source of ozone (Alm 1999).

Nitrogen dioxide is also an important atmospheric gas, in addition to health effects, also because it absorbs visible solar

radiation and, therefore contributes to atmospheric visibility, it could have a role in global climate change, it is with

nitric oxide (NO), a main regulator of the oxidising capacity of the free troposphere and ozone (O

) concentrations in

the troposphere (WHO 2000).

There is no clear evidence for a concentration-response relationship for NO

. Asthmatics and persons with chronic

obstructive pulmonary disease (COPD) are the most susceptible populations to acute changes in lung function, airway

responsiveness and other respiratory symptoms. (WHO 2000).

2. Physical and Chemical properties

Molecular weight (g/mol) 46.01

Melting point (°C) -11.2

Boiling point (°C) 21.2

Density (at 20 °C, 1 atm) 1.448

Relative density (air =1) 1.59

Solubility Soluble in sulphuric and nitric acid

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 1.912 µg/m

1 µg/m

= 0.523 ppb

Sources: Alm (1999), Verschueren 2001, HSDB (2003)

3. Indoor Air Exposure assessment

Emission sources

The most important indoor sources of NO

include gas appliances such as stoves, ovens, space and water heaters, and

unflued kerosene heaters. Especially, gas stoves with pilot light have been found strong indoor sources of NO

Typically, NO

concentrations in homes with gas stoves are 2-5 times higher than in homes with electric stoves. Indoor

concentrations of NO

are typically higher in winter than in summer, probably due to increased use of gas heating,

lower ventilation rates, and higher outdoor concentrations (Alm 1999).

The INDEX project Final report

The main ambient sources of nitrogen oxides emissions include intrusion of stratospheric nitrogen oxides, bacterial and

volcanic action, and lightning. Fossil fuel power stations, motor vehicles and domestic combustion appliances emit

nitric oxide (NO), which is a reactive compound forming rapidly NO

. Other contributions of nitrogen dioxide to the

atmosphere come from specific non-combustion industrial processes, such as the manufacture of nitric acid, the use of

explosives and welding. Ambient concentrations of NO

in urban areas peak during the morning and evening rush hours

due to increased emissions of NO from motor vehicles (COMEAP 1997, WHO 1997, WHO 2000).

Indoor air and exposure concentrations

Indoor concentrations vary widely depending on the presence of special indoor sources of NO

. Concentrations in

homes without NO

sources are typically lower than outdoor concentrations and in those cases indoor levels are driven

by outdoor sources. In the recently published ECRHS II study, carried out in 21 European cities annual ambient NO

concentrations ranged from 4.9 µg/m

in Reykjavik, Iceland to 72 µg/m

in Turin, Italy (Hazenkamph-von Arx et al

2004). The average winter/summer ratio showed considerably higher concentrations in winter than in summer, being

1.5. According to WHO (2000) annual mean ambient concentrations in urban areas throughout the world typically

range from 20 to 90 µg/m

. Typical indoor air concentrations in homes with gas cooking vary between 25 and 200

µg/m

over a period of several days. Maximum indoor 1-hour peaks may reach up to 2000 (COMEAP 1997, WHO

2000).

In Helsinki and Prague average indoor concentrations were lower than the respective outdoor concentrations that ranged

from 24 to 61 µg/m

(Kousa et al 2001). Mean indoor concentrations in Europe ranged from 13 µg/m

in UK to 43

µg/m

in Prague according to the results by Kousa et al (2001) and COMEAP (1997). Average personal exposures were

equal or slightly higher.

In an Italian population based study, the highest weekly indoor NO

concentrations were measured in a rural area of Po

Delta (Simoni et al 2002). Weekly mean indoor concentration in kitchen in winter was higher than the respective

concentration in summer, being 62 µg/m

and 38 µg/m

. The presence of gas-furnace heating was identified as the

determining factor for the elevated NO

concentrations.

In a Spanish study of 340 dwellings between 1996 and 1999 (Garcia-Algar et al 2003), average annual indoor

concentrations of NO

did not vary significantly, ranging only from 12.5 to 14.7 µg/m

. Respective outdoor air

concentrations were slightly higher in 1996 and 1998 and slightly lower in 1997 and 1999, thus typical indoor-outdoor

ratios were close to 1. The principal indoor sources of NO

in Spanish homes were the use of gas cooker, the absence of

an extractor fan when cooking, the absence of central heating and cigarette smoking.

Levy et al (1998) studied NO

concentrations in homes in 18 cities in 15 countries reporting 2-day means ranging from

10 µg/m

to 81 µg/m

and personal exposures from 21 µg/m

to 97 µg/m

. Use of a gas stove was found the dominant

activity influencing indoor concentrations with an increase in indoor-outdoor ratios from 0.7 to 1.2 for participants

using gas stoves. Results showed also the importance of combustion space heaters to elevated NO

concentrations. On

the contrary, wood-burning appliances were not related to elevated NO

concentrations in a Canadian study carried out

in 49 houses (Levesque et al 2001).

Brown et al (2002) studied emissions of unflued gas heaters in room chamber (32.4 m

) tests. NO

concentrations in the

chamber ranged from 180 µg/m

to 530 µg/m

due to emissions from the gas heaters. Lee and Wang (2004) studied

nitrogen dioxide emissions from incense burning also in the chamber (18 m

) tests. They found that maximum nitrogen

dioxide concentrations peaked up to 91 µg/m

during incense burning.

Other studies showed indoor concentrations of NO

in homes without gas stoves ranging from 13 µg/m

to 40 µg/m

and with gas stoves from 25 µg/m

to 70 µg/m

in UK (COMEAP 1997). Determinants of NO

exposures were studied

with preschool children in Finland. Preschool located in down town, gas stoves and parent smoking at home were found

the most important factors of the increased NO

exposures for those children (Alm 1999).

Emission sources and microenvironments where high concentrations of NO

occur are reviewed in Table 3.1. Elevated

concentrations were typically related to gas cooking, gas heating, and incense burning. High NO

concentrations were

also found in some public buildings. The highest maximum values up to 7530 µg/m

were measured in indoor ice

arenas (Pennanen et al 1997). Also in kitchens with gas appliances high concentrations, 3808 µg/m

, were determined

(Lebret et al 1987).

Residential indoor air concentrations of NO

in European urban populations are presented in Figure 3.1 and Figure 3.2.

Respective exposure concentrations are presented in Figure 3.3.

The INDEX project Final report

100

0 20 40 60 80 100 120 140

Indoor concentration (µg/m

)

Cumulative frequency (%)

Hel

Bas

Oxf

Pra

PoD

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of nitrogen dioxide in Helsinki (Hel, n=

175), Basel (Bas, n= 50), Prague (Pra, n= 33) (Kousa et al 2001), Oxford (Oxf, n=40) (Hanninen et al 2004, EXPOLIS

2002) and in Po Delta (PoD) (n = 112, AM = 62 µg/m

, std = 32 µg/m

) region in Northern Italy (Simoni et al 2002).

100

0 102030405060708090100

Indoor concentration (µg/m

)

Cumulative frequency (%)

NO2 14-day

Figure 3.2. Cumulative frequency distribution of 14-day indoor air concentrations of nitrogen dioxide in the kitchens in

England (GM= 21.8 µg/m

, max 620 µg/m

, n=845, Coward et al 2002).

The INDEX project Final report

100

0 20406080100120140

Exposure (µg/m

)

Cumulative frequency (%)

Hel

Bas

Oxf

Pra

Figure 3.3. Cumulative frequency distributions 48-hour personal exposure concentrations of nitrogen dioxide in

Helsinki (Hel, n= 177), Basel (Bas, n= 50), Prague (Pra, n= 35) (Kousa et al 2001) and Oxford (Oxf, n=42) (Hanninen

et al 2004, EXPOLIS 2002).

Table 3.1. Short time NO

concentrations related to specific microenvironments or emission sources.

Environment/

emission source

Location Averaging time Reference

M GM median max

Indoor ice arenas Finland max 15-min 320 - 7530 Pennanen et al 1997

max 1-hour 270 - 7440

Gas appliances The Netherlands max 1-min kitchen 400 - 3808 Lebret et al 1987

max 1-hour kitchen 230 - 2055

max 24-hour kitchen 53 - 478

Gas stoves The Netherlands kitchen 2500 Noy et al 1990

Indoor ice arenas Finland 1-week 2 - 1838 Pennanen et al 1998

ice resurfacers propane 396

gasoline 283

elecrtic 25

Homes with gas appliances several hours 200 WHO 2000

1-hour 2000

Gas cooking UK 1-week gas stove 28 - 107 Ross 1996

max 1-hour 342 - 1585

1-week electric stove 23 - 26

max 1-hour 38 - 55

Unflued gas heaters room chamber tests 180 - 530 Brown et al 2002

Homes Sweden 24-hour urban dwellings 6.7 11 Sakai et al 2004

Japan 24-hour urban dwellings 98 369

Indoor ice ring Hong Kong, China 15-min gasoline-fueled 58 - 91 Guo et al 2004

15-min propane-fueled 242

Homes 15 countries 2-day indoor 10 - 81 Levy et al 1998

exposure 21 - 97

Incense burning Chamber tests during burning 17-91 Lee and Wang 2004

Exposure, schoolchildren Sweden 24-hour urban 13 Berglund et al 1994

rural 7

AM = arithmetic mean, GM = geometric mean, max = maximum value

Concentration (µg/m

)

The INDEX project Final report

4. Toxicokinetics

Absorption

Inhaled nitrogen dioxide (NO

) reaches easily the lower respiratory tract (bronchioli and alveoli of the lungs) and

dissolves in lung fluids, hydrating slowly to nitrous (HNO

) or nitric acid (HNO

). These acids dissociate into nitrite

(NO

) or nitrate (NO

) ions, which are rapidly taken up in the lung epithelium (Goldstein ate al., 1977; Postlethwait

and Bidani 1989). NO

acts as a strong oxidant and likely reacts with unsaturated lipids and functional proteins in cells

either within the lung epithelial lining fluid or in the epithelial cell membrane. This may result in loss of cell

permeability and control. Furthermore, the substance activates peroxide detoxification pathways (WHO97).

In Bauer et al. (1986) fifteen asthmatic subjects were exposed to 0.6 mg/m

via mouthpiece for 20 minutes at rest,

followed by 10 minutes of exercise. Expired NO

concentrations were measured continuously. NO

deposition was

72±2% at rest, increasing to 87±1% with moderate exercise. These findings indicate that the NO

dose to the distal

airways and alveolar space, and therefore toxic effects in this region, would be substantially increased by exercise.

Mathematical modelling studies show that the deposition of nitrogen dioxide in the tissues of the lower respiratory tract

is predicted to be maximal at about the junction of the conducting airways and the gas exchange region of the lungs in

humans, rats, guinea pigs and rabbits. Although the actual tissue dose at a similar starting tracheal concentration differs

across species, the shape of the deposition curves is roughly equivalent. The region predicted to receive the maximal

dose is that where the typical nitrogen-dioxide-induced morphometric lesion is observed in several species of animals.

Using this mathematical model, it also can be predicted that, as tidal volume increases in humans (e.g. in exercise), the

dose to the tissue of the gas exchange region increases substantially more than the dose to the conducting airways

(Miller et al., 1982; Overton, 1984; Overton et al., 1987).

Distribution

After passing the lungs, nitrogen dioxide products are distributed to all parts of the body via the bloodstream. Goldstein

et al. (Goldstein ate al., 1977) exposed two female rhesus monkeys to radiolabelled nitrogen dioxide (

) for 6 to 9

minutes through a facemask. The investigators found the radioactive tracer back in the circulation. They also reported

that the time-concentration relationship of

N in arterial blood correlated well with the time-concentration relationship

in the lungs. In addition, the investigators showed that the gas was distributed throughout the lungs.

Elimination

No human data have been found. Saul and Archer (1983) observed that nitrogen dioxide or its products (nitrite, nitrate)

were excreted as nitrate in the urine of Sprague-Dawley rats. Also, they showed that the amount of urinary nitrate

correlated linearly to nitrogen dioxide exposure levels. The urinary values returned to normal after four days.

5. Health effects

Nitrogen dioxide is thought to damage lungs in three ways: (1) it is converted to nitric and nitrous acids in the distal

airways, which directly damages certain structural and functional lung cells; (2) it initiates free radical generation,

which results in protein oxidation, lipid peroxidation, and cell membrane damage; and (3) it reduces resistance to

infection by altering macrophage and immune function.

Effects of short-term exposure

Numerous controlled clinical studies have examined lung function, airway responsiveness to pharmacological, physical

(e.g. cold air) or natural (i.e. allergens) bronchoconstrictors and the defence against viral and bacterial airway infections

in human subjects exposed to nitrogen dioxide.

Generally, concentrations in excess of 1880 µg/m

(1 ppm) are necessary during acute controlled exposures to induce

changes in pulmonary function in healthy adults (U.S.EPA, 1993; Berglund et al., 1993; Wagner, 1985). Because these

concentrations rarely occur in indoor air, concern about the effects of nitrogen dioxide has been focused on people with

The INDEX project Final report

pre-existing lung disease. There have been numerous studies of people with asthma, chronic obstructive pulmonary

disease, or chronic bronchitis showing that exposure to low levels of nitrogen dioxide can cause small decrements in

forced vital capacity and forced expiratory volume in 1 second (FEV

) or increases in airway resistance.

The lowest level of nitrogen dioxide exposure reported in more than one laboratory to show a direct effect on

pulmonary function in asthmatics was a 30-minute exposure, with intermittent exercise, to 560 µg/m

(0.3 ppm) (Roger

at al., 1990; Bauer et al., 1986). Although similar but statistically non significant trends have been observed in other

controlled human studies performed at lower concentrations (Bylin et al., 1985; Hazucha et al., 1983; Orehek et al.,

1976), the small size of the decrements and questions regarding the statistical significance of some of these results

together suggest that caution should be exercised in accepting these findings as demonstrating acute effects.

Orehek and colleagues (Orehek et al., 1976) were the first to report that relatively brief exposures of asthmatics to low-

level NO

(0.19 mg/m

) might enhance subsequent responsiveness to challenge with a broncho-constricting drug.

Although NO

alone caused an increase in airway resistance in only 3 of 20 asthmatics, bronchial responsiveness to

carbachol increased in 13 of these 20 subjects. However, this report was challenged because of the retrospective

separation of responding from non-responding subjects.

Hazucha and colleagues (Hazucha et al., 1983) failed to confirm these results in a study of 15 asthmatic subjects.

Although there were some differences in techniques and patient selection between the Orehek and Hazucha studies, it

seems likely that the findings of Orehek and coworkers reflect a retrospective stratification of subjects into “responder”

and “non-responder” groups that was not justified a priori.

Other investigators have also been unable to confirm effects of 0.19–0.38 mg/m

(0.1-0.2 ppm) NO

on lung function in

either asthmatic adolescents (Koenig et al., 1985; Koenig et al., 1988) or in mildly asthmatic adults (Koenig et al., 1985;

Bauer et al., 1986; Orehek et al., 1976; Bylin et al., 1985; Hazucha et al., 1983; Koenig et al., 1988; Kleinman et al.,

1983; Linn et al., 1986; Mohsenin & Gee, 1987; Morrow & Utell, 1989; Roger et al., 1990).

Kleinman and colleagues (Kleinman et al., 1983) evaluated the response of lightly exercising asthmatic subjects to

inhalation of 0.38 mg/m

for 2 hours, during which resting minute ventilation was doubled. Although NO

did not

cause alterations in flow rates or airways resistance, approximately two-thirds of the subjects experienced increased

responsiveness to methacholine after inhalation of NO

compared with clean air, as assessed by specific airway

resistance.

In view of the inconclusive findings at 0.19 and 0.38 mg/m

, Bauer and colleagues (Bauer et al., 1986) studied the

effects of mouthpiece exposure to 0.56 mg/m

for 30 minutes (20 minutes at rest followed by 10 minutes of

exercise at approximately 40 L/min) in 15 asthmatics. At this level, NO

inhalation produced significant decrements in

forced expiratory flow rates after exercise, but not at rest. Furthermore, after airway function was allowed to return to

baseline during a 1-hour recovery period, isocapneic cold-air hyperventilation elicited increased airway responsiveness

in the asthmatics who had earlier been exposed to NO

Roger and coworkers (Roger et al., 1990), in a comprehensive, concentration response experiment, were unable to

confirm the results of a previous pilot study suggesting airway responses in asthmatic subjects. Twenty-one male

asthmatics exposed to NO

at 0.28, 0.38, and 1.13 mg/m

for 75 minutes did not experience significant effects on lung

function or airway responsiveness compared with air exposure. Bylin and coworkers (Bylin et al., 1985) found

significantly increased bronchial responsiveness to histamine challenge compared with sham exposure in 8 atopic

asthmatics exposed to 0.56 mg/m

for 20 minutes. Five of 8 asthmatics demonstrated increased reactivity, while 3

subjects showed no change, as assessed by specific airway resistance. Mohsenin (Mohsenin, 1987) reported enhanced

responsiveness to methacholine in eight asthmatic subjects exposed to 0.94 mg/m

at rest for 1 hour; airway

responsiveness was measured by partial expiratory flow rates at 40% vital capacity, which may have increased the

sensitivity for detecting small changes in airway responsiveness.

Strand et al. (Strand et al., 1996) found increased responsiveness to histamine among 19 asthmatic subjects 5 hours after

a 30 minute exposure to 0.49 mg/m

, with intermittent exercise. The inconsistent results of these studies have not

been satisfactorily explained. It is evident that a wide range of responses occur among asthmatics exposed to NO

. This

variation may in part reflect differences in subjects and exposure protocols: mouthpiece vs. chamber, obstructed vs.

non-obstructed asthmatics, sedentary vs. exercise, and requirements for medication. Identification of factors that

predispose to NO

responsiveness requires further investigation.

These studies have typically involved volunteers with mild asthma; data are needed from more severely affected

asthmatics who may be more susceptible. Overall, there is little convincing evidence that short-term exposures to NO

at outdoor ambient concentrations significantly alter lung function or nonspecific airway responsiveness in most people

The INDEX project Final report

with mild asthma. However, outdoor levels influence indoor concentrations, which may reach peak levels that are

clinically important for some adults and children with asthma.

Effects on Allergen Responsiveness:

The potential for NO

exposure to enhance responsiveness to allergen challenge in asthmatics deserves special mention.

Several recent studies have reported that low-level exposures to NO

,both at rest and with exercise, enhance the

response to specific allergen challenge in mild asthmatics. Tunnicliffe et al. (Tunnicliffe et al., 1994) reported exposures

of 8 subjects with asthma to 400 ppb NO

for only 1 hour at rest, and found increased responsiveness to a fixed dose of

allergen, both during the early and late phases of the response. No significant effect was seen at 100 ppb, but the data

suggested an exposure-response relationship. Davies’ group from the U.K., in two reports (Devalia et al., 1994;

Rusznak et al., 1996), described an effect of exposure to the combination of 400 ppb NO

and 200 ppb SO

, but not

either pollutant alone, on subsequent allergen challenge in mild asthmatics.

Strand and colleagues (Strand et al., 1998) from Sweden demonstrated increases in both the early and late phase

responses to allergen following 4 daily repeated exposures to 260 ppb NO

for 30 minutes, at rest. Finally, Jenkins et al.

(Jenkins et al., 1999) exposed asthmatic subjects to NO

, ozone, and their combination using two different protocols

that varied time of exposure and gas concentration, but kept the total exposure constant. All three exposures of the high

concentration regimen (200 ppb ozone, 400 ppb NO

, and the combination for 3 hours), but not the low concentration

regimen, enhanced subsequent responsiveness to allergen.

Additional data from both animal exposure and in vitro exposure studies provide support for enhancement of allergen

responsiveness by NO

exposure. Gilmour (Gilmour, 1995) has reviewed the evidence in animal models. Of particular

interest is a rat model of house-dust-mite sensitivity in which a 3-hour exposure to 9.4 mg/m

, after a priming

injection and pulmonary challenge with antigen, increased the specific immune response and immune-mediated

pulmonary inflammation. NO

exposure also enhanced lymphocyte proliferation responses to allergen in both the spleen

and mediastinal lymph nodes.

Schierhorn et al. (Schierhorn et al., 1999) observed increased histamine release by cultured human nasal mucosa from

surgical resections in response to exposure to NO

at 200 and 800 µg/m3 (106 and 424 ppb) for 24 hours. The

magnitude of the effect was more pronounced than for ozone.

These recent studies involving allergen challenge appear relatively consistent in demonstrating effects at concentrations

that occur indoors, and suggest that NO

may enhance both allergen sensitization and its associated inflammatory

response. Confirmation of these findings is needed from other centers. However, the rising incidence, prevalence, and

mortality from asthma make these observations particularly important and timely. Additional work is needed in

understand more completely the exposure-response characteristics, effects of exercise, relationship to severity of

asthma, role of asthma medications, and other clinical factors. Animal and in vitro studies are needed to establish the

precise mechanisms involved.

Chronic Obstructive Pulmonary Disease:

Few studies have examined responses to NO

in subjects with chronic obstructive pulmonary disease (COPD). In a

group of 22 subjects with moderate COPD, Linn and associates (Linn et al., 1985a) found no pulmonary effects of 1-

hour exposures to 0.94, 1.88, and 3.76 mg/m

. In a study by Morrow and colleagues (Morrow et al., 1992), 20

subjects with COPD were exposed for 4 hours to 0.56 mg/m

in an environmental chamber, with intermittent

exercise. Although progressive decrements in lung function occurred during the exposure, significant decreases were

not found for FVC until the end of the exposure. The decrement in lung volume occurred without changes in flow rates.

The difference in results between the Linn and Morrow studies may reflect the difference in duration of exposure. It is

worth noting that changes in lung function were typical of the “restrictive” pattern seen with ozone rather than the

obstructive changes described by some with NO

exposure in asthmatics.

Effects on Host Defense:

More recently, investigators have sought to evaluate the effects of nitrogen dioxide on measures other than on

pulmonary function. The key studies are summarized here. Analysis of lung lavage from healthy humans indicated that

high levels (5640–7520 µg/m

; 3–4 ppm) reduce the activity of alpha-1-protease inhibitor, a protein that acts to protect

the lung from the proteolytic enzyme elastase by inhibiting connective tissue damage. However, 2820 µg/m

(1.5 ppm)

had no such effect.

Goings et al. (Goings et al., 1989) exposed healthy volunteers to either 1.9-5.6 mg/m

or to air for 2 hours per day

for 3 consecutive days. A live, genetically engineered influenza A vaccine virus was administered intranasally to all

subjects after exposure on day 2. Infection was determined by virus recovery from nasal washings, a 4-fold or greater

increase in antibody titer, or both. The findings of this study were inconclusive, in part because of limitations in sample

The INDEX project Final report

size. In addition, the attenuated, cold-adapted virus used in the study was incapable of infecting the lower respiratory

tract, where NO

may have its greatest impact on host defense.

Another approach has been to obtain lavaged cells from NO

-exposed individuals and examine their handling of

infectious virus in vitro. Several NO

exposure scenarios, including continuous low-level exposure or intermittent peak

exposures have been examined (Frampton et al., 1989). Alveolar macrophages obtained by BAL 3 1/2 hours after a 3-

hour continuous exposure to 1.1 mg/m

tended to inactivate influenza in vitro less effectively than cells collected

after air exposure. The effect was observed in cells from 4 of the 9 subjects studied; alveolar macrophages from these 4

subjects increased release of interleukin-1 after exposure to NO

, whereas cells from the remaining 5 subjects decreased

release of interleukin-1 following exposure. However, in a subsequent study (Azadniv et al., 1998) involving 3.8 mg/m

exposures for 6 hours with intermittent exercise, no effect on alveolar macrophage function or inactivation of

influenza virus was observed, either immediately or 18 hours after exposure.

Airway Inflammation:

Unlike ozone exposure, NO

exposure at near-ambient levels (i.e., less than 3.8 mg/m

) does not cause a significant

influx of polymorphonuclear leukocytes (PMN) into the airways and alveoli (Frampton et al., 1989). NO

appears to be

much less potent than ozone in eliciting a neutrophilic inflammatory response.

However, prolonged exposure to NO

at concentrations only slightly above peak levels occurring indoors can cause

mild airway inflammation. Healthy volunteers exposed to 3.8 mg/m

for 6 hours with intermittent exercise

(Azadniv et al., 1998) showed a slight increase in the percentage of PMN obtained in bronchoalveolar lavage fluid 18

hours after exposure (air: 2.2±0.3%; NO

: 3.1±0.4%). In a separate group of subjects, no effects of this exposure

protocol were found on alveolar macrophage phenotype or expression of the adhesion molecule CD11b or receptors for

IgG when assessed immediately after exposure (Gavras et al., 1994). Blomberg et al. (Blomberg et al., 1997) reported

that 4-hour exposures to 3.8 mg/m

resulted in an increase in interleukin-8 and PMN in the proximal airways of

healthy subjects, although no changes were seen in bronchial biopsies. This group also studied the effects of repeated 4-

hour exposures to 3.8 mg/m

on 4 consecutive days, with BAL, bronchial biopsies, and BAL fluid antioxidant

levels assessed 1.5 hours after the last exposure (Blomberg et al., 1999). The bronchial wash fraction of BAL fluid

showed a two-fold increase in PMN and a 1.5-fold increase in myeloperoxidase, indicating persistent mild airway

inflammation with repeated NO

exposure. Interestingly, small but significant decrements in FVC and FEV1 were

observed after the first exposure, which returned to baseline following subsequent exposures.

There is evidence from both animal and human studies that exposure to NO

may alter lymphocyte subsets in the lung

and possibly in the blood. Lymphocytes, particularly cytotoxic T cells and NK cells, play a key role in host defense

against respiratory viruses by eliminating infected host cells. Richters and colleagues (Damji & Richters, 1989)

(Richters & Damji, 1988; Richters & Richters, 1989; Kuraitis & Richters, 1989) showed that mice exposed to NO

levels as low as 7.5 mg/m

for eight hours demonstrate reductions in populations of CD8+ (cytotoxic/suppressor)

lymphocytes in the spleen. In humans, Sandstrom et al. (Sandstrom et al., 1991) observed a significant, dose-related

increase in lymphocytes and mast cells recovered by BAL 24 hours after a 20-minute exposure to NO

at 4.23 – 10.3

mg/m

. Rubinstein et al. (Rubinstein et al., 1991) found that a series of 4 daily 2- hour exposures to 1.1 mg/m

resulted in a small increase in NK cells recovered by BAL. In contrast, repeated exposures to 2.8 or 7.5 mg/m

for

20 minutes every 2nd day on six occasions resulted in decreased CD16+56+ and CD19+ cells in BAL fluid, 24 hours

after the final exposure (Sandstrom et al., 1992b; Sandstrom et al., 1992a). No effects were seen on PMN or total

lymphocytes. Finally, Azadniv et al. (Azadniv et al., 1998) observed a small but significant reduction in CD8+ T

lymphocytes in peripheral blood, but not BAL, 18 hr following single 6 hour exposures to 3.8 mg/m

Differing exposure protocols and small numbers of subjects among these studies may explain the varying and

conflicting findings. Furthermore, the clinical significance of transient, small changes in lymphocyte subsets is unclear.

However, even small changes in susceptibility to respiratory viruses resulting from exposure to NO

may have a

significant public health impact because of the large number of individuals exposed in the home, both to NO

and to

respiratory viruses. However, clinical studies provide little evidence for effects on lung function, airway inflammation,

or host defense impairment in healthy subjects at outdoor ambient exposure concentrations.

Induction of Emphysema:

Clinical emphysema in humans has been linked with deficient proteinase inhibitor activity in the lung, presumably via

inactivation by cigarette smoke. One mechanism by which chronic NO

exposure may result in structural lung injury is

through inactivation of lung proteinase inhibitors. Animal models involving prolonged exposure to relatively high levels

of NO

have found pathological changes of emphysema (Evans et al., 1976; Lafuma et al., 1987). Mohsenin and Gee

(Mohsenin and Gee, 1987) exposed healthy volunteers to 5.6 or 7.5 mg/m

for 3 hours and observed a 45%

decrease in the functional activity of a1-proteinase inhibitor in BAL fluid. Supplementation with vitamins C and E prior

to exposure abrogated the effect of 7.5 mg/m

on elastase inhibitory capacity of the alveolar lining fluid

The INDEX project Final report

(Mohsenin, 1991). In contrast, Johnson et al. (Johnson et al., 1990) found no effect of exposure for 3 hours to

continuous 2.8 mg/m

or intermittent peaks of 3.8 mg/m

on either the concentration (immunoassay) or functional

activity of a1-proteinase inhibitor in BAL fluid. The absence of an effect in the Johnson study may reflect the lower

exposure levels used.

Frampton et al. (Frampton et al., 1989) observed a 47% increase in a2-macroglobulin, a metalloproteinase inhibitor

released by alveolar macrophages, in BAL fluid 3 and 1/2 hours following 3-hour exposures to 1.1 mg/m

. This

protein may have local immunoregulatory effects as well as provide local protection against proteinases. Its increase

following NO

exposure suggests a protective response. However, no change in BAL fluid levels of a2-macroglobulin

was observed following similar exposures to 2.8 mg/m

of NO

(Frampton et al., 1989).

Summary of short-term exposure effect levels

NOAEL

mg/m

LOAEL

mg/m

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

0.47

increase in airway reactivity

sensitive humans

(asthmatics), 1h

Review: California Air

Resources Board

(CARB), 1992

OEHHA 1999

REL: 0.47

0.19 0.38-0.56 increased responsiveness to bronchoconstrictors Mild asthmatics, 0.5h WHO 2000

GV: 0.2 (1h)

0.96 possible detrimental respiratory effects in both

normal and asthmatic subjects

Short-term human

clinical studies

Health Canada

1995

ASTER: 0.48 (1h)

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of Acute Reference Exposure Level (protective against mild adverse effects): 0.25 ppm (470 mg/m³)

California Ambient Air Quality Standard

Study California Air Resources Board (CARB), 1992

Study population sensitive humans (asthmatics)

Exposure method inhalation

Critical effects increase in airway reactivity

LOAEL

NOAEL 0.25 ppm

Exposure duration 1 hour

Extrapolated 1 hour concentration 0.25 ppm

LOAEL uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 1

Cumulative uncertainty factor 1

Reference Exposure Level 0.25 ppm (0.47 mg/m³; 470 mg/m³)

WHO Guidelines (2001)

Despite the large number of acute controlled exposure studies on humans, several of which used multiple

concentrations, there is no evidence for a clearly defined concentration–response relationship for nitrogen dioxide

exposure. For acute exposures, only very high concentrations (1990 µg/m3; > 1000 ppb) affect healthy people.

Asthmatics and patients with chronic obstructive pulmonary disease are clearly more susceptible to acute changes in

lung function, airway responsiveness and respiratory symptoms. Given the small changes in lung function (< 5% drop

in FEV1 between air and nitrogen dioxide exposure) and changes in airway responsiveness reported in several studies,

375–565 µg/m3 (0.20 to 0.30 ppm) is a clear lowest-observed-effect level. A 50% margin of safety is proposed because

of the reported statistically significant increase in response to a bronchoconstrictor (increased airway responsiveness)

with exposure to 190 µg/m3 and a meta-analysis suggesting changes in airway responsiveness below 365 µg/m3. (The

significance of the response at 190 µg/m3 (100 ppb) has been questioned on the basis of an inappropriate statistical

analysis.)

On the basis of these human clinical data, a 1-hour guideline of 200 µg/m

is proposed. At double this recommended

guideline (400 µg/m3), there is evidence to suggest possible small effects in the pulmonary function of asthmatics.

Should the asthmatic be exposed either simultaneously or sequentially to nitrogen dioxide and an aeroallergen, the risk

The INDEX project Final report

of an exaggerated response to the allergen is increased. At 50% of the suggested guideline (100 µg/m3, 50 ppb), there

have been no studies of acute response in 1 hour.

Effects of long-term exposure

Studies with animals have clearly shown that several weeks to months of exposure to NO

concentrations of less than

1880 µg/m

(1 ppm) cause a plethora of effects, primarily in the lung but also in other organs such as the spleen, liver

and blood. Both reversible and irreversible lung effects have been observed. Structural changes range from a change in

cell types in the tracheobronchial and pulmonary regions (lowest reported level 640 µg/m

) to emphysema-like effects

(at concentrations much higher than ambient). Biochemical changes often reflect cellular alterations (lowest reported

levels for several studies 380–750 µg/m

(0.2–0.4 ppm) but isolated cases of lower effective concentrations). Nitrogen

dioxide levels as low as 940 µg/m

(0.5 ppm) also increase susceptibility to bacterial and viral infection of the lung.

Several epidemiological studies show significant relationships between ambient (urban) NO

levels and health effects,

including respiratory symptoms, episodes of respiratory illness, lung function, and even mortality. However, because

shares sources with other pollutants, one of the difficulties in deriving quantitative estimates of health risks is in

separating the relative contributions of NO

from those of other major pollutants (ozone, sulfur dioxide, particulate

matter, etc.); hence nitrogen dioxide might best be considered as just one indicator of polluted ambient air, especially

where it is present in traffic-dominated urban air pollution mixes. Great care is therefore needed in interpreting

available outdoor epidemiology studies, so as not to ascribe an undue degree of confidence to reported health

associations with nitrogen dioxide concentrations as being specifically due to that compound as compared to the overall

ambient air mix.

Indoor studies have compared groups exposed to nitrogen dioxide emitted from the combustion of gas inside buildings

to groups in homes without such sources and hence with lower levels of nitrogen dioxide. The limited studies of adults

have tended to show no relationship between the use of gas for cooking and respiratory symptoms or lung functions.

Thus, the following discussion focuses on children.

Hasselblad et al. (1992) performed a meta-analysis of studies in homes with gas stoves that met several criteria. The

endpoint was the presence of lower respiratory symptoms and disease in children aged 5–12 years. Given the nature of

the studies evaluated, the goal was to estimate the odds ratio of an increase in nitrogen dioxide concentration of

28.3 µg/m

(15 ppb). Although two analytic models (fixed and hierarchical) were used, they gave about the same

results. The combined odds ratio was about 1.2, with confidence intervals ranging from about 1.1 to 1.3. Thus, the

analysis estimated an increased risk of approximately 20% for respiratory symptoms and disease for each increase of

28.3 µg/m

(2-week average), where average weekly bedroom concentrations were between 15 and 122 µg/m

(8 and

65 ppb) (Hasselblad et al., 1992). Several uncertainties exist in this analysis, and exposure measurement errors are

present in the studies.

A number of additional studies have been published since 1990, with continued mixed results. A prospective cohort

study of infants conducted in Albuquerque, New Mexico (Samet et al., 1993) attempted to address many of the issues of

previous studies related to sample size and exposure misclassification. Exposures to NO

and respiratory illnesses were

monitored prospectively from birth to 18 months of age in a cohort of 1205 infants living in homes with gas and electric

cooking stoves, without smoking. NO

exposures were estimated from serial measurements of bedroom NO

concentrations. Respiratory illnesses were quantified from reports of symptoms and illnesses from mothers and

validated by home visits. No consistent trends in incidence or duration of illness were observed by level of NO

exposure at the time of illness or during the prior month, or by type of stove. However, indoor NO

levels were very

low in this study.

To further illustrate the indoor epidemiological findings, the study of Neas et al. (1991) is described in further detail

here. It was selected because the individual symptom results are consistent with the magnitude of effects found in the

British studies and other analyses of the data from six United States cities. The authors evaluated 1286 white children

(7–11 years) from the larger six-city study. For the children selected there was complete covariate information and at

least one indoor measurement of nitrogen dioxide (Palmes passive diffusion tubes for 2 weeks during the heating season

and 2 weeks during the cooling season at three sites: kitchen, activity room and the child’s bedroom). Parents completed

a questionnaire on symptoms during the previous year. An increase in symptoms was estimated for an additional

exposure of 28.3 µg/m3 (15 ppb). A multiple logistic model was used. The odds ratios were as follows: shortness of

breath, 1.23 (95% confidence interval (CI), 0.93–1.61); persistent wheeze, 1.16 (CI, 0.89–1.52); chronic cough, 1.18

(CI, 0.87–1.60); chronic phlegm, 1.25 (CI, 0.94–1.66); and bronchitis, 1.05 (CI, 0.75–1.47).When the authors

performed a multiple logistic regression of the combined lower respiratory symptom measure (i.e. the presence of any

The INDEX project Final report

of the above-mentioned symptoms), they obtained an odds ratio of 1.40 (CI, 1.14–1.72). Table 5.2 shows the adjusted

odds ratios for combined lower respiratory symptoms for ordered nitrogen dioxide exposure categories. These findings

are consistent with a linear concentration–response relationship.

Table 5.2: Odds ratio and 95% confidence interval for the effect of nitrogen dioxide exposure

on the prevalence of lower respiratory symptoms in children (Neas et al., 1991)

Nitrogen dioxide concentration

(µg/m

)

Range Mean

Number of children

Odds ratio

95%Confidence

interval

0–9 7 263 1.00

9–19 14 360 1.06 0.71–1.58

19–37 27 317 1.36 0.89–2.08

38–147 58 346 1.65 1.03–2.63

More recent studies have utilized personal monitoring methods in an attempt to improve exposure classification.

Mukala et al. (Mukala et al., 1999) prospectively studied personal exposure to NO

for periods of 13 weeks among 163

preschool children in Helsinki, using individual passive diffusion monitors. Daily diaries of symptoms were kept by the

parents, and in a subset of 53 children, peak expiratory flow rates were measured in the morning and evening. Co-

variates considered in the model included allergy, education, smoking, stove type, and outdoor pollutant concentrations

(NO, NO

, O

, SO

, and total suspended particles). The median personal NO

exposure was 21.1 µg/m

(11 ppb), with a

maximum of 99 µg/m

(50 ppb). An increased risk of cough was associated with increasing NO

exposure (risk ratio =

1.52; 95% confidence interval 1.00-2.31). There were no significant effects on other respiratory symptoms or peak flow.

In Australia, where unvented natural gas cooking and heating are common, Pilotto et al. (Pilotto et al., 1997) queried

respiratory symptoms and school absences among 388 children from 6 to 11 years of age, and monitored indoor NO

levels at their schools, which were chosen for having either unvented gas heating or electric heating. Classroom

monitoring of NO

levels was conducted intermittently over several months. A significant increase in sore throat, colds,

and absences from school were found for children in environments with hourly peak levels of 150 µg/m

, compared

with background levels of 38 µg/m

. Exposure-response relationships were evident for each outcome. However, no

measurements of other pollutants, either indoor or outdoor, were provided. Caution must be used in interpreting the

findings from cross-sectional studies, because many factors other than pollutant levels may influence differences

between populations.

In the Latrobe Valley of Australia (Garrett et al., 1998), NO

levels were monitored in eighty homes, on 5 separate

occasions for 4 days each, and health questionnaires administered to the 148 children residing in those homes. 58 of the

children were asthmatic, although the diagnostic criteria were not provided. Children underwent allergy prick testing

and monitored their peak flow rates for a 2-week period in the winter and spring. The indoor median NO

concentration

was 11 µg/m

(6.0 ppb), with a maximum of 241 µg/m

(128 ppb). Respiratory symptoms were more common in

children exposed to a gas stove (odds ratio 2.3, CI 1.00-5.2), even after adjusting for NO

levels (odds ratio 2.2, CI 1.0-

4.8). Atopic children tended to have a greater risk than non-atopic children. NO

concentration was not a significant risk

factor for symptoms. The authors conclude that gas stoves may pose a risk apart from NO

. However, the relative

paucity of NO

monitoring data for each home may have provided insufficient statistical power to demonstrate an

association. More important weaknesses in the study are the inclusion of homes with cigarette smokers, and the failure

to monitor other pollutants, either inside or outside the home. These factors may have confounded the findings.

Jarvis et al. (1996) studied symptoms, lung function, and atopy in 15000 adults aged 20-44 years in Britain, as part of

the European Community Respiratory Health Survey. Women, but not men, who reported cooking with gas had an

increased risk for symptoms consistent with asthma, such as wheezing (odds ratio (OR) 2.07, CI 1.41- 3.05), waking

with shortness of breath (OR 2.32, CI 1.25-4.34), and “asthma attacks” (OR 2.60, CI 1.20-5.65). Lung function was

measured in a subset of subjects, and FEV1 was reduced 3.1% of predicted for women cooking with gas compared to

those using other means, after adjusting for age, smoking, and town of residence. Total and specific IgE levels were not

associated with gas stove use. There was no protective effect associated with use of an exhaust fan. The authors boldly

concluded from their estimate of the population attributable risk fraction that “the prevalence of wheeze with

breathlessness in young women would be reduced by between 8% and 48% if cooking with gas were abandoned.”

Although studies such as this are limited by the potential for exposure misclassification and the influence of other

environmental and biological factors, the findings are consistent with women spending more time at cooking than men,

and with reports of increased responsiveness to allergen challenge following NO

exposure (see below).

The INDEX project Final report

Taken together, studies of the health effects of exposure to NO

indoors fail to make a convincing case for association

with respiratory illness in either children or adults. The findings of the Hasselblad meta-analysis (Hasselblad et al.,

1992) must be interpreted with caution because the 11 studies used in the analysis employed varying methodologies and

study populations. Small sample size, potential for misclassification, inclusion of smokers in many of the studies, and

failure to consider potential effects of outdoor pollution, or other indoor pollutants, may bias many of the studies. For

example, burning of natural gas in gas stoves emits ultrafine particles in addition to NO

, and the cooking process is

also a source of particles. It is possible that observed health effects associated with gas stove use may represent health

effects of particle exposure, or of particles combined with NO

. This may explain why Garrett et al. (Garrett et al.,

1998), found a significant relationship between respiratory symptoms in children and gas stove use, but not indoor NO

levels. The Samet et al. study of infants in Albuquerque (Samet et al., 1993) provides convincing evidence that indoor

, at the very low concentrations found in that study, are not associated with respiratory illnesses in children under

18 months of age.

Recently reported time-series analyses of ambient air pollution effects from a European multi-city epidemiology project

(Katsouyanni et al., 1996) also found the following very mixed results.

- There was no association with any cause of death for nitrogen dioxide or ozone in Lyon, France, where daily nitrogen

dioxide concentrations averaged 70 µg/m

(maximum 324 µg/m

) and both sulfur dioxide and particulate matter were

significantly related to mortality from cardiovascular and respiratory causes (Zmirou et al., 1996).

- No relationship between daily nitrogen dioxide levels and daily mortality was found in Cologne, Germany, where the

all-year median for daily nitrogen dioxide was 45 µg/m

(maximum 176 µg/m

) and both sulfur dioxide and

particulate matter were significantly related to increased mortality risk (Spix and Wichmann, 1996).

- In Paris, there were no significant effects of photo-oxidant pollutants (nitrogen dioxide, ozone) on daily mortality

risk, but there was a significantly increased hospital asthma admission relative risk (RR) of 1.175 per 100 µg/m

increase in nitrogen dioxide concentration above the reference value. The mean 24-hour nitrogen dioxide

concentration averaged 45 µg/m

(99th centile 108 µg/m

) and the mean 1-hour maximum was 74 µg/m

(99th centile

203 µg/m

). Particulate matter and sulfur dioxide significantly increased daily mortality from respiratory causes and

hospital admissions for all respiratory disease (Dab et al., 1996).

- Negative (sometimes statistically significant) effects of nitrogen dioxide in decreasing respiratory hospital admissions

(RR = 0.86–0.89) were found in Amsterdam for adults (aged 15–64 years) and in chronic obstructive pulmonary

disease admissions for all ages (RR = 0.795–0.948), but there were some positive effects on respiratory admissions

for the elderly (65+ years) and on asthma admissions (RR = 0.902–1.062). The mean 24-hour nitrogen dioxide

concentration averaged 50 µg/m

(95th centile 84 µg/m

) and the mean 1-hour maximum was 75 µg/m

(95th centile

127 mg/m

). In addition, ozone showed nonsignificant positive effects on summer respiratory admissions for the

elderly (65+ years) but neither particulate matter nor sulfur dioxide showed any clear effects on hospital admissions

(Schouten et al., 1996).

- Rotterdam showed contrasting results, i.e. predominantly positive (albeit few significant at P < 0.05) effects of

nitrogen dioxide in terms of relative risk increases for all respiratory admissions for all age groups and increases

(mostly significant) for chronic obstructive pulmonary disease admissions for all ages. The mean 24-hour nitrogen

dioxide concentration averaged 54 µg/m

(95th centile 86 µg/m

) and the mean 1-hour maximum averaged 82 µg/m

(95th centile 138 µg/m

). Ozone and particulate matter (as black smoke) generally increased the risk of hospital

admissions, and sulfur dioxide showed mixed results (Schouten et al., 1996).

- Ambiguities regarding the effects of ambient nitrogen on hospital admissions for asthma were found in Helsinki

(daily mean in the range 34–44 µg/m

). Pönkä & Virtanen (1996) reported asthma admissions for children (aged 0–14

years) to be related to 8-hour ozone levels and for adults (15–64 years) to 24-hour sulfur dioxide levels, but provided

no results for nitrogen dioxide. This was in contrast to an earlier report (Pönkä, 1991) of significant associations

between hospital admissions for asthma and nitrogen dioxide, sulfur dioxide, ozone and total suspended particulates

based on different statistical modelling.

- In London, the effects of nitrogen dioxide and sulfur dioxide on daily mortality were significant but smaller and less

consistent than the effects of ozone and particulate matter (as black smoke) on respiratory, cardiovascular and all-

cause mortality, the effects being greater in the warm season (April–September) and independent of other pollutants

(Anderson et al., 1996).

The INDEX project Final report

Summary of long-term exposure effect levels

NOAEL

mg/m

LOAEL

mg/m

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

0.038-

0.056

excess of lower respiratory illness Children 5-12 y old,

annual avg conc. , also

basing on background

level of 0.015 mg/m³

taken from EHC 188 WHO 2000

GV: 0.04 (1y)

WHO Guidelines (2001)

Although there is no particular study or set of studies that clearly support selection of a specific numerical value for an

annual average guideline, the database nevertheless indicates a need to protect the public from chronic nitrogen dioxide

exposures. For example, indoor air studies with a strong nitrogen dioxide source, such as gas stoves, suggest that an

increment of about 30 µg/m3 (2-week average) is associated with a 20% increase in lower respiratory illness in children

aged 5–12 years. However, the affected children had a pattern of indoor exposure that included peak exposures higher

than those typically encountered outdoors. Thus the results cannot be readily extrapolated quantitatively to the outdoor

situation. Outdoor epidemiological studies have found qualitative evidence of ambient exposures being associated with

increased respiratory symptoms and lung function decreases in children (most clearly suggestive at annual average

concentrations of 50–75 µg/m3 or higher and consistent with findings from indoor studies), although they do not

provide clear exposure–response information for nitrogen dioxide. In these epidemiological studies, nitrogen dioxide

has appeared to be a good indicator of the pollutant mixture. Furthermore, animal toxicological studies show that

prolonged exposures can cause decreases in lung host defences and changes in lung structure. On these grounds, it is

proposed that a long-term guideline for nitrogen dioxide be established. Selecting a well-supported value based on the

studies reviewed has not been possible, but it has been noted that a prior review conducted for the Environmental

Health Criteria document on nitrogen oxides recommended an annual value of 40 µg/m3 (WHO, 1997). In the absence

of support for an alternative value, this figure is recognized as an air quality guideline.

Carcinogenic and mutagenic effects

To date, there are no reports that nitrogen dioxide causes malignant tumours or teratogenesis (U.S.EPA, 1993; Berglund

et al., 1993; Witschi, 1988). Limited genotoxicity studies have produced mixed results with in vitro and high-

concentration in vivo studies (e.g. 50 000 µg/m

, 27 ppm) (Victorin, 1994).

Interaction with other chemicals

Environmental exposures to NO

do not occur singly, but rather as a complex mixture of pollutants, and failure to

consider the presence of other pollutants may confuse interpretation of the observed effects. Recent data suggest

exposure to low concentrations of NO

at rest may enhance the response to allergen inhalation in subjects with asthma.

When considering mixtures of anthropogenic pollutants, It may be impossible to separate the effects of one component

from those of other components, particularly with the possibilities of synergistic or antagonistic interactions. In

considering the health effects of mixtures, potential causal pathways should be carefully delineated. For example, some

reports have suggested that HONO may contribute to the health effects attributed to indoor NO

(Spengler et al., 1990).

Efforts have been made to study effects of NO

-ozone mixtures on pulmonary function. These studies have generally

revealed no interactive effects; the observed pulmonary function decrements appear to reflect the ozone component of

the mixtures. Hazucha et al., (Hazucha et al., 1994) found that pre-exposure of healthy women to 1.1 mg/m

(0.6 ppm)

for 2 hours enhanced the development of nonspecific airway responsiveness induced by a subsequent 2-hour

exposure to 0.6 mg/m

(0.3 ppm) ozone, with intermittent exercise.

Relatively high-level, prolonged (6 hours/day, up to 90 days) exposure to NO

(27 mg/m

) and ozone (1.6 mg/m

)

results in a syndrome of progressive pulmonary fibrosis in rats (Rajini et al., 1993), associated with a sustained increase

in procollagen gene expression in the central acini (Farman et al., 1999). This does not occur with either gas alone,

indicating a true synergistic effect. The relevance of this observation for human ambient exposures is not clear, given

the high exposure concentrations used in the study, and absence of evidence for alveolar fibrosis or restrictive lung

disease in epidemiological studies.

The INDEX project Final report

Bermudez et al. (Bermudez et al., 1999) examined DNA strand breaks in BAL cells from rats exposed to ozone (0.6

mg/m

), to NO

(2.3 mg/m

), and the combination. Ozone and the combination exposure increased DNA strand breaks

to a similar degree compared with air exposure, but NO

alone had no effect.

The effects of NO

exposure on SO

-induced bronchoconstriction have been examined, but with inconsistent results.

Jorres and Magnussen (Jorres & Magnussen, 1991) found an increase in airways responsiveness to SO

in asthmatic

subjects following exposure to 0.47 mg/m

for 30 minutes, yet Rubinstein et al. (Rubinstein et al., 1990) found no

change in responsiveness to SO

inhalation following exposure of asthmatics to 0.56 mg/m

for 30 minutes.

Overall, there are little definitive data suggesting that NO

interacts with other pollutants in causing human health

effects. However, human clinical studies have not systematically addressed the effects of pollutant combinations

containing NO

, in part because of the complexity of the experimental design and the difficulty in studying the most

susceptible subjects.

Odour perception

Source: Devos et al. 1990

Odour thresholds: 185 µg/m

Source: Berglund et al., 1993

Nitrogen dioxide has a stinging, suffocating odour. The odour threshold has been placed by various authors at between

94 µg/m

(0.05 ppm) and 410 µg/m

(0.22 ppm). Owing to adaptation, however, no odour was perceived during a

gradual (15-minute) increase in concentration from 0 to 51 mg/m

(0–27 ppm).

Source AIHA, 1989 (Odor thresholds for chemicals with established occupational health standards. American Industrial

Hygiene Association, 1989. p. 25, 70-71)

Minimum perceptible value: 207–263 µg/m

(0.11-0.14 ppm).

Source: Amoore and Hautala, 1983

Odour thresholds: 733 µg/m

(0.39 ppm)

The INDEX project Final report

Summary of Nitrogen dioxide Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Does not exist in water (food), reacting instantaneously to form nitric and nitrous

acids

Toxicokinetics: 70-90% absorbed in the respiratory tract (increased during exercise). Oral breathing increases the

delivery to the lower respiratory tract. Produces oxidation of either membrane lipids or proteins resulting in the loss of

cell permeability control.

Health effect levels of short- and long-term exposure

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

0.47

increase in airway reactivity

sensitive humans

(asthmatics), 1h

Review: California

Air Resources Board

(CARB), 1992

OEHHA 1999

REL: 0.47

0.19 0.38-0.56 increased responsiveness to

bronchoconstrictors

Mild asthmatics, 0.5h WHO 2000

GV: 0.2 (1h)

0.96 possible detrimental respiratory effects in both

normal and asthmatic subjects

Short-term human

clinical studies

Health Canada

1995

ASTER: 0.48

(1h)

Long-term exposure

0.038-

0.056

excess of lower respiratory illness Children 5-12 y old,

annual avg conc. ,

also basing on

background level of

0.015 mg/m³

taken from EHC 188 WHO 2000

GV: 0.04 (1y)

Carcinogenicity: No evidence

Genotoxicity: Limited evidence in vitro and in vivo at high concentrations (50 mg/m³)

Odour threshold: Range: 0.09-0.41 mg/m³ (Berglund), 0.185 mg/m³ (Devos), 0.2 mg/m³ (AIHA). Owing to

adaptation, however, no odour was perceived during a gradual (15-minute) increase in concentration from 0 to 51

mg/m

Susceptible population:

- Asthmatic individuals appear to be the most susceptible members of the population. Exposure of asthmatics to NO

could increase airway responsiveness to a variety of provocative mediators, including cholinergic and histaminergic

chemicals, SO

, and cold air.

- Individuals with chronic obstructive pulmonary disease (COPD: emphysema and chronic bronchitis).

- Children are at increased risk for respiratory illness; of concern because repeated lung infections in children can cause

lung damage later in life (greater lung surface area/body weight ratios and increased minute volumes/weight ratios;

increased exposure of children is finally expected due to higher levels of nitrogen dioxide found nearer to the

ground).

Remarks: High uncertainties in exposure-response relationships for both acute (<3-hour) and long-term exposure; high

uncertainties in establishing an appropriate margin of protection.

The INDEX project Final report

6. Risk characterization

Health hazard evaluation of short- and long-term exposure

[Adopted from WHO, 2000]

Despite the large number of acute controlled exposure studies on humans, several of which used multiple

concentrations, there is no evidence for a clearly defined concentration–response relationship for nitrogen dioxide

exposure. For acute exposures, only very high concentrations (2000 µg/m

) affect healthy people. Asthmatics and

patients with chronic obstructive pulmonary disease are clearly more susceptible to acute changes in lung function,

airway responsiveness and respiratory symptoms. Given the small changes in lung function (< 5% drop in FEV

between air and nitrogen dioxide exposure) and changes in airway responsiveness reported in several studies, 375-565

µg/m

(0.20 to 0.30 ppm) is a clear lowest-observed-effect level. A 50% margin of safety is proposed because of the

reported statistically significant increase in response to a bronchoconstrictor (increased airway responsiveness) with

exposure to 190 µg/m

and a meta-analysis suggesting changes in airway responsiveness below 0.36 mg/m

. (The

significance of the response at 190 µg/m

(0.1 ppm) has been questioned on the basis of an inappropriate statistical

analysis).

On the basis of these human clinical data, a 1-hour exposure limit of 200 µg/m

is proposed. At double this

recommended guideline (400 µg/m

), there is evidence to suggest possible small effects in the pulmonary function of

asthmatics. Should the asthmatic be exposed either simultaneously or sequentially to nitrogen dioxide and an

aeroallergen, the risk of an exaggerated response to the allergen is increased. At 50% of the suggested guideline

(100 µg/m

, 50 ppb), there have been no studies of acute response in 1 hour.

Although there is no particular study or set of studies that clearly support selection of a specific numerical value for an

annual average guideline, the database nevertheless indicates a need to protect the public from chronic nitrogen dioxide

exposures. For example, indoor air studies with a strong nitrogen dioxide source, such as gas stoves, suggest that an

increment of about 30 µg/m

(2-week average) is associated with a 20% increase in lower respiratory illness in children

aged 5–12 years. However, the affected children had a pattern of indoor exposure that included peak exposures higher

than those typically encountered outdoors. Thus the results cannot be readily extrapolated quantitatively to the outdoor

situation. Outdoor epidemiological studies have found qualitative evidence of ambient exposures being associated with

increased respiratory symptoms and lung function decreases in children (most clearly suggestive at annual average

concentrations of 50–75 µg/m

or higher and consistent with findings from indoor studies), although they do not

provide clear exposure–response information for nitrogen dioxide. In these epidemiological studies, nitrogen dioxide

has appeared to be a good indicator of the pollutant mixture. Furthermore, animal toxicological studies show that

prolonged exposures can cause decreases in lung host defences and changes in lung structure.

On these grounds, it is proposed that a long-term guideline for nitrogen dioxide be established. Selecting a well-

supported value based on the studies reviewed has not been possible, but it has been noted that a prior review conducted

for the Environmental Health Criteria document on nitrogen oxides recommended an annual value of 40 µg/m

. In the

absence of support for an alternative value, this figure is recognized as an air quality guideline.

Exposure data of relevance for the European population

Table 6.1: Short-term NO

peak levels related to gas appliances (adapted from Chapter 1.3)

Studies based on NO

sources µg/m

Gas appliances in kitchen, 1h-TWA (Lebret. 1987) 230 – 2055

Gas stoves in kitchen, 1h-TWA (Noy, 1990) 2500

Gas cooking, 1h-TWA (Ross, 1996) 342 – 1585

Homes with gas appliances, 1h-TWA (WHO, 2000) 2000

Unflued gas heaters, ?-TWA (Brown, 2002) 180 – 530

The INDEX project Final report

Table 6.2: Percentage of population exposed beyond given guideline values for NO

Available exposure data

Indoor concentrations (48h-PEM)

WHO recommended value

1-y average

German GV II

1-week average

Description (Study, Year) N 40 µg/m

* 60 µg/m

Helsinki (Expolis, 96-98) 175 10% (15 %) <

Basel (Expolis, 96-98) 50 15% (15 %) 5 %

Prague (Expolis, 96-98) 33 50% (50 %) 25 %

Oxford (Expolis, 98-00) 40 20% (20 %) 10 %

Northen Italy-TWA (Simoni, 2002) 112 75% 45 %

England 2-w TWA (BRE, 97-99) 796 25% 20 %

percentages put in parenthesis result from “through the day” 48h-personal exposure monitorings

* also EU annual limit value for the protection of human health (total of NO

and oxides of nitrogen) to be met the 01/01/2010 (see

margins of tolerance hereafter)

< out of the evaluation range (i.e. <5% of the environments investigated)

Guideline Value II (GV II): GV II is a health-related value based on current toxicological and epidemiological knowledge. If the concentration

corresponding to GV II is reached or exceeded immediate action must be taken because permanent stay in a room at this concentration level is likely

to represent a threat to health, especially for sensitive people. In this context, taking action means an immediate examination of the situation with

regard to a need for control measures. It may include the evacuation of the room in question. If by measurement GV II has been found to be exceeded,

the results should be checked by repetitive measurements carried out immediately under normal conditions of occupancy. If possible and deemed

meaning ful, biological monitoring of the occupants should be carried out in addition.

EU limit values for nitrogen dioxide NO

and oxides of nitrogen NO

– Directive 1999/30/EC

Averaging period Limit value

Margin of tolerance

Date by which limit

value is to be met

1. Hourly limit value

for the protection

of human health

1 hour 200 µg/m

not to be exceeded

more than 18 times a

calendar year

50 % on the entry into force of this

Directive, reducing on 1 January 2001

and every 12 months thereafter by

equal annual percentages to reach 0%

by 1 January 2010

1 January 2010

2. Annual limit value

for the protection of

human health

Calendar year 40 µg/m

50 % on the entry into force of this

Directive, reducing on 1 January 2001

and every 12 months thereafter by

equal annual percentages to reach 0%

by 1 January 2010

1 January 2010

3. Annual limit value

for the protection of

vegetation

Calendar year 30 µg/m

None 19 July 2001

Alert threshold for nitrogen dioxide

400 µg/m3 measured over three consecutive hours at locations representative of air quality over at least 100 km 2 or an entire zone or agglomeration,

whichever is the smaller.

Minimum details to be made available to the public when the alert threshold for nitrogen dioxide is exceedeed

Details to be made available to the public should include at least:

- the date, hour and place of the occurrence and the reasons for the occurrence, where known;

- any forecasts of:

- changes in concentrations (improvement, stabilisation, or deterioration), together with the reasons for those changes,

- the geographical area concerned,

- the duration of the occurrence;

- the type of population potentially sensitive to the occurrence;

- the precautions to be taken by the sensitive population concerned

The INDEX project Final report

Result

Reported maximum nitrogen dioxide (NO

) levels associated with the use of gas appliances in homes (gas cooking and

heating) are in the range 180-2500 µg/m

. Exposure at these levels could generate effects in the pulmonary function of

asthmatics, considered to be the subjects most susceptible to acute NO

exposure, with the lower end of the range

approximating the WHO guideline (200 µg/m

, 1-hour average), established for the protection of asthmatic individuals

and the upper end starting to affect health in normal individuals.

For long-term exposures, increased respiratory symptoms and lung function decreases in children were shown to be the

most sensitive effect in the general population. Measured background levels in European homes indicate that a

remarkable portion of the population is exposed at NO

levels higher than current guideline values protecting from

respiratory effects in children. In up to 25% of the investigated residences (45% in an Italian study) NO

levels

exceeded the German indoor-related guideline value (GV II: 60 µg/m

, 1-week average), what would have resulted in

immediate action i.e. the examination of the situation with regard to a need for control measures.

On the other hand, safe levels in homes, i.e. < 40 µg/m

(following the WHO recommended annual value), are not

likely to be achievable everywhere (e.g. in areas with intense automotive traffic) given that ventilation alone may

introduce outdoor air containing such concentrations.

The INDEX project Final report

Benzene

Synonyms: Annulene, benzine, benzol, benzole, benzol coal naphtha, cyclohexatriene,

mineral naphtha, motor benzol, phenyl hydride, pyrobenzol, pyrobenzole

CAS Registry Number: 71-43-2

Molecular Formula: C

1. Compound identification

Benzene is a naturally occurring colourless liquid at room temperature and ambient pressure, and it has a characteristic

aromatic odour. It is widely used toxic, volatile, flammable organic solvent. It is used as an industrial solvent for

example in paints, varnishes, lacquer thinners and gasoline (1-4%) (IARC, 1989, Holmberg and Lundberg, 1985).

Presently it is used as a raw material in the synthesis of styrene, phenol, cyclohexane, aniline, and alkyl benzenes in the

manufacture of various plastics, resins, and detergents. Syntheses of many pesticides and pharmaceuticals also involve

benzene as a chemical intermediate (HSDB, 2003). The tire industry and shoe factories use benzene extensively in their

manufacturing processes. Benzene causes central nervous system damage acutely and bone marrow damage chronically

and is carcinogenic. It was formerly used as parasiticide (SIS 2003).

2. Physical and Chemical properties

Molecular weight (g/mol) 78.11

Melting point (°C) 5.5

Boiling point (°C) 80.1

Density (g/l at 20 °C, 1 atm) 0.88

Relative vapour density (air =1) 2.8

Solubility: Water (g/l) 1.8

Miscible in most organic solvents.

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 3.242 µg/m

1 µg/m

= 0.308 ppb

Sources: GDCh (1988), RIVM (1988) and ATSDR (1989), Verschueren 2001, HSDB (2003).

3. Indoor Air Exposure assessment

Emission sources

Probably the most important indoor source is cigarette smoking. Other remarkable indoor sources are emissions from

consumer products, including off-gassing from particle board. Similarly, living in the vicinity of hazardous waste sites

or industrial facilities may increase exposure to benzene (ATSDR, 1991).

In a German study, being a subject to environmental tobacco smoke together with driving a car and refuelling were

found significant determinants of exposure to benzene (Hoffman et al 2000). Similar results were reported in the US

TEAM study (Wallace, 1989). It appeared that personal sources (use of products emitting benzene, driving or riding in

automobiles), automobile exhaust and smoking (active and passive) were major sources of benzene to the general

population. Inhalation exposures accounted for more than 99% of the general population exposure to benzene in US

according to Hattemer-Frey et al (1990).

The INDEX project Final report

The average daily intake for an adult in Canada was estimated to be about 200 µg/day in total: 14 µg from ambient air,

140 µg from indoor air, 1.4 µg each from food and drinking water and 50 µg from automobile-related activities. The

corresponding calculated average daily intake in the USA is 320 µg, with a daily intake from ambient and indoor air

ranging between 180 - 1300 µg. Cigarette smoking may add to that as much as 1800 µg/day and passive smoking 50

µg/day. Based on assumptions of spending 2 hours per day in urban ambient air at 7 µg/m

, 21 hours per day in indoor

air at 4 µg/m

, and 1 hour per day inside a vehicle at 50 µg/m

(typical for large roads and heavy traffic), calculated

relative uptakes from urban ambient air, indoor air, air inside cars, and intake from food were 9; 53; 30 and 8%,

respectively (WHO, 1996).

Indoor air and exposure concentrations

Mean indoor concentrations are typically higher than the respective ambient levels all over Europe. The trend showing

increasing ambient concentrations of benzene from northern to southern European cities (Cocheo et al 2000) seems to

be similar for the indoor air concentrations (Jantunen et al 1999). In Northern Europe (Helsinki) the mean indoor

concentration, 2.1 µg/m

, was slightly higher than the respective outdoor concentration, 1.7 µg/m

, if tobacco smoke

was not present indoors (Edwards et al 2001). When tobacco smoke was present in indoor air, indoor concentration of

benzene was more than twofold compared to outdoor level. Personal exposures were typically higher than indoor

concentrations. The highest mean exposure in Helsinki was found due to active smoking being, 2.4 times higher than

the mean indoor concentration in homes without tobacco smoke. In addition to smoking, residential indoor benzene

concentration, time spent in a car, workplace concentration and time spent in home workshop determined personal

exposures to benzene (Edwards et al 2001, Edwards and Jantunen 2001). Similar pattern between mean indoor, outdoor

and personal exposure concentrations could be seen in central and southern European cities, although the concentrations

were higher. Mean indoor concentrations ranged between 2.3 and 12 µg/m

in central European cities such as Basel,

Erfurt, Hamburg, Prague and Antwerp (Jantunen et al 1999, Schneider et al 1999 and 2001, Cocheo et al 2000) and

between 10 and 13 µg/m

in southern cities like Milan and Athens (Jantunen et al 1999, Cocheo et al 2000). The mean

personal exposures to benzene in central Europe ranged from 6 µg/m

in Basel, Switzerland (Jantunen et al 1999) to 14

µg/m

in Germany (Hoffman et al 2000). Carrer et al (2000) studied personal exposures to benzene of 100 office

workers in Milan showing an average exposure of 21.1 µg/m

. The highest reviewed mean population exposure to

benzene in Southern Europe was measured in Murcia being 23 µg/m

(Cocheo et al 2000).

In a German population based study of children and teenagers (GerES IV), weekly mean exposure to benzene was

higher than the mean residential indoor concentration being 2.5 µg/m

and 1.6 µg/m

respectively. The highest exposure

and indoor concentrations were 7.7 µg/m

and 7.0 µg/m

respectively (Ullrich et al 2002).

Based on these results we can conclude, that in addition to outdoor sources of benzene, there are important indoor

sources like smoking, but clearly also emissions from other indoor sources.

Cumulative frequency distributions of the indoor and 48-hour personal exposure concentrations in European studies are

presented in Figure 3.1, Figure 3.2, Figure 3.3 and Figure 3.4.

Source related short time concentrations of benzene are presented in Figure 3.5. Lee and Wang (2004) studied

emissions of incense burning in chamber (18 m

) tests. They found that all 10 tested incense types caused

concentrations exceeding the Recommended Indoor Air Quality Objectives for Office Buildings and Public Places in

Hong Kong (HKIAQO) of 16.1 µg/m

. Maximum benzene concentrations peaked up to 117 µg/m

during incense

burning. Brown (2002) studied short time (30-50 min) concentrations of benzene in established and new buildings in

Australia, reporting mean concentrations of 7 µg/m

and <30 µg/m

respectively. The highest measured indoor

concentration was 81 µg/m

. Emission sources of this high concentration could not be identified, but high indoor

concentrations (16 – 19 µg/m

) in two other houses were related to attached garages. In general, indoor concentrations

of VOCs were much higher in new and recently renovated houses persisting above ‘baseline level’ for several weeks.

Concentration decay rates correlated with molecular volume, indicating the role of diffusion processes for limiting the

emissions. Elke at al (1998) measured short time concentrations of benzene inside buildings (60-min average), inside a

train and a car (30-min average). Slightly higher concentrations were measured during driving a car than when exposed

to tobacco smoke in a smoker compartment of a train or in a smoker day room.

Short time exposure scenarios for potential consumer activities were analysed in European Union Risk Assessment

Report (EU 2003). The highest benzene concentration, 150 – 240 µg/m

were expected to be present in a mainstream

tobacco smoke. Exposure of passive smokers to benzene from smoking was assumed to be 7 µg/m

, exposure from

painting and from car interior accessories 17 µg/m

and 12 µg/m

respectively.

The INDEX project Final report

100

0 10203040506070

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of 30-hour indoor air concentrations of benzene in Athens (Ath, n=42),

Basel (Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=41) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS

2002).

100

0 10203040506070

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ant

Ath

Cop

Mur

Pad

Rou

Figure 3.2. Cumulative frequency distributions of 5-day indoor air concentrations of benzene in Antwerp (Ant), Athens

(Ath), Copenhagen (Cop), Murcia (Mur), Padova (Pad) and Rouen (Rou) (n= 50 in each city) measured in the

MACBETH study (Cocheo et al 2000).

The INDEX project Final report

100

0 5 10 15 20

Indoor concentration (µg/m

)

Cumulative frequency (%)

Benzene 28-day

Figure 3.3. Cumulative frequency distribution of 28-day indoor air concentrations of benzene in England (GM= 3.0

µg/m3, max 93.5 µg/m3, n=796, Brown et al 2002).

100

0 10203040506070

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.4. Cumulative frequency distributions of 48-hour personal exposure concentrations of benzene in Athens

(Ath), Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean concentrations

of the German Survey GerES II (Hoffman et al 2000).

The INDEX project Final report

100

120

140

Incence

burning

New buildings In a car Room with

ETS

Train with

ETS

Concentration (µg/m

)

Benzene

Figure 3.5. Short time benzene concentrations related to specific microenvironments or emission sources. Sampling

times of ‘New buildings’ (Brown 2002), ‘In a car’, ‘Room with ETS’, and ‘Train with ETS’ (Elke at al 1998) were 30-

50 min, 30-min, 60-min, 30-min respectively. During incense burning grab samples were obtained (Lee and Wang

2004).

4. Toxicokinetics

Absorption

Benzene is readily absorbed from oral and inhalation exposures. Less than 1% is absorbed through the skin. Pekari et al.

(1992) studied respiratory absorption of benzene in three males exposed to 5 and 32 mg/m

(1.7 and 10 ppm) benzene

for 4 hours each. Absorption of benzene was determined by measuring differences in concentration between inhaled and

exhaled air. The average absorption was 52% at the low concentration and 48% at the high concentration.

Most human respiratory absorption studies were conducted at concentrations probably conducing to saturation of the

benzene metabolism. The observed decline in absorption with increasing exposure time is apparently due to respiratory

excretion of unmetabolized benzene. Animal studies indicate that the metabolism of benzene is saturated at exposure

concentrations in excess of 32 mg/m

. Because absorption at low dose levels is most relevant for derivation of exposure

limits, the estimates of 48–52% retention derived by Pekari et al. (1992) are the most relevant values for evaluating

absorption in humans. Absorption of benzene through the oral route is complete (100%).

Distribution

Results from test animal studies indicate that absorbed benzene is widely distributed throughout the body,

independently of the route of the route of exposure. The parent compound is preferentially stored in the fat, although the

relative uptake in tissues also appears to be dependent on the perfusion rate of tissues by blood. Steady-state benzene

concentrations in rats exposed via inhalation to 1600 mg/m3 (500 ppm) for 6 hours were blood, 1.2 mg%; bone marrow,

3.8 mg%; and fat, 16.4 mg% (Rickert et al., 1979). Benzene was also found in the kidney, lung, liver, brain, and spleen.

Levels of the benzene metabolites phenol, catechol, and hydroquinone were higher in the bone marrow than in blood,

with phenol being eliminated more rapidly than catechol or hydroquinone after exposure. Benzene also has been shown

to cross the human placenta and has been found in the cord blood in amounts equal to or greater than those in maternal

blood (Dowty et al. 1976).

Metabolism and elimination

Despite extensive research, the metabolism of benzene is still not thoroughly understood. It is generally accepted that

benzene itself is not directly responsible for causing the toxic effects; however, the metabolic products responsible for

the noncancer and carcinogenic effects of benzene exposure have not been clearly defined. The evidence suggests that

The INDEX project Final report

several metabolites, as well as interactions between these metabolites, may be necessary to explain the toxic effects of

benzene.

Reviews of benzene metabolism are given in ATSDR, 1997; Snyder et al., 1993b; Snyder and Hedli, 1996; Ross, 1996;

Witz et al.,1996 (see also Figure 4.1). The liver is the major site of benzene metabolism. Benzene is converted to a

number of metabolites (phenol, catechol, hydroquinone, 1,2,4-benzenetriol, benzoquinone and trans,trans-

muconaldehyde) by the cytochrome P-450 dependent mixed-function oxidase system. The P-450 enzyme CYP 2E1

appears to exhibit the greatest affinity for benzene and is the most active in benzene metabolism. CYP 2B1 is also

capable of hydroxylating benzene, but contributes to the metabolism only at higher concentrations. Oxidative

metabolism by CYP 2E1 is required for manifestation of the haematotoxic and genotoxic effects of benzene as

confirmed recently by studies on the benzene in vivo metabolism in transgenic CYP 2E1 knockout mice (Valentine et

al. 1996).

In the attempt to explain bone marrow toxicity the hypothesis is brought forward that phenol, catechol, and

hydroquinone are transported to the bone marrow and converted by peroxidase-mediated reactions to electrophilic

compounds. For example, in the bone marrow, phenol is rapidly oxidized by the marrow peroxidases to reactive,

protein-binding species such as biphenols and biphenoquinones (Sawahata et al. 1985). Hydroquinone also is oxidized

in the marrow via a peroxidase-catalyzed reaction to an extremely short-lived semiquinone radical intermediate. This

forms p-benzoquinone, a reactive metabolite, known to cause DNA damage and cytotoxic effects in benzene-exposed

animals (Smith et al. 1989). However, trans-trans-muconaldehyde may also play a role (Snyder et al. 1989).

Pharmacokinetic models have been developed for rats and mice and used for risk assessments based on animal cancer

data (WHO, 1993; Cox and Ricci, 1992). Physiologically based toxicokinetic models have also been developed using

human data (WHO, 1993; Watanabe et al., 1994), but so far these have not been used for risk assessment.

The metabolism of benzene can be inhibited by toluene, leading to decreased toxicity. On the other hand, ethanol can

increase the metabolism of benzene, primarily by inducing xenobiotic metabolizing enzymes (Medinsky at al., 1994).

The average half-time of benzene in humans is 28 hours (WHO, 1987).

Following inhalation exposure to benzene, the major route of elimination of unmetabolized benzene in humans and

animals is in exhaled air. Most benzene is metabolized and the metabolites are excreted after phase-II-conjugation

predominantly in the urine. Phenolic metabolites are conjugated with sulfate or glucuronic acid (Henderson et al. 1989).

Small amounts of the glucuronides may enter the bile and are found in the feces.

5. Health effects

Effects of short-term exposure

The acute toxicity of benzene is low. In general, acute symptoms are dependent on both the concentration and duration

of exposure. Exposure to 24 g/m

for 30 min is life-threatening; 5 g/m

for 60 min produces significant symptoms; 150-

500 mg/m

for 5 hr results in headache and weakness; whereas exposure to 80 mg/m

or less for 8 hr results in no

demonstrable acute effect (Irons, 1992). Exposure at the odor threshold (2.8 mg/m³) for a brief duration has been

reported to enhance the electropotential of the brain (Haley, 1977).

Major concerns of systemic benzene toxicity include aplastic anemia and acute myelogenous leukemia (IARC, 1987;

Reprotext, 1993). Both of these conditions are typically seen in the chronic and subchronic exposures, but may be of

concern following acute exposures as well. Myeloid and erythroid components of the bone-marrow are specific targets

of benzene toxicity, which leads to aplastic anemia (IARC, 1982).

The long elimination half-life for benzene (13.1 hours for the longer phase of a 2-compartment model) is due to the

formation of catechol, quinone, and hydroquinone in the bone marrow. These reactive metabolites are not readily

excreted, and are cytotoxic to the germinal cells in the bone marrow (Greenlee et al., 1981). A 3-compartment model

was fitted to human data on benzene disposition and bonemarrow metabolism (Watanabe et al., 1994). The general

relationship between cumulative quantity of metabolites produced and inhalation concentration was not linear, but was

S-shaped, inflecting upward at low concentrations, and saturating at high concentrations.

The INDEX project Final report

Figure 4.1: Benzene Metabolism (taken from OEHHA, 2001)

Male C57BL/6J mice (7-Wgroup) were exposed to benzene 0, 33, 99, 319 or 961 mg/m

(0, 10.2, 31, 100 or 301 ppm)

in wholebody dynamic inhalation chambers for 6 hours/day for 6 consecutive days (Rozen et al. 1984). Control mice

were exposed to filtered, conditioned air only. Erythrocyte counts were depressed in C57BL/6 mice only at 319 and 961

mg/m

. At exposures of 33 and 99 mg/m

, respectively, depressions of the proliferative responses of bone-marrow-

derived B-cells and splenic T-cells occurred in C57BL/6J mice without causing a concomitant depression in number of

splenic T- or B- cells. Peripheral lymphocyte counts were depressed at all levels. No correlation was made of reduced

lymphocytes with general effects. These results demonstrate that short-term inhaled benzene even at low exposure

concentrations can cause reduction in certain immune associated processes.

Results from subacute exposures further illustrate the hematotoxic effects of benzene and the potential for

immunotoxicity. Inhalation of 334 mg/m³ (103 ppm) benzene for 6 hours/day for 7 days by mice caused decreased

spleen and marrow cellularities and decreased spleen weights (Green et al., 1981). Benzene inhalation at concentrations

of 0, 32, 96, 319, and 958 mg/m³ (0, 10, 30, 100, and 300 ppm) for 6 hours/day for 5 days resulted in a decreased host-

resistance to bacterial infection by Lysteria monocytogenes (Rosenthal and Snyder, 1985). The numbers of L.

monocytogenes bacteria isolated from the spleen were increased in a dose-dependent manner on day 4 of infection. The

total numbers of T- and B-lymphocytes in the spleen and the proliferative ability of the splenic lymphocytes were

decreased in a dose-dependent manner by benzene exposures of 96 mg/m³ or greater. In this study, no decrement in host

resistance or immune response was observed at 32 mg/m³ benzene. Later studies in mice have also shown that exposure

to 32 mg/m³ for a subacute duration does not significantly alter hematological parameters in blood, spleen, thymus, or

bone marrow (Farris et al., 1996; 1997). Farris et al. (1997) reported the hematological consequences of benzene

inhalation in B6C3F1 mice exposed to 3.2 , 16, 32, 319 and 639 mg/m³ (1, 5, 10, 100, and 200 ppm) benzene for 6

hr/day, 5 days/week for 1, 2, 4, or 8 weeks and a recovery group. There were no significant effects on hematopoietic

parameters from exposure to 10 ppm benzene or less. Thus 32 mg/m³ was a NOAEL for 1 week of exposure (and

longer). Exposure to 319 and 639 mg/m³ benzene reduced the number of total bone marrow cells, progenitor cells,

differentiating hematopoietic cells, and most blood parameters. Replication of primitive progenitor cells in the bone

marrow was increased during the exposure period as a compensation for the cytotoxicity. At 639 mg/m³, the primitive

progenitor cells maintained an increased percentage of cells in S-phase through 25 days of recovery compared with

controls.

The INDEX project Final report

100

In addition to myelotoxicity, acute exposure to benzene may disrupt erythropoiesis and result in genotoxicity.

Erythropoiesis, as measured by uptake of radiolabeled iron in the bone-marrow, has been shown to be inhibited by

subcutaneous injection of 10 mmol/kg benzene in mice (Bolcsak and Nerland, 1983). Inhalation of 9.6 mg/m³ benzene

for 6 hours by rats resulted in a significant increase over controls in the frequency of sister chromatid exchanges in

peripheral blood lymphocytes (Erexson et al., 1986).

Reproductive and developmental effects have been reported following benzene exposure. Coate et al. (1984) exposed

groups of 40 female rats to 3.2 , 32, 128, 319 and 639 mg/m³ (0, 1, 10, 40, and 100 ppm) benzene for 6 hours/day

during days 6-15 of gestation. In this study, teratologic evaluations and fetotoxic measurements were done on the

fetuses. A significant decrease was noted in the body weights of fetuses from dams exposed to 319 mg/m³. No effects

were observed at a concentration of 128 mg/m³.

Keller and Snyder (1986) reported that exposure of pregnant mice to concentrations as low as 16 mg/m³ (5 ppm)

benzene on days 6-15 of gestation (6 hr/day) resulted in bone-marrow hematopoietic changes in the offspring that

persisted into adulthood. However, the hematopoietic effects (e.g., bimodal changes in erythroid colony-forming cells)

in the above study were of uncertain clinical significance.

Keller and Snyder (1988) exposed Swiss Webster mice to benzene in utero from day six through day 15 of gestation.

Pregnant mice were administered benzene via inhalation at doses of 5, 10 or 20 ppm. On day 16 of gestation, two days

after birth and six weeks after birth, offspring were tested for levels of blood haemoglobin and levels of hematopoietic

cells in the blood, bone marrow and liver. No adverse effects were observed for fetuses at day 16 of gestation. However,

two days after birth, offspring exhibited reduced numbers of circulating erythroid precursor cells (all doses), increased

numbers of hematopoietic blast cells in the liver (20 ppm only), increased granulopoietic precursor cells (20 ppm only)

and decreased numbers of erythropoietic precursor cells (20 ppm only). Mice examined six weeks after birth also

demonstrated increased granulopoiesis in the 20 ppm dose group.

An exposure of 1600 mg/m³ (500 ppm) benzene through days 6-15 of gestation was teratogenic in rats while 160 mg/m³

resulted in reduced fetal weights on day 20 of gestation. No fetal effects were noted at an exposure of 32 mg/m³ (Kuna

and Kapp, 1981). An earlier study by Murray et al. (1979) showed that inhalation of 1600 mg/m³ benzene for 7

hours/day on days 6-15 and days 6-18 of gestation in mice and rabbits, respectively, induced minor skeletal variations.

Tatrai et al. (1980) demonstrated decreased fetal body weights and elevated liver weights in rats exposed throughout

gestation to 150 mg/m³ (47 ppm).

A recent study did not find any increased risk of spontaneous abortion among the wives of 823 male workers

occupationally exposed to benzene levels of less than or around 15 mg/m

(Strucker et al., 1994).

An attempt by Nielsen and Alarie (1982) to determine the inhalation RD

(the concentration of a compound inducing a

40% decrease in respiratory rate in male Swiss-Webster mice) for benzene was not successful. These investigators

showed that inhalation of 18,5 g/m³ (5800 ppm) benzene in mice caused an increase in respiratory rate beginning at 5

minutes following onset of exposure. They speculated that the stimulation of respiratory rate resulted from the action of

benzene on the central nervous system. In this study, benzene was not irritating to the upper airways of the animals.

Summary of short-term exposure effect levels (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

128

EXP

319

EXP

Developmental: decreased fetal body weights

(severe adverse effect!)

Rats, 6 hours per day

during days 5-16 of

gestation

Coate et al., 1984;

(supported by Kuna

and Kapp, 1981;

Keller and Snyder,

1988)

OEHHA 1999

REL: 1.3

EXP

Developmental: Altered hematopoiesis in

offspring

(high degree of uncertainty!)

Mice, 6 hours per day

during days 5-16 of

gestation

Keller and Snyder,

1988

OEHHA 2001

MADL

inhal.

: 1.3

µg/day

EXP

ADJ

HEC

Haematological: depressed peripheral

lymphocytes and mitogen-induced blastogenesis

of femoral B-lymphocytes

(less Serious LOAEL) .

Mice, 6 hours per day

(for 6 consecutive

days)

Rozen et al. 1984

ATSDR 1997

MRL: 0.16

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

The INDEX project Final report

101

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration; MADL

inhal.

maximum allowable daily level (OEHHA)

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of Acute Reference Exposure Level (for a 6-hour exposure) (Level Protective Against Severe Adverse

Effects):

Study Coate et al., 1984; (supported by Kuna and Kapp, 1981; Keller

and Snyder, 1988)

Study population pregnant female rats

Exposure method inhalation of 0, 1, 10, 40, or 100 ppm

Critical effects decreased fetal body weights

LOAEL 100 ppm

NOAEL 40 ppm

Exposure duration 6 hours per day (for 5 days)

LOAEL uncertainty factor 1

Interspecies uncertainty factor 10

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 100

Reference Exposure Level 0.4 ppm (1.3 mg/m³; 1,300 µg/m³)

While benzene exposure results in decreased immune response and hematopoietic effects in laboratory animals

following 5 day exposures, it was problematic to extrapolate from these repeated dose studies for these endpoints. Thus,

no level protective against mild adverse effects for one-hour is being recommended. The REL is based on

developmental toxicity, a severe adverse effect.

Pregnant female rats (40 per group) were exposed for 6 hours/day on days 6-15 of gestation to benzene concentrations

of 0, 1, 10, 40, and 100 ppm (0, 3.24, 32.4, 129.6, and 324 mg/m³) (Coate et al., 1984). The mean fetal weights from the

females treated with 100 ppm benzene were significantly decreased (p < 0.05) compared to controls. No teratogenic,

fetotoxic, or maternally toxic effects were observed in rats exposed to 40 ppm (129.6 mg/m³) benzene or less. The 40

ppm (129.6 mg/m³) concentration is considered a NOAEL for reduced fetal weight. The value of 40 ppm for a 6-hour

exposure was extrapolated to a 1-hour exposure using the equation Cn * T = k, where n = 2. The resulting 100 ppm

extrapolated value was used to determine the level protective against severe adverse effects using uncertainty factors of

10 for intraspecies and 10 for interspecies variation. The level protective against severe adverse effects for benzene is

therefore 1.0 ppm or 3.24 mg/m³.

Kuna and Kapp (1981) found direct teratogenic effects measured as decreased crown-rump length, exencephaly, and

angulated ribs in rats when pregnant females were exposed 6 hours/day during days 6-15 of gestation to a concentration

of 500 ppm. In this study, a concentration of 50 ppm during gestation resulted in lower fetal weights measured on day

20 of gestation. No fetal effects were noted at an exposure of 10 ppm (32 mg/m³). Keller and Snyder (1988) reported a

NOAEL of 10 ppm for developmental hematopoietic effects in mice. The highest reported NOAEL (i.e., 40 ppm)

consistent with reported LOAEL values was chosen for the derivation of the Reference Exposure Level (severe adverse

effect level, in this case) for benzene.

OEHHA (2001)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

From all the available animal studies the most sensitive study was a developmental toxicity study in mice using

endpoints reflecting altered hematopoiesis (blood cell formation) (Keller and Snyder, 1988). Exposures were by

inhalation at 0, 5, 10 and 20 ppm benzene concentrations in air, 6 h/day, during gestation days 6 to 15. The LOEL for

developmental hematopoietic effects was 5 ppm based on a significantly reduced relative number of early nucleated red

cells in a sample of nucleated cells from the peripheral blood of 2-day old neonates that had been exposed to benzene in

utero. Because the Keller and Snyder study provided a LOEL rather than a NOEL, a NOEL for purposes of assessment

was calculated by dividing the LOEL of 5 ppm by 10, resulting in a NOEL of 0.5 ppm.

The following calculations were performed to derive the Maximum Allowable Daily Level (MADL) for benzene:

Conversion of air concentration in ppm to mg/m

using a molecular weight for benzene of 78.11 daltons and a partial

molar volume of 24.45 at 25°C (=1.6 mg/m

). Conversion of air concentration for 6 hour (h) exposure to a 24 h day

The INDEX project Final report

102

(=0.40 mg/m

). Calculation of NOEL dose for 30 g mouse with an inhalation rate of 0.063 m

/day (Bond et al. 1986,

Depledge 1985) (=0.84 mg/kg/day). Calculation of NOEL dose for a 58 kg woman (=49 mg/day).

The MADL is derived by dividing the NOEL by one thousand. Thus, the adjusted NOEL was divided by 1000 to obtain

the MADL

inhalation

= 49 µg/day

ATSDR (1997)

Agency for Toxic Substances and Disease Registry

Derivation of the Minimal Risk Level (MRL):

ATSDR derived an acute-duration inhalation MRL of 0.05 ppm for benzene based on a LOAEL of 10 ppm for

immunotoxicity (reduced lymphocyte proliferation) following mitogen stimulation in mice (Rozen et al. 1984). The

animal LOAEL was converted to a human equivalent concentration (LOAEL

HEC

) of 14.7 ppm and divided by an

uncertainty factor of 300 (10 for the use of a LOAEL, 3 for extrapolation from animals to humans, and 10 for human

variability) to yield the MRL. The mice were exposed 6 hours/day for 6 days.

Effects noted in the study (Rozen et al. 1984) and corresponding doses: 10.2 ppm: No adverse effect on erythrocytes,

depressed peripheral lymphocytes and mitogen-induced blastogenesis of femoral B-lymphocytes (Less Serious

LOAEL). 31 ppm: No adverse effect on erythrocytes, depression of mitogen-induced blastogenesis of splenic T-cells.

100 ppm: Depressed erythrocyte counts. Dose end point used for MRL derivation: 10.2 ppm. Uncertainty factors used

in MRL derivation: 300, 10 for use of a LOAEL, 3 for extrapolation from animals to humans, 10 for human variability.

The ratio of the blood/gas partition coefficients was assumed to be 1. The concentration was adjusted for intermittent

exposure by multiplying the LOAEL (10.2 ppm) by 6/24 to correct for less than a full day of exposure. The resulting

number (adjusted LOAEL) is 2.50 ppm.

The human equivalent concentration (HEC) was calculated using Formula 4-l1, from Interim Methods for Development

of Inhalation Reference Concentrations, EPA 1990h:

LOAEL

[HEC]

(mg/m

or ppm) = LOAEL [ADJ] (mg/m

or ppm) x (V

/BW)

/ (V

/BW)

where: LOAEL

[HEC]

= the LOAEL human equivalent concentration

LOAEL

[ADJ]

= the LOAEL adjusted for duration (see above)

/BW)

/ (V

/BW)

= the ratio of the alveolar ventilation rate (mL/min or L/hr) divided by Body weight (kg) of the

animal species to the same parameters for humans.

Values for this ratio were taken from EPA 1988e, as follows: (V

)

= ventilation rate for male B6C3F1 mice,

subchronic, 0.053 m

/day x 1,000 L/m

x 1 day/24 hr= 2.208 L/hr = 2.21 L/hr. (BW)

= body weight for male B6C3F1

mice, subchronic, 0.0316 kg (V

)

= ventilation rate for human adult male, 20 m

/day x 1,000 L/m

x 1 day/24 hr=

833.33 L/hr = 833 L/hr. (BW)

= body weight for human adult male, 70 kg. Resulting in a LOAEL

[HEC]

given by: 2.50

ppm x (2.21 L/hr / 0.0316 kg ) / (833 L/hr / 70 kg ) = 14.7 ppm (47 mg/m

Other additional studies or pertinent information that lend support to this MRL: Increased number of MN-PCEs,

decreased numbers of granulopoietic stem cells (Toft et al. 1982) lymphopenia (Cronkite et al. 1985), lymphocyte

depression, and increased susceptibility to bacterial infection (Rosenthal and Snyder 198.5) are among the adverse

hematological and immunological effects observed in several other acute-duration inhalation studies. The study by

Rozen et al. (1984) shows benzene immunotoxicity (reduced mitogen-induced lymphocyte proliferation) at a slightly

lower exposure level than these other studies. C57BI/6J mice were exposed to 0, 10.2, 31, 100, and 301 ppm benzene

for 6 days at 6 hrs/day. Control mice were exposed to filtered, conditioned air only. Lymphocyte counts were depressed

at all exposure levels while erythrocyte counts were elevated at 10.2 ppm, equal to controls at 31 ppm and depressed at

100 and 301 ppm. Femoral B-lymphocyte and splenic B-lymphocyte numbers were reduced at 100 ppm. Levels of

circulating lymphocytes and mitogeninduced blastogenesis of femoral B-lymphocytes were depressed after exposure to

10.2 ppm benzene for 6 days. Mitogen-induced blastogeneses of splenic T-lymphocytes were depressed after exposure

to 31 ppm of benzene for 6 days. In another study, mice exhibited a 50% decrease in the population of erythroid

progenitor cells (CFU-E) after exposure to 10 ppm benzene for 5 days, 6 hours per day (Dempster and Snyder 1991). In

a study by Wells and Nerland (1991), groups of 4-5 male Swiss-Webster mice were exposed to 0, 3, 25, 55, 105, 199,

303, 527, 1,150, or 2,290 ppm benzene for 6 hours/day for 5 days. The number of leukocytes in peripheral blood and

spleen weights were significantly decreased compared with untreated controls at all concentrations ¡Ý25 ppm.

Therefore, 3 ppm was the NOAEL and 25 ppm was the LOAEL for these effects. Other endpoints were not monitored

in this study. These data support the choice of Rozen et al. (1984) as a critical study.

The INDEX project Final report

103

Effects of long-term exposure (noncancer)

The primary toxicological effects of chronic benzene exposure are on the hematopoeitic system. Neurological and

reproductive/developmental toxic effects are also of concern at slightly higher concentrations. Impairment of immune

function and/or various anemias may result from the hematotoxicity. The hematologic lesions in the bone marrow can

lead to peripheral lymphocytopenia and/or pancytopenia following chronic exposure. Severe benzene exposures can

also lead to life-threatening aplastic anemia. These lesions may lead to the development of leukemia years after

apparent recovery from the hematologic damage (DeGowin, 1963).

Several types of blood dyscrasia, including pancytopenia, aplastic anaemia, thrombocytopenia, granulocytopenia and

lymphocytopenia have been noted following exposure to benzene. As in experimental animals, the primary target organ

for benzene that results in haematological changes is the bone marrow. It has been suggested that the cells at highest

risk are the rapidly proliferating stem cells (WHO, 1993). Most of these cases were reported several years ago, when

benzene was used as a solvent in different workplaces. An increased frequency of anaemia was detected among shoe

workers, rotogravure workers and rubber factory workers with prolonged exposure to high benzene concentrations

(hundreds of milligrams of benzene per m

) (WHO, 1993).

Kipen et al. (1988) performed a retrospective longitudinal study on a cohort of 459 rubber workers, examining the

correlation of average benzene exposure with total white blood cell counts taken from the workers. These researchers

found a significant (p < 0.016) negative correlation between average benzene concentrations in the workplace and white

blood cell counts in workers from the years 1940-1948. A reanalysis of these data by Cody et al. (1993) showed

significant decreases in RBC and WBC counts among a group of 161 workers during the 1946-1949 period compared

with their pre-exposure blood cell counts. The decline in blood counts was measured over the course of 12 months

following start of exposure. During the course of employment, workers who had low monthly blood cell counts were

transferred to other areas with lower benzene exposures, thus potentially creating a bias towards non-significance or

removing sensitive subjects from the study population. Since there was a reported 75% rate of job change within the

first year of employment, this bias could be highly significant. In addition, there was some indication of blood

transfusions used to treat some “anemic” workers, which would cause serious problems in interpreting the RBC data,

since RBCs have a long lifespan in the bloodstream. The exposure analysis in this study was performed by Crump and

Allen (1984). The range of monthly median exposures was 30-54 ppm throughout the 12-month segment examined.

Despite the above-mentioned potential biases, workers exposed above the median concentrations displayed significantly

decreased WBC and RBC counts compared with workers exposed to the lower concentrations using a repeated

measures analysis of variance.

Tsai et al. (1983) examined the mortality from all cancers and leukemia, in addition to hematologic parameters in male

workers exposed to benzene for 1-21 years in a refinery from 1952-1978. The cohort of 454 included maintenance

workers and utility men and laborers assigned to benzene units on a “regular basis”. Exposures to benzene were

determined using personal monitors; the median air concentration was 0.53 ppm in the work areas of greatest exposure

to benzene. The average length of employment in the cohort was 7.4 years. The analysis of overall mortality in this

population revealed no significant excesses. Mortality from all causes and from diseases of the circulatory system was

significantly below expected values based on comparable groups of U.S. males. The authors concluded the presence of

a healthy worker effect. An internal comparison group of 823 people, including 10% of the workers who were

employed in the same plant in operations not related to benzene, showed relative risks for 0.90 and 1.31 for all causes

and cancer at all sites, respectively (p < 0.28 and 0.23). A subset of 303 workers was followed for medical surveillance.

Up to four hematological tests per year were conducted on these workers. Total and differential white blood cell counts,

haemoglobin, hematocrit, red blood cells, platelets and clotting times were found to be within normal (between 5% and

95% percentile) limits in this group. Although the exposure duration averaged only 7.4 years, the study was considered

to be chronic since 32% of the workers had been exposed for more than 10 years.

Collins et al. (1997) used routine data from Monsanto’s medical/industrial hygiene system to study 387 workers with

daily 8-hour time-weighted exposures (TWA) averaging 0.55 ppm benzene (range = 0.01 – 87.69 ppm; based on 4213

personal monitoring samples, less than 5% of which exceeded 2 ppm). Controls were 553 unexposed workers. There

was no increase in the prevalence of lymphopenia, an early, sensitive indicator of benzene toxicity, among exposed

workers (odds ratio = 0.6; 95% confidence interval = 0.2 to 1.8), taking into account smoking, age, and sex. There also

was no increase in risk among workers exposed 5 or more years (odds ratio = 0.6; 95% confidence interval = 0.2 to 1.9).

There were no differences between exposed and unexposed workers for other measures of hematotoxicity, including

mean corpuscular volume and counts of total white blood cells, red blood cells, haemoglobin, and platelets.

Rothman et al. (1996) compared hematologic outcomes in a cross-sectional study of 44 male and female workers

heavily exposed to benzene (median = 31 ppm as an 8-hr TWA) and 44 age and gender-matched unexposed controls

from China. Hematologic parameters (total WBC, absolute lymphocyte count, platelets, red blood cells, and hematocrit)

The INDEX project Final report

104

were decreased among exposed workers compared to controls; an exception was the red blood cell mean corpuscular

volume (MCV), which was higher among exposed subjects. In a subgroup of 11 workers with a median 8 hr TWA of

7.6 ppm (range = 1-20 ppm) and not exposed to more than 31 ppm on any of 5 sampling days, only the absolute

lymphocyte count (ALC) was significantly different between exposed workers and controls (p = 0.03). Among exposed

subjects, a dose response relationship with various measures of current benzene exposure (i.e., personal air monitoring,

benzene metabolites in urine) was present only for the total WBC count, the absolute lymphocyte count, and the MCV.

Their results support the use of the absolute lymphocyte count as the most sensitive indicator of benzene-induced

hematotoxicity.

Ward et al. (1985) exposed male and female CD-1 mice and Sprague-Dawley rats to 0, 1, 10, 30 or 300 ppm (0, 3.2, 32,

96 or 960 mg/m3) benzene, 6 hours/day, 5 days/week for 91 days and measured various hematological endpoints. The

study identified both a LOAEL of 300 ppm and a NOAEL of 30 ppm. The male mouse appeared to be the most

sensitive sex/species in this study. The exposure-response relationships for the different hematological endpoints for the

male mouse were modeled using a BMD modeling approach and decreased hematocrit (i.e., volume percentage of

erythrocytes in whole blood) was chosen as the critical effect.

An examination of 32 patients, who were chronically exposed to benzene vapors ranging from 150 to 650 ppm for 4

months to 15 years, showed that pancytopenia occurred in 28 cases. Bone marrow punctures revealed variable

hematopoeitic lesions, ranging from acellularity to hypercellularity (Aksoy et al., 1972).

In an evaluation of the literature, a WHO Task Group (WHO, 1993) drew the conclusion that bone marrow depression

or anaemia would not be expected to occur in workers exposed for 10 years to 3.2 mg/m

(1 ppm) or less.

Intermediate-duration inhalation and oral exposures to benzene induced neurological effects in animals that included

reduced limb grip strength, behavioral disturbances, and changes in brain levels of monoamine transmitters and

acetylcholinesterase (Dempster et al. 1984; Frantik et al. 1994; Hsieh et al. 1988; Li et al. 1992).

Older studies on workers exposed to benzene, toluene and xylene have shown decreased levels of agglutinins and IgG

and IgA immunoglobulins, and increased levels of IgM.

A loss of leukocytes has been observed in several studies in highly exposed workers as well as reduced numbers of T

lymphocytes. In a study of refinery workers exposed to benzene concentrations of less than 32 mg/m

, no such

immunological effects were seen (WHO, 1993).

Summary of long-term exposure effect levels (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

1.7

Study

0.60

ADJ

Haematological effects (total and differential

white blood cell counts, haemoglobin,

hematocrit, red blood cells, platelets and clotting

times)

Occupational, 7.4y Tsai et al. 1983

OEHHA 1999

REL: 0.06

BMC

BMCL

8.2

ADJ-BMCL

Decreased lymphocyte counts Occupational, 6.3y

(0.7-16y)

Rothman et al. 1996 EPA-IRIS 2003

RfC: 0.03

(3.2) 32

EXP

Blood cell depression (lymphocytes) Mouse, 6-178d

NOAEC has been

derived, not observed

Dempster and Snyder,

1990, Rozen et al.

1984, Baarson et al.,

1984, Green et al.,

1981a,b, and other

studies

ECB 2003

2.5

EXP

1.05

HEC

Neurological: increased (rapid response time)

(Less Serious minimal LOAEL).

Mice, 30d

(2h/d, 6d/w)

Li et al. 1992

intermediate duration

study !

ATSDR 1997

MRL

INT

: 0.013

(intermediate!)

The INDEX project Final report

105

OEHHA (1999)

Office of Environmental Health Hazard Assessment

Derivation of Chronic Reference Exposure Level (REL)

Study Tsai et al. (1983)

Study population 303 Male refinery workers

Exposure method Occupational exposures for 1-21 years

Critical effects Hematological effects

LOAEL Not observed

NOAEL 0.53 ppm

Exposure continuity 8 hr/day (10 m

3 per 20 m3 day), 5 days/week

Exposure duration 7.4 years average (for the full cohort of 454); 32% of the workers were

exposed for more than 10 years

Average occupational exposure 0.19 ppm

Human equivalent concentration 0.19 ppm

LOAEL uncertainty factor 1

Subchronic uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 10

Inhalation reference exposure level 0.02 ppm (20 ppb; 0.06 mg/m

3; 60 µg/m3)

Both the animal and human databases for benzene are excellent. The human data presented by Tsai et al. (1983) and

associates were selected over animal studies because the collective human data were considered adequate in terms of

sample size, exposure duration, and health effects evaluation. Although the study by Tsai et al. (1983) is a free-standing

NOAEL, the endpoint examined is a known sensitive measure of benzene toxicity in humans. In addition, the LOAEL

for the same endpoint in workers reported by Cody et al. (1993) help form a dose-response relationship and also yield

an REL which is consistent with that derived from Tsai et al. (1983). The study by Cody et al. (1993), since it failed to

identify a NOAEL and was only for a period of 1 year, contained a greater degree of uncertainty in extrapolation to a

chronic community Reference Exposure Level. The recent results of Collins et al. (1997) that included a NOAEL of

0.55 ppm and of Rothman et al. (1996) that included a LOAEL of 7.6 ppm are consistent with those of Tsai et al.

Therefore the study by Tsai et al. (1983) was used as the basis for the chronic REL for benzene.

In the Cody et al. (1993) study, significant hematological effects, including reduced RBC and WBC counts, were

observed in 161 male rubber workers exposed to median peak concentrations (i.e., only the peak concentrations for any

given exposure time were reported) of 30-54 ppm or more for a 12-month period during 1948. The 30 ppm value was

considered a 1-year LOAEL for hematological effects. The 30 ppm value is the low end of the range of median values

(30-54 ppm) reported by Crump and used in the Kipen et al. (1988) and Cody et al. (1993) studies. An equivalent

continuous exposure of 10.7 ppm can be calculated by assuming that workers inhaled 10 m

3 of their total 20 m3 of air

per day during their work-shift, and by adjusting for a normal 5 day work week. Application of uncertainty factors for

subchronic exposures, estimation of a NOAEL, and for protection of sensitive subpopulations (10 for each) results in an

REL of 0.01 ppm (10 ppb; 30 µg/m

3). This is comparable to the REL based on Tsai et al. (1983).

Ward et al. (1996) determined a relationship between occupational exposures to benzene and decreased red and white

cell counts. A modeled dose-response relationship indicated a possibility for hematologic effects at concentrations

below 5 ppm. However, no specific measures of the actual effects at concentrations below 2 ppm were taken, and the

Tsai et al. (1983) data were not considered in their analysis. The purpose of this study was to characterize the trend for

effects at low concentrations of benzene. A NOAEL or LOAEL was not identified in the study. The selection of a

NOAEL of 0.53 ppm is therefore not inconsistent with the results of the Ward et al. (1996) study.

For comparison with the REL of 20 ppb based on human data, OEHHA estimated a REL based on the chronic

inhalation study in mice by Baarson et al. (1984), which showed that bone-marrow progenitor cells were markedly

suppressed after intermittent exposures (6 hr/day, 5 days/week) to 10 ppm benzene for 6 months. An extrapolation of

this value to an equivalent continuous exposure resulted in a concentration of 1.8 ppm. Application of an RGDR of 1

for a systemic effect and uncertainty factors of 3 and 10 for inter- and intraspecies variability, and 10 for estimation of a

NOAEL from the LOAEL would result in an REL of 6 ppb (20 µg/m

3). The Farris et al. (1997) 8 week study indicated

a LOAEL of 100 ppm and a NOAEL of 10 ppm for hematological effects. Application of an RGDR of 1 and UFs of 10

for subchronic, 3 for interspecies and 10 for intraspecies extrapolation (total UF = 300) also resulted in an estimated

REL of 6 ppb, in reasonable agreement with the proposed REL of 20 ppb. One could also crudely approximate an

inhalation REL from the oral NTP bioassay where a dose of 25 mg/kgday was associated with hematological effects.

The INDEX project Final report

106

The concentration approximately equivalent to a 25 mg/kg dose for a 70 kg human breathing 20 cubic meters per day is

27 ppm. Assuming this is a LOAEL and applying an RGDR of 1 for systemic effects, a 3 fold UF for extrapolation to

humans, a 10-fold UF for LOAEL to NOAEL extrapolation and a 10-fold UF for intraspecies extrapolation yields a

REL of 90 ppb. There are a number of uncertainties to this approach, yet it comes within a factor of 5 of the proposed

REL based on human studies.

IRIS (2003)

Integrated Risk Information System (IRIS) – U.S.EPA

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC)

Critical effect Exposures* UF MF RfC

Decreased lymphocyte count (Human

occupational inhalation study of

Rothman et al., 1996)

BMCL = 8.2 mg/m

300 1 0.03 mg/m

*Conversion factors: MW = 78.11. BMCL = 7.2 ppm, 8-hour TWA. Assuming 25ºC and 760 mm Hg, BMCL (mg/m

) = 7.2 ppm x MW/24.45 = 23.0

mg/m

. BMCL

ADJ

= 23.0 mg/m

x 10 m

/20 m

x 5 days/7days = 8.2 mg/m

. (The BMC was based on a benchmark response of one standard deviation

change from the control mean.)

The RfC is based on BMD modeling of the absolute lymphocyte count (ALC) data from the occupational epidemiologic

study of Rothman et al. (1996), in which workers were exposed to benzene by inhalation. A comparison analysis based

on BMD modeling of hematological data from the Ward et al. (1985) subchronic experimental animal inhalation study

was also conducted. In addition, comparison analyses using the LOAEL from the Rothman et al. (1996) study and the

NOAEL from the Ward et al. (1985) study were performed.

BMD modeling of the ALC exposure-response data from Rothman et al. (1996) was done using U.S. EPA's Benchmark

Dose Modeling Software (version 1.20). The data are rather supralinear, that is, the change in ALC per unit change in

exposure decreases with increasing exposure; therefore, in order to fit the data with one of the available continuous

models, the exposure levels were first transformed according to the equation d’ = ln(d+1). Then the exposure-response

data were fitted using the continuous linear model, which provided a good fit (p=0.54). A two-degree polynomial and a

power model also fit the data, but the linear model was selected because it is the most parsimonious. The parameters

were estimated using the method of maximum likelihood. A constant variance model was used.

In the absence of a clear definition for an adverse effect for this continuous endpoint, a default benchmark response of

one standard deviation change from the control mean was selected, as suggested in EPA's draft Benchmark Dose

Technical Guidance Document (U.S. EPA, 2000). This default definition of a benchmark response for continuous

endpoints corresponds to an excess risk of approximately 10% for the proportion of individuals below the 2nd

percentile (or above the 98th percentile) of the control distribution for normally distributed effects (see U.S. EPA,

2000). A 95% lower confidence limit (BMCL) on the resulting BMC was calculated using the likelihood profile

method. Transforming the results back to the original exposure scale yields a BMC of 13.7 ppm (8-hr TWA) and a

BMCL of 7.2 ppm (8-hr TWA).

As suggested in the draft technical guidance document (U.S. EPA, 2000), the BMCL is chosen as the point of departure

for the RfC derivation. An adjusted BMCL is calculated by converting ppm to mg/m3 and adjusting the 8-hour TWA

occupational exposure to an equivalent continuous environmental exposure. The BMCL is first converted to mg/m3

using the molecular weight of 78.11 for benzene and assuming 25ºC and 760 mm Hg: 7.2 ppm × 78.11/24.45 = 23.0

mg/m3. The converted value is then adjusted from the 8-hour occupational TWA to a continuous exposure

concentration using the default respiration rates (U.S. EPA, 1994): BMCLADJ = 23.0 mg/m3 × (10 m3/20 m3) × 5

days/7 days = 8.2 mg/m3.

The RfC is then derived by dividing the adjusted BMCL by the overall UF of 300: RfC = BMCLADJ/UF = 8.2 mg/m3

÷ 300 = 3 × 10-2 mg/m3. The overall UF of 300 comprises a UF of 3 for effect-level extrapolation, 10 for intraspecies

differences (human variability), 3 for subchronic-to-chronic extrapolation, and 3 for database deficiencies (see Section

I.B.3).

For comparison, an RfC was also calculated based on the LOAEL of 7.6 ppm (8-hr TWA) from the Rothman et al.

(1996) study. Converting the units and adjusting for continuous exposure as above results in a LOAEL

ADJ

of 8.7

mg/m3. The LOAEL

ADJ

is then divided by an overall UF of 1000 to obtain the RfC: 8.7 mg/m3 ÷ 1000 = 9 × 10-3

mg/m3. The combined UF of 1000 represents UFs of 10 to account for the use of a LOAEL because of the lack of an

appropriate NOAEL, 10 for intraspecies differences in response (human variability), 3 for subchronic-to-chronic

The INDEX project Final report

107

extrapolation, and 3 for database deficiencies. The value of 9 × 10-3 mg/m3 is in good agreement with the RfC of 3 ×

10-2 mg/m3 calculated from the BMC.

ATSDR (1997)

Agency for Toxic Substances and Disease Registry

Derivation of the Minimal Risk Level (MRL):

The intermediate-duration inhalation MRL of 0.004 ppm was derived from a LOAEL value of 0.78 ppm for

neurological effects of intermediate-duration inhalation exposure of mice to benzene (Li et al. 1992). The ratio of the

blood/gas partition coefficients was assumed to be 1. The dose was adjusted for intermittent exposure by multiplying

the LOAEL (0.78 ppm) by 2 hours/24 hours and by 6 days/7 days to adjust to continuous exposure. The resulting

adjusted LOAEL 0.056 ppm was then converted to a human equivalent concentration (HEC) by multiplying by the ratio

of the mouse ventilation rate (2.21 L/hour) to its body weight (0.0316 kg) divided by the ratio of the human ventilation

rate (833 L/hour) to human body weight (70 kg) (formula 4-l1, EPA 1990h; ventilation rate and body weight values

taken from EPA 1988e). The resulting LOAEL (HEC) of 0.33 ppm was then divided by an uncertainty factor of 90 (3

for the use of a minimal LOAEL, 3 for extrapolation from animals to humans after adjusting for the human equivalent

concentration, and 10 for human variability), yielding the MRL of 0.004 ppm (see Appendix A). ATSDR (1997) did not

derive a chronic-duration inhalation MRL for benzene due to lack of suitable data.

ECB (2003)

European Chemical Bureau - Institute for Health and Consumer Protection - European Union Risk Assessment Report

In its recent risk assessment draft report (2003) ECB concludes that chronic benzene exposure leads to depression of

white blood and red blood cells. This effect is reversible after long time exposures (years) with low concentrations

(reported concentration range: > 32-64 mg/m³ = 10-20 ppm). Exposure to 192 mg/m³ (60 ppm) of benzene for about a

week may be associated with an increased proportion of large granular lymphocytes, and not severe narrow effects nor

specific cytopenias. At higher concentrations, benzene may lead to aplastic anemia which can be fatal. Jandl's (1977)

review suggests a fatal outcome in 13% of the cases (as opposed to 85% for idiopathic aplastic anaemia).

The prevalence of leucopenia correlates with the concentration of benzene as shown by data of Yin et al. (1987) as well

as by Kipen et al. (1988, 1989). Taken these data, the LOAEC (lowest observed adverse effect concentration) for

leucopenia is in the range between 40 mg/m³ (12.5 ppm) and 64 mg/m³. A higher prevalence for leucopenia is given at

concentrations above 320 mg/m³ (100 ppm).

The case control studies presented recently by Rothman et al. (1996a, 1996b) and Dosemeci et al. (1997) have shown,

that the most sensitive reaction in humans to chronic benzene exposure is lymphopenia. The data show that a collective

of workers exposed to benzene concentrations in a range between 1 and 31 ppm had significantly reduced lymphocyte

counts as compared to a cohort of non-exposed workers.

Thus, for blood cell depression an overall LOAEC is suggested to be 32 mg/m³ (10 ppm). For depression of

lymphocytes, a NOAEC of 3.2 mg/m³ (1 ppm) has been derived. In conclusion, the NOAEC for effects of benzene is

assumed to be 3.2 mg (1 ppm).

Genotoxic effects

WHO (2000) Air Quality Guidelines for Europe 2000

There are numerous studies on chromosomal effects in exposed workers (2, 22). Both structural and numerical

chromosome aberrations have been observed. In most cases the benzene exposure has also been high enough to produce

haematological effects. The chromosomal effects in these studies are evident at concentrations of around 320 mg/m

(100 ppm) or higher, but in some studies effects were reported in workers chronically exposed to levels of around 32

mg/m

(10 ppm) (2, 23, 24). Tompa et al. (25) reported that the frequency of chromosome aberrations decreased when

exposure levels decreased from 3–69 mg/m

to 1–18 mg/m

. In the study by Karacic et al. (24), a decrease in sister

chromatid exchanges but not in chromosomal aberrations was noted in a group of female workers when examined with

a 5-year interval during which the mean benzene concentration had decreased from 26 to 16 mg/m

. Smoking did not

influence the results (23, 24, 26).

Somatic mutations as an endpoint of benzene-induced genotoxic effects in heavily exposed workers was studied

recently by Rothman et al. (27). They used the glycophorin A (GPA) mutation assay. The results suggested that benzene

induces gene-duplicating mutations, presumably through recombination mechanisms, but not gene-inactivating

mutations due to point mutations or deletions.

The INDEX project Final report

108

RIVM (2001)

National Institute of Public Health and the Environment - The Netherlands

The genotoxicity of benzene has been studied extensively. Benzene appears to have a peculiar genotoxic profile.

Benzene is hardly, if at all, able to induce gene mutations in in vitro test systems, despite its potent clastogenic

properties in vitro and in vivo. Several studies have demonstrated induction of both numerical and structural

chromosomal aberrations, sister chromatid exchanges, and micronuclei in experimental animals and humans after in

vivo benzene exposure (various routes of exposure). After short-term exposure to benzene levels of 3 to 30 mg/m

experimental animals showed chromosomal effects (Vermeirre et al. 1993) and some studies in humans have

demosnstrated chromosomal effects at mean workplace exposures as low as 4 to 7 mg/m

(EU 1999).

There appears to be a lack of evidence for direct DNA interaction under normal in vivo conditions: despite the

demonstrated ability of several benzene metabolites to form DNA adducts under in vitro conditions, DNA adduct

formation in bone marrow cells in experimental animals in vivo could only be demonstrated under quite specific and

very high exposure regimes (EU 1999; Health Council of the Netherlands 1997).

IRIS (2000)

Integrated Risk Information System (IRIS) – U.S.EPA

Current evidence indicates that benzene-induced myelotoxicity and genotoxicity result from a synergistic combination

of phenol with hydroquinone, muconaldehyde, or catechol. Molecular targets for the action of these metabolites,

whether acting alone or in concert, include tubulin, histone proteins, topoisomerase II, and other DNA-associated

proteins. Damage to these proteins would potentially cause DNA strand breakage, mitotic recombination, chromosomal

translocations, and malsegregation of chromosomes to produce aneuploidy. If these effects took place in stem or early

progenitor cells, a leukemic clone with selective advantage to grow could arise as a result of protooncogene activation,

gene fusion, and suppressor-gene inactivation. Epigenetic effects of benzene metabolites on the bone marrow stroma,

and perhaps the stem cells themselves, could then foster development and survival of a leukemic clone. Since these

plausible events have not been conclusively demonstrated, this remains a hypothesis (Smith, 1996).

UK Environment Agency (2003)

Benzene induces both numerical and structural chromosomal changes. Chromosome aberrations and micronuclei have

been commonly reported in laboratory animals exposed to benzene either in the atmosphere or by oral administration.

Often the positive results are seen throughout the tested dose range (ATSDR, 1997; WHO, 1993). Chromosomal

damage has been detected in the peripheral blood of workers by many teams of investigators (ATSDR, 1997), and

chromosomal effects have been claimed even at exposures of 4–7 mg m–3 (WHO, 2000).

Chromosomal aberrations may act as predictors of cancer risk, and their role in subsequent leukaemia is thought to

involve chromosome numbers 5 and 7, and also possibly numbers 8 and 21 (EBS, 1996; Irons and Stillman, 1996).

Aneuploidy may also be a precursor of leukaemogensis (EBS, 1996; Irons and Stillman, 1996). Other indicators of

genetic damage are sister chromatid exchange (SCE), DNA crosslinking, DNA adduct formation and DNA repair

(reviewed in Paustenbach et al, 1993; Snyder et al, 1993; EBS, 1996). Although benzene is not mutagenic in the Ames

test, there are a number of reports of treatment-related mutations in the tissues of mice exposed by inhalation (ATSDR,

1997).

Carcinogenic effects

Both IARC and the US EPA have assigned benzene their highest cancer classification, Group 1 (“is carcinogenic to

humans”) (IARC, 1987) and Category A (“known human carcinogen”) (USEPA, 2002), respectively. A large number of

epidemiological studies provide conclusive evidence of a causal association between benzene exposure and leukaemia,

acute non-lymphocytic leukaemia (ANLL) certainly, and also possibly chronic non-lymphocytic and chronic

lymphocytic leukaemias (USEPA, 2002). There have also been occasional reports of an increased risk in humans of

other neoplastic changes in the blood or lymphatics, including Hodgkin’s and non-Hodgkin’s lymphoma and

myelodysplastic syndrome (USEPA, 2002).

The vast majority of the evidence for an association with ANLL derives from occupational epidemiology, where

benzene is often a constituent of a complex mixture. These include studies in the shoe-making, printing, petrochemical,

chemical, coke production and rubber manufacturing industries (Aksoy et al, 1974; Christie et al, 1991; Decouflé et al,

1983; Hurley et al, 1991; Paxton et al, 1994a; Rushton, 1993; Tsai et al, 1983; Vigliani and Forni, 1976; Yin et al,

1989). Many of the workforces were exposed to benzene at levels markedly higher than those experienced in these

industries today.

The INDEX project Final report

109

There have been a number of cohort and nested case control studies that have estimated exposure for each worker.

Attempts have been made to develop dose–response relationships using cumulative exposure, which is, assuming a

symmetrical contribution of exposure concentration and exposure duration. The cancer mortality seen in the “Pliofilm”

cohort, which included workers from three factories manufacturing rubber film in the USA, has been most often used in

cancer risk assessment. A wide range of exposures were encountered, with relatively few other chemicals involved, and

a clear dose–response was seen between leukaemia risk and benzene exposure. There have been several publications

from this cohort giving mortality results (Infante et al, 1977; Paxton et al, 1994a; Rinsky et al, 1981, 1987; Wallace,

1996) and risk assessments using different methods of exposure estimation and different mathematical models (Austin

et al, 1988; Brett et al, 1989; Byrd and Barfield, 1989; Crump, 1994, 1996; Crump and Allen, 1984; Paustenbach et al,

1992; Paxton, 1996; Paxton et al, 1994b; Rinsky et al, 1989; Schnatter et al, 1996a,b). For example, the analysis of

Paxton et al (1994a), based on 14 cases of leukaemia in the exposed male workers and three previous estimates of

worker exposure (Crump and Allen, 1984; Paustenbach et al, 1992; Rinsky et al, 1981), reported standard mortality

ratios in the highest exposure group (>1600 mg m

–3

–1

) of between 10.3 and 20.

Four other cohort studies have quantified benzene exposures for all workers. Bond et al (1986), in a follow-up of a

study by Ott et al (1978) of 956 Dow Chemical workers, found three deaths from leukaemia compared with 1.9

expected. Wong (1987a,b) followed up 7676 chemical workers, 3536 of whom were continuously exposed to benzene.

Six leukaemia deaths were reported in the exposed group compared with 4.5 expected. An update of a subgroup of this

population has been reported by Ireland et al (1997). There were three cases of leukaemia observed at exposures at or

above 6 ppm (19.2 mg m–3) compared with 0.65 expected.

Increased risks of non-haematopoietic cancers have been found in a few studies of occupational cohorts exposed to

benzene or mixtures containing benzene. These include bladder, stomach, prostate and lung neoplasms (ATSDR, 1997),

although, in general, the risks are only moderately raised; also, in many studies, exposure to multiple chemicals

complicates the interpretation of the results.

Benzene has been implicated as a potential risk factor for the development of childhood leukemia (OEHHA, 1997; Ries

et al., 1997; Smith and Zhang, 1998; U.S. EPA, 1998).

There are little data on the effects of direct exposure of children to benzene. Because of their small size, increased

activity, and increased ventilation rates compared to adults, children may have greater exposure to benzene in the air, on

a unit body weight basis (U.S. EPA, 1998). Although no human cancer studies of benzene-exposed children are

available, there is good reason to believe that childhood exposure to benzene would also contribute to adult-onset

leukemias. Also, there is some evidence to suggest that exposure to benzene is associated with childhood leukemia.

Paternal exposure to benzene prior to conception in humans has been associated in some studies with increased

childhood leukemia, especially of the acute nonlymphocytic type, findings that are supported by observations in animals

of benzene-induced DNA damage to sperm. Maternal exposure to benzene in humans also has been associated with

increased incidences of childhood leukemia. These findings are supported by observations in animals of benzene-

induced transplacental genotoxicity, altered hematopoiesis, and of carcinogenicity, following exposure in utero and

continuing until weaning. It should be noted that other epidemiological studies that represented a large number of cases

of various subtypes of leukemia (Kaatsch et al., 1998) or acute lymphocytic leukemia only (Shu et al., 1999) did not

find an association with paternal benzene exposure. Thus a casual relationship would be difficult to establish.

Summary of cancer (leukemia) risk values

Risk value Unit Risk specific concentration

µg/m

(at an excess lifetime

cancer risk of 10

-4

)

Risk value name Study Source; Ref.

Value mg/m³

6 · E-6

(µg/m

)

-1

17 Unit risk (UR) and guideline

value (GV)

Pliofilm cohort

Crump, 1994

WHO 1995

2.2 - 7.8 · E-6 (µg/m

)

-1

13 - 45 UR Pliofilm cohort

multiple: Rinsky et al., 1981, 1987;

Paustenbach et al., 1993; Crump and

Allen, 1984; Crump, 1992, 1994;

U.S. EPA, 1998.

US-EPA 1998

5 · E-8

6 · E-6

(µg/m

)

-1

20-36

lowest plausible estimate

highest plausible estimate

multiple:

Wong and Raabe (1995)

Crump, 1994 (Pliofilm cohort)

EU 1999

1.5 · E+1 mg/m

30 TC

(Tumorigenic

concentration associated

with a 5% increase in

incidence) - acute

myelogenous leukemia

Pliofilm cohort

Rinsky et al., 1987

Health Canada

1991

The INDEX project Final report

110

2.4 · E-2 mg/m

24 CR

inhal.

(Lifetime inhalatory

excess cancer risk at the 1 in

level) - haematopoietic

system

multiple:

Wong and Raabe (1995)

RIVM 1999

0.054 (mg/kg

-1

4.1 using mean children’s

breathing rate of 0.45m

/kg-

day;

7.9 using mean adult

breathing rate of

0.235m

/kg-day

Inhalation cancer potency Pliofilm Cohort: Rinsky et al., 1987;

Paxton et al., 1994

Chinese Worker Cohort: Hayes et al.,

1997

OEHHA 2003

2 · E-5 (ppm)

-1

16 Lifetime leukemia (AML)

risk equivalent

Pliofilm cohort

multiple

Crump, 1994

USEPA, 1998

NICNAS 2001

WHO (1995)

Air Quality Guidelines for Europe 2000

The WHO derived in 1995 a Unit risk (UR) for leukemia of 6 · 10

–6

(µg/m

)

-1

on the basis of risk calculations of Crump

(1994). Based on this UR, air quality guidelines corresponding to excess lifetime cancer risks of 10

-4

, 10

-5

and 10

-6

are

17, 1.7 and 0.17 µg/m

, respectively (WHO, 2000)

US EPA (1998)

The US EPA is currently in the process of reviewing the evidence relating to the risk of benzene. Their provisional

conclusion is that there is insufficient evidence to reject a linear dose-response curve for benzene carcinogenicity at

low-dose exposures and that the approach of using a linear dose-response curve is still to be recommended. The US

EPA gives a range of risk estimate for leukemia at 1 µg/m

from 2.2 10

–6

to 7.8 10

–6

European Commission (1999)

The EU-Working Group concluded that it is reasonable to assume that benzene-induced effects at low exposure levels

differ quantitatively as well as qualitatively from those induced at high occupational exposures. It is not clear whether

the underlying mechanism of benzene carcinogenicity is direct DNA interaction. In view of this uncertainty the EU

Working Group decided, on precautionary grounds, to hold on to a non-threshold extrapolation method. Though it was

not possible to give a precise estimate of the risk associated with benzene it was possible to define a range within which

that risk was likely to lie. The UR calculated by the WHO (6 · 10

–6

) was considered to result in the highest plausible

estimate of risk. The lowest UR which the EU Working Group felt was likely to be plausible was in the order of 5 · 10

–

, an estimate which is consistent with the Dutch Health Council’s analysis of the Wong and Raabe data. This range of

URs means that an excess lifetime risk for leukemia of 10

–6

is associated with annual average concentrations varying

between 0.17 and 20 µg/m

(EU 1999).

On the basis of this risk estimation the European Commission has adopted an ambient air quality limit value for

benzene of 5 µg/m

to be met on 1 January 2010 (Directive 2000/69/EC). This limit value has been proposed by the EU

working group taking into account practical considerations; 5 µg/m

is as low as reasonably achievable in 2010.

Health Canada (1991)

Health Canada based their risk values on the incidence of leukemia in occupationally-exposed humans, resulting in a

(Tumorigenic concentration associated with a 5% increase in incidence) of 1.5E+1 mg/m

. The Health Canada

can be divided by 5000 to represent a 1 in 100,000 risk level of 3 · 10

–3

mg/m

. It should be noted that the Health

Canada TC

is computed directly from the dose-response curve within or close to the experimental range.

RIVM (1997)

National Institute of Public Health and the Environment - The Netherlands

The figure of 24 µg/m

(risk specific concentration at an excess lifetime cancer risk of 10

-4

) was derived from the

estimate as presented by the Health Council of 12 µg/m

being associated with an excess lifetime cancer risk of 10

-6

The figure of 12 µg/m

was based on the unit risk estimate of the US-EPA of 8.3 · 10

–6

as presented in the Dutch

Integrated Criteria Document on benzene (RIVM, 1987) increased with a factor of 100. In combination with the unit

risk of the US EPA the 95% upper confidence limit of the cancer risk in the Pliofilm cohort was calculated to be 0.052

The INDEX project Final report

111

(=6234* · 8.3 · 10

–6

). Linear interpolation between the two points on the dose response curve, i.e. [12 µg/m

, 10

–6

] and

[6234 µg/m

, 0.052] gives an 10

–4

excess cancer risk at 24 µg/m

* The average exposure in the Pliofilm cohort was 128 mg/m

. Converting this value to a continuous lifetime exposure

for the general population results in a value of 6234 µg/m

(128 mg/m

· 10/75 years/lifetime · 5/7 days/week · 48/52

weeks/year · 10/18 inhalation volume/day).

European Commission (1999)

The EU-Working Group concluded that it is reasonable to assume that benzene-induced effects at low exposure levels

differ quantitatively as well as qualitatively from those induced at high occupational exposures. It is not clear whether

the underlying mechanism of benzene carcinogenicity is direct DNA interaction. In view of this uncertainty the EU

Working Group decided, on precautionary grounds, to hold on to a non-threshold extrapolation method. Though it was

not possible to give a precise estimate of the risk associated with benzene it was possible to define a range within which

that risk was likely to lie. The UR calculated by the WHO (6 · 10

–6

) was considered to result in the highest plausible

estimate of risk. The lowest UR which the EU Working Group felt was likely to be plausible was in the order of 5 10

–8

an estimate which is consistent with the Dutch Health Council’s analysis of the Wong and Raabe data. This range of

URs means that an excess lifetime risk for leukemia of 10–6 is associated with annual average concentrations varying

between 0.17 and 20 µg/m

(EU 1999).

On the basis of this risk estimation the European Commission has adopted an ambient air quality limit value for

benzene of 5 µg/m

to be met on 1 January 2010 (Directive 2000/69/EC). This limit value has been proposed by the EU

working group taking into account practical considerations; 5 µg/m

is as low as reasonably achievable in 2010.

OEHHA (2003)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

The inhalation cancer potency of benzene was estimated from observations of leukemia among U.S. rubber

hydrochloride workers exposed to benzene (Pliofilm Cohort) (Rinsky et al., 1987; Paxton et al., 1994) and benzene-

exposed workers from various industries in China (Chinese Worker Cohort) (Hayes et al., 1997). In estimating potency,

observations were assumed to follow a Poisson distribution. Relative risk was assumed to increase linearly with

cumulative exposure, and leukemia onset was assumed unaffected by benzene exposure after 30 years. Cancer potency

estimates, obtained from the occupational data, were adjusted using lifetable analyses to those expected for the

California population from constant exposure to benzene.

Studies in human volunteers, workers, and animals suggest that humans retain approximately 50 percent of inhaled

benzene and absorb 100 percent of ingested benzene at environmentally relevant levels of exposure (OEHHA, 2001a).

Based on these differences in absorption of benzene by route of exposure, the oral cancer potency was derived from the

inhalation cancer potency.

The evidence regarding the shape of the exposure-response curve for benzene-induced leukemia supports the use of

linear risk models to extrapolate to low doses (U.S. EPA, 1998; OEHHA, 2001a,b). Saturation of metabolism and

formation of reactive compounds may result in nonlinearity at high exposures.

OEHHA (2001a) reviewed the scientific evidence for associations of benzene and various human cancers. The strongest

association is for acute non-lymphocytic leukemia and myelodysplastic syndromes (ANLL/MDS). MDS, if not fatal,

often progresses to ANLL. Evidence suggests that benzene causes other forms of leukemia as well. Total leukemia

(e.g., all subtypes of leukemia as a related class of diseases) was judged to be appropriate as the basis of risk

assessment.

Cancer potency estimates were derived for all exposures and after removal of workers with the highest cumulative

exposures. Highly exposed workers were removed from the analysis because of reduced confidence in their exposure

assignments or because of observed non-linearity of response at higher exposures.

Upper-bound estimates of cancer potency from Pliofilm Cohort or Chinese Worker Cohort were very similar after

removal of the workers with the highest exposures from each cohort. Cancer potency estimates in units of ppm

-1

were

converted to units of (mg/kg-d)

-1

using default intake parameters of 70 kg body weight and 20 m

/d for breathing rate:

Potency/(mg/kg-d) = (upper-bound potency/ppm) x (ppm/3.19 mg/m

) x (70 kg) x (1/20 m

/d).

The INDEX project Final report

112

Thus, inhalation cancer potency estimates from the Pliofilm Cohort based on the Rinsky et al. or Crump and Allen

exposure estimates are 0.048 and 0.049 (mg/kg-day)

-1

, respectively, resulting in a geometric mean of 0.048 (mg/kg-

day)

-1

after rounding. The cancer potency estimate based on the Chinese Worker Cohort is 0.061 (mg/kg-day)

-1

. The

geometric mean estimate of these two cohort studies is 0.054 (mg/kg-day)

-1

. Cancer potency estimates from inhalation

exposures were converted to those expected from oral exposures by applying a factor of two since 50 percent of inhaled

benzene is absorbed and 100 percent of ingested benzene is absorbed (OEHHA, 2001a). The resulting oral cancer

potency estimates are 0.11 (mg/kg-day)

-1

The cancer potency estimates obtained from the Pliofilm Cohort and Chinese

Worker Cohort are consistent with estimates obtained from animal cancer bioassays of benzene, other epidemiological

studies of benzene-exposed workers, and epidemiological studies of leukemia and cigarette smoke (of which benzene is

a constituent) (OEHHA, 2001a).

The NO SIGNIFICANT RISK LEVEL (NSRL) for Proposition 65 is the intake associated with a lifetime cancer risk of

-5

. The cancer potency estimates derived above were used to calculate NSRLs for benzene, which are 13 µg/day for

inhalation exposures and 6.4 µg/day for oral exposures.

NICNAS (2001)

National Industrial Chemicals Notification and Assessment Scheme – Australia

NICNAS considered the Pliofilm cohort as the most reliable data set on which to establish the risk estimates. Data from

the most recent follow-up of the Pliofilm cohort indicate that the risk for leukaemia is significantly elevated at

cumulative exposures >50 ppm-years. However, this finding derives from a single cohort study with insufficient

statistical power to rule out the possibility of some increase in leukaemia risk at lower exposures. The additional risk for

leukaemia attributable to environmental exposure to benzene can be predicted by low-dose extrapolation of the

quantitative estimates for occupational exposure to benzene. Crump (1994) and USEPA (1998a) both predict the

number of additional leukaemia cases at two lifetime exposure levels, namely 1 ppm and 1 ppb, however, only the latter

is relevant for exposure to benzene in the general environment.

The risk characterisation is based on the most conservative prediction of the Pliofilm cohort (Crump, 1994 and USEPA,

1998a), that is, a lifetime leukaemia (AML) risk equivalent to 2 additional cases of leukaemia per 100,000 population at

1 ppb. By extrapolation, the lifetime leukaemia (AML) risk equivalent for increasing exposure levels can be calculated

as follows:

1 ppb ～ 2 additional leukaemia/ 10

population

10 ppb ～20 additional leukaemia / 10

population

20 ppb ～ 40 additional leukaemia / 10

population

The NICNAS estimated that the average 24 hour lifetime exposure to benzene, from all sources (including an estimated

ambient air component of 11%), of an individual not exposed to environmental tobacco smoke living in an urban area

of an Australian city is 5.2 ppb. The predicted excess lifetime risk of leukaemia is therefore 1/10,000, or 1.2% of the

lifetime risk of contracting leukaemia of any cause (1 in 118, or 85/10,000 population, based on 1996 incidence figures

for Australia (AIHW, 1999).

AIHW. 1999. Cancer in Australia 1996 – incidence and mortality data for 1996 and selected data for1997 and 1998. Bruce, ACT,

Australian Institute of Health and Welfare.

NICNAS. 2001. National Industrial Chemicals Notification and Assessment Scheme, Benzene, Priority Existing Chemical

Assessment Report No. 21 (2001)

Susceptible population

The potential magnitude or range of susceptibilities cannot be accurately quantified at this time; however, current

evidence suggests individual differences in susceptibility may be two or three orders of magnitude.

Benzene’s toxicity is intimately tied to its complex metabolism and distribution. Key enzymes involved in the

metabolism of benzene, including CYP2E1, the quinone reductase NQO1, and myeloperoxidase are polymorphic. For

example, a 7.6-fold difference in benzene-induced hematotoxicity in workers was observed among gene variants of

CYP2E1 and NQO1 (Rothman et al., 1997). Thus the expression of genetic polymorphisms may modulate the

sensitivity of an individual or ethnic group to the effects of benzene exposure.

Aksoy et al. (1976) suggested familial susceptibility as one of the main factors in the frequency of chronic benzene

toxicity. Two patients with pancytopenia associated with chronic exposure to benzene were cousins. One of these two

The INDEX project Final report

113

patients died of aplastic anemia; the son of this patient presented with leukopenia a short time after exposure to this

chemical. Genetic polymorphism with regard to the P-450 enzymes that metabolize benzene to its reactive metabolites

may render some individuals at higher risk to the toxic effects of benzene (Kato et al. 1995). Although there are

possible indications of genetic susceptibility to benzene, it is believed that all humans are susceptible to the

pancytopenic effects of this agent.

Children may represent a subpopulation at increased risk due to factors that could increase their susceptibility to effects

of benzene exposure (e.g., activity patterns), on key pharmacokinetic processes (e.g., ventilation rates, metabolism rates,

and capacities), or on key pharmacodynamic processes (e.g., toxicant-target interactions in the immature hematopoietic

system).

There is a possible relationship between thalassemia (abnormal haemoglobins) and exposure to benzene. Aksoy (1989)

presented a series of 44 pancytopenic patients, 4 of which had β-thalassemia heterozygotes. Two of these four patients

had high levels of fetal haemoglobin, while a third patient had pseudo-Pelger-Huet anomaly. Thus, some forms of β-

thalassemia may increase the deleterious effects of benzene on the hematopoietic system.

The occurrence of aplastic anemia in chronic benzene toxicity may be accelerated in individuals with viral hepatitis

(Aksoy 1989). Furthermore, children and fetuses may be at increased risk-because their hematopoietic cell populations

are expanding and dividing cells are at a greater risk than quiescent cells.

In vitro studies of benzene-caused chromosomal aberrations in human lymphocytes taken from people with different

health practices suggests that “unhealthy lifestyles, ” including smoking, excessive alcohol consumption, inadequate

physical exercise and sleep, excessive stress, and inadequate nutritional balance, could make cells more sensitive to the

production of chromosomal aberrations after exposure (Morimoto et al. 1983).

Interactions with other chemicals

It has been reported that high exposures to gasoline (300 to 2000 ppm) which contains benzene may be less hazardous

than an equivalent exposure to benzene alone, based on studies in mice and from modeling predictions (Bond et al.,

1997). Volatile aromatics in gasoline including toluene, ethylbenzene and xylenes compete with benzene for CYP2E1,

thus these co-exposures significantly reduce the overall metabolism of benzene relative to equivalent exposures of pure

benzene. However, these experiments employed relatively high concentrations of test material (1) 300 ppm gasoline

plus six ppm benzene, and (2) 2000 ppm gasoline plus 40 ppm benzene), levels at which metabolic saturation is

expected. At low levels of exposure, competition for the P450 binding sites would be minimal; thus, reduction in

benzene toxicity due to co-exposures would also be minimal.

Odour perception

Source: Devos et al. (1990)

Odour threshold: 1.2 mg/m

Source: Haley, 1977

Odour threshold: 2.8 mg/m³

Source: AIHA (1989) - Odor thresholds for chemicals with established occupational health standards.

Detection: 195 mg/m

(61 ppm); Recognition 309 mg/m

(97 ppm). Reported values range from 2.5 – 510 mg/m

(0.78-

160 ppm).

The INDEX project Final report

114

Summary of Benzene Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: ingestion of contaminated tap water, potential exposure from dermal contact during

showering and bathing (additional inhalatory exposure due to volatilized benzene). Dietary and endogenous sources of

phenol, hydroquinone and other primary metabolites of benzene confer potentially large differences in susceptibility to

benzene toxicity.

Toxicokinetics: at low levels (<10 ppm) about half of the inhaled dose is retained, the rest being exhaled. Metabolism

of benzene occurs largely in the liver, although some metabolism may take place in the bone marrow. Benzene can

cross the placenta. Not benzene itself, but several metabolites, as well as interactions between these metabolites, are

thought to explain the toxic effects of benzene. The proportion of the putatively toxic metabolites (benzoquinone and

muconaldehyde) formed is greater at low doses than at high doses (saturable process at relatively low doses).

Health effect levels of short- and long-term exposure (noncancer):

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

128

EXP

319

EXP

Developmental: decreased fetal body weights

(severe adverse effect!)

Pregnant female rats,

6 hours per day (for 5

days)

Coate et al., 1984;

(supported by Kuna

and Kapp, 1981;

Keller and Snyder,

1988)

OEHHA 1999

REL: 1.3

EXP

Developmental: Altered hematopoiesis in

offspring

(high degree of uncertainty!)

Mice, 6 hours per

day during days 5-16

of gestation

Keller and Snyder,

1988

OEHHA 2001

MADL

inhal

.: 1.3

µg/day

EXP

ADJ

HEC

Haematologica-immunological: depressed

peripheral lymphocytes and mitogen-induced

blastogenesis of femoral B-lymphocytes (Less

Serious LOAEL) .

Mice, 6 hours per

day (for 6

consecutive days)

Rozen et al. 1984

ATSDR 1997

MRL: 0.16

Long-term exposure

1.7

Study

0.60

ADJ

Haematological effects (total and differential

white blood cell counts, haemoglobin,

hematocrit, red blood cells, platelets and

clotting times)

Occupational, 7.4y Tsai et al. 1983

OEHHA 1999

REL: 0.06

BMC

BMCL

8.2

ADJ-

BMCL

Decreased lymphocyte counts Occupational, 6.3y

(0.7-16y)

Rothman et al. 1996 EPA-IRIS 2003

RfC: 0.03

(3.2) 32

EXP

Blood cell depression (lymphocytes) Mouse, 6-178d

NOAEC has been

derived, not observed

Dempster and

Snyder, 1990, Rozen

et al. 1984, Baarson

et al., 1984, Green et

al., 1981a,b, and

other studies

ECB 2003

2.5

EXP

1.05

HEC

Neurological: increased rapid response time

(Less Serious minimal LOAEL).

Mice, 30d

(2h/d, 6d/w)

Li et al. 1992

intermediate duration

study !

ATSDR 1997

MRL

INT

: 0.013

(intermediate!)

Carcinogenicity: IARC: 1 ; U.S.EPA: A ; Genotoxic carcinogen

Unit risks derived from Pliofilm cohort study - leukemia in occupational settings.EPA (1998): 2.2 - 7.8 · 10

-6

(µg/m³)

-1

;

WHO(1995): 6.0 · 10

-6

(µg/m³)

-1

; OEHHA(2003; Pliofilm cohort and Chinese worker cohort) with respect to breathing

rate: children 24 · 10

-6

(µg/m³)

-1

and adults 12 · 10

-6

(µg/m³)

-1

Genotoxicity: Induction of both numerical and structural chromosomal aberrations, sister chromatid exchanges and

micronuclei in experimental animals and humans after in vivo benzene exposure (as low as 4-7 mg/m³). Lack of

evidence for direct DNA interaction under normal in vivo conditions. Genotoxicity may result from a synergistic

combination of phenol with hydroquinone, muconaldehyde, or catechol.

Odour threshold: 1.2 mg/m³ (Devos); 2.8 mg/m³ (Haley, 1977)

Susceptible sub-populations: The range of susceptibility to benzene toxicity within the general population could be

very large (two or three orders of magnitude). Significant sources of variability are:

The INDEX project Final report

115

- genetic polymorphisms in key enzymes involved in the metabolism of benzene,

- dietary and endogenous sources of primary metabolites of benzene,

- factors such as pre-existing inflammation, radiation exposure, and ethanol consumption can potentiate the toxic

effects of exposure to benzene.

- Benzene has been implicated as a potential risk factor for the development of childhood leukemia.

Remarks:

The INDEX project Final report

116

6. Risk Characterization

Cancer risk evaluation

The most severe human responses to benzene exposure are those related to the blood-forming organs. There is a wealth

of evidence demonstrating benzene’s ability, at high concentrations, to cause bone marrow damage giving rise to

clinical outcomes such as various anemias, myelodysplastic syndromes, and leukemia. Additional evidence suggests

that benzene may also be associated with increased incidence of childhood leukemia. Benzene is regarded as a

genotoxic chemical, causing chromosomal damage and mutations in test systems in vitro, in animals in vivo, and in

occupationally exposed humans.

The present risk characterization is adressing benzene-induced leukemia and is supported by an evaluation on

haematological effects, considered a promotional condition to tumour development. Since no safe exposure level can be

established for genotoxic carcinogens, the quantitative risk estimate is here based on a non-threshold linear

extrapolation at low levels using the WHO Unit Risk guideline set as 6 · 10

–6

(µg/m

)

-1

and described as the geometric

mean of the range of estimates of the excess lifetime risk of leukemia at an air concentration of 1 µg/m

. This factor,

applied to a single exposure value or to a population distribution of exposure values, provides an estimate of the

individual risk or the population-based risk, the latter indicating the potential excess cancers among a million people

continuously exposed to the given benzene concentration over a 70-year lifetime.

Population cancer risk estimates of benzene-induced leukemia

In Table 6.1 the available exposure data (see cumulative frequency distributions in Chapter 1.3) are listed along with the

lifetime cancer risk estimates at three different exposure levels. Indoors, risk estimates are based on the median and the

percentile of the cumulative exposure distribution, with the 90

percentile describing a “worse case“ scenario

among each indoor survey. Both indoor levels were corrected following the assumption that people spend 95% of their

time in their homes and resting time outdoors. Additionally, risk estimates are refered to median outdoor concentrations,

a scenario describing indoor exposure experienced in the absence of indoor sources of benzene.

Table 6.1 Excess lifetime cancer risks per one million people derived from available exposure data

Population-based studies

Description (Study, Year)

Median estimate

Indoors

% estimate

Indoors

Median estimate

Outdoors

Athens, 30h-TWA (Expolis, 96-98) 42 48 112 45

Basel 30h-TWA (Expolis, 96-98) 47 13 28 8

Helsinki 30h-TWA (Expolis, 96-98) 188 10 25 9

Milan 30h-TWA (Expolis, 96-98) 41 60 216 55

Oxford 30h-TWA (Expolis, 98-00) 40 16

n.a.

Prague 30h-TWA (Expolis, 96-98) 46 41 80 28

England 28d-TWA (BRE, 97-99) 796 18

n.a.

Germany 4w-TWA (GerES IV pilot

study, spring+summer 2001)

44 8 16 7

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

109

n.a.

supposing that people spend 95% of their time at home

supposing outdoor air as the unique source of indoor benzene exposure

not adjusted for outdoor exposure (0.05%)

n.a. data not available

The corresponding benzene concentrations are given in Table 6.2, with an additional column pointing out the

percentage of individuals exposed at levels higher than the current (2004) and the finalised (2010) EU-limit value of 10

and 5 µg/m

, respectively (1-y average; Directive 2000/69/EC). This limit value has been proposed by the EU working

group taking into account the developed risk estimation and practical considerations (5 µg/m

is as low as reasonably

considered achievable in 2010). Also, in Figure 6.1 the outcomes of the available indoor surveys are presented together

with the results of the MACBETH project.

The INDEX project Final report

117

0,0

0,5

1,0

1,5

2,0

2,5

(

) G

y 20

inte

(

[SF] 1996-98

(2) Bas el [CH] 1996-98

(

) F

(

)

K] 19

(

)

(4) England 1

97-99

(3) Pa

ova [I] 19

(2) Prague [CZ] 1996

(

)

wer

[

(

]

)

Ath e

s (

(

2) Athens [E]

1996-98

(3) M

rcia [E]

7-98

)

rma

(

) Mi

n [I

]

Population excess lifetime cancer risk per 10.000 people

2004 EU Limit Value

2010 EU Limit Value

Table 6.2: Benzene concentrations (µg/m

) measured and percentages of population exposed indoors at levels higher

than the EU-limit value (1y avg.): 10 µg/m

(current) and 5 µg/m

to be met the 01/01/10

% exceeding

Population-based studies Description

(Study, Year)

Median

Indoors

Median

Outdoors

10 µg/m

current

5 µg/m

01/01/10

Athens, 30h-TWA (Expolis, 96-98) 42 8,0 19,2 7,5 35% 75%

Basel 30h-TWA (Expolis, 96-98) 47 2,3 4,8 1,3 < 10%

Helsinki 30h-TWA (Expolis, 96-98) 188 1,7 4,3 1,5 < 10%

Milan 30h-TWA (Expolis, 96-98) 41 10,0 37,5 9,1 50% 75%

Oxford 30h-TWA (Expolis, 98-00) 40 2,6 7,3 - 5% 20%

Prague 30h-TWA (Expolis, 96-98) 46 6,9 13,8 4,6 25% 75%

England 28d-TWA (BRE, 97-99) 796 3,0 9,0 - 15% 30%

Germany 4w-TWA (GerES IV pilot study,

spring+summer 2001)

44 1,3 2,7 1,1 < <

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

109

2,3 5,0 - < 10%

n.a. data not available;

< out of the evaluation range (i.e. <5% of the environments investigated)

Figure 6.1:Cancer risk estimates of benzene induced leukemia (last minute implementation)

based on arithmetic mean and 90%ile (---) indoor levels measured among European surveys

(1) GerES II ; (2) EXPOLIS ; (3) MACBETH ; (4) BRE National survey ; (5) GerES IVP ; (6) French National

survey

The INDEX project Final report

118

Health hazard evaluation (noncancer)

Both epidemiology and animal experiments show that carcinogenic effects occur at benzene levels also expected to be

haematotoxic. Together with the observation that susceptibility to haematotoxicity by benzene evidently predisposes to

myelodysplastic syndromes as well as haematopoietic malignancies, this target organ toxicity is considered a

promotional condition to tumour development.

A complementary health hazard evaluation is here based on haematological effects (see Table 6.3) with an observed

chronic NOAEL of 600 mg/m

(OEHHA, corrected for continuous exposure) and an assessment factor of ten for inter-

individual variability. In the occupational study refered to, total and differential white blood cell counts, hemoglobin,

hematocrit, red blood cells, platelets and clotting times of the population studied were found to be within normal limits

(between 5% and 95% percentile) during an exposure duration averaged 7.4 years.

Table 6.3: Haematological effects:

Population exposed beyond given threshold levels (µg/m

)

NOAEL

Population-based studies

Description (Study, Year)

N 600 60

Athens, 30h-TWA (Expolis, 96-98) 42 < <

Basel 30h-TWA (Expolis, 96-98) 47 < <

Helsinki 30h-TWA (Expolis, 96-98) 188 < <

Milan 30h-TWA (Expolis, 96-98) 41 < 5%

Oxford 30h-TWA (Expolis, 98-00) 40 < <

Prague 30h-TWA (Expolis, 96-98) 46 < <

England 28d-TWA (BRE, 97-99) 796 < <

Germany 4w-TWA (GerES IV pilot

study, spring+summer 01)

44 < <

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

109

< <

n.a. data not available; < out of the

evaluation range (i.e. <5% of the

environments investigated)

The studies

The EXPOLIS study has been designed to produce direct population based measurements of exposures for major air

pollutants, the target population being the adult urban population of Europe (measurements performed from fall 1996 to

winter 1997/1998). (Selection bias (Oglesby et al., 2000) In Basel, participants of direct monitoring as compared to non-

participants were more likely to live at streets with low traffic volume; although in Helsinki, traffic volume was neither

significantly related to participation in direct nor indirect monitoring, the point estimates indicate a tendency to

decreased participation with increasing traffic intensity at home.)

The Building Research Establishment (BRE, UK) has conducted a national representative survey of air pollutants in

876 homes in England, designed to increase knowledge of baseline pollutant levels and factors associated with high

concentrations (17 months, October 1997 to February 1999).

The GerES IVP is a one-year pilot study (February 2001 - March 2002) of the German Environmental Survey (GerES),

and includes about 560 children/teenagers (0 to 17 years) with a sub collective of 112 for personal, indoor and outdoor

air analyses. The aim of the study was to collect information on parameters influencing the response rate and to test the

suitability of the different instruments intended to be used for the main study. The here presented benzene exposure data

are based on indoor air measurements averaged over 4 weeks.

The French National Survey: Study details in

A nationwide survey on indoor air quality has been st up in France in 2003-2004, with the aim of assessing household

exposure to indoor pollutants. The target population is the national housing stock of approximately 24 million

permanently occupied housing units. For further details refer on Golliot et al., 2003 and Kirchner et al., 2004.

The INDEX project Final report

119

Comments

A widespread characterization of residential exposures to benzene is given by the outcome of the Expolis-study,

considered here as indicative of the urban population in Europe. Within this study, a specific bias has been reported

(Oglesby et al. 2000) concerning the selection of residences (participants) in Basel and Helsinki, with these data more

likely to be collected in lower trafficked areas of the towns.

In two further surveys (BRE and GerES IVP), one-month average benzene levels were measured in randomly

distributed residences in England and Germany. With the exception of the GerES IV pilot study, all studies took into

account possible seasonal effects on indoor benzene concentration.

Tendency in ambient benzene levels

Since outdoor air is representing a source of indoor benzene levels, the adoption, in most member states, of

environmental policies regarding benzene emissions, raises the question on whether the available/selected exposure data

could be considered descriptive of the nowadays situation, the comment requiring a brief introduction to the policies

adopted and on-going tendencies of benzene ambient levels.

Combustion processes are the largest source of benzene in the atmosphere, with road traffic generally the biggest single

source. Especially in urban locations, the emission by vehicles is expected to be the principal cause of increased

benzene concentrations. Emissions are resulting from

direct emission from the exhausts and from the

evaporation of fuels either by car or from fuel

distribution and refuelling. These emissions have been

declining in recent years, owing to legislation on

vehicle emissions, industrial emissions and fuel

distribution. In particular, important results were

obtained with the introduction of the automobile

catalyst converter in the last two decades and the

reduction of the benzene content of gasoline more

recently. The Air Quality Report of the first Auto-Oil

Programme (AOP1) estimated a 56% reduction in

urban emissions of benzene between 1990 and 2010.

As an example, the control of benzene in gasoline has

recently been described as the main reason for the

decrease of indoor benzene levels in Germany, together

with the reduction of the use of benzene in many

consumer products (Seifert, 2002; Schleibinger et al., 2001). Levels decreased by a factor of about 4 over one decade in

the Berlin metropolitan area (see Figure). Although an example of a successful application of environmental control

measures in Germany, a tendencial decrease is expected to occur in all member states adopting these policies. As a

result, both the Expolis and the BRE data, collected more than 5 years ago, are expected to describe a pessimistic

picture of the actual situation, with respect to the contribution of ambient benzene levels.

When examining the exposure data (Table 6.2) with respect to the contribution of outdoor air contamination, two

exposure scenarios could be depicted as follows:

- In highly trafficked urban areas (outdoor benzene levels >5 µg/m

), outdoor air is well contributing to indoor

benzene levels, as shown in the case of Milan and Athens, where median indoor-levels exceed outdoor-levels for

not more than 9%. In these locations the relative contribution of indoor specific sources (e.g. tobacco smoke) is

likely to be impaired.

- At low outdoor levels (benzene <5 µg/m

), indoor sources moderately (Basel) or marginally (GerES IVP,

Helsinki) contribute to exposure, with median indoor levels not more than three-fold the background benzene

concentration (0.6 to 1.9 µg/m

in European continental pristine air,

as reported by ECB 2004).

Benzene levels measured in Prague and Oxford (EXPOLIS) and in England (BRE) are expected to fall between the two

scenarios described.

As pointed out in Table 6.1, the additional lifetime cancer risk in Europe is expected to range from 8 to 216 · 10

–6

(µg/m

)

-1

resulting in a 27-fold increase experienced by worse-case residentials in Milan (37 µg/m

) with respect to

average residentials in Germany (1.3 µg/m

), the latter considered as the European background cancer risk estimate,

with 7.5 cases of benzene-induced leukemia per one million people. Moreover, almost the entire bulk of data is below

the exposure level (60 µg/m

) considered safe with respect to haematological effects, the most severe toxicological

enpoint other than cancer.

The INDEX project Final report

120

EU limit value for benzene in ambient air – Directive 2000/69/EC

Limit value for the protection of human health: 5 µg/m

(averaging period of a calendar year)

Margin of tolerance: 5 µg/m

(100 %) on 13 December 2000, reducing on 1 January 2006 and every 12 months thereafter by 1 µg/m

to reach 0 % by

1 January 2010

Date by which limit value is to be met: 1 January 2010, except within zones and agglomerations within which a time-limited extension has been

agreed in accordance with Article 3(2) of the directive. The limit value for benzene to be granted during conditional extension for a period of up to

five years shall, however, not exceed 10 µg/m

Result

Benzene is ubiquitous in the athmosphere, mainly due to anthropogenic sources (90%), with concentrations in the

European continental pristine air ranging from 0.6 to 1.9 µg/m

. It is a genotoxic carcinogen and hence no safe level of

exposure could be recommended. Results from nine monitoring surveys indicate that the European population is

experiencing in their homes an increased risk in contracting benzene induced leukemia, with respect to the estimated

background lifetime risk of 7-8 cases per one million people (considering the WHO unit risk factor). Based on the

available exposure data (Median levels±sd: 4.2±3.2 µg/m

; 90th levels±sd: 11.5±11.1 µg/m

; N = 9) two main scenarios

could be described as follows:

- People living in highly trafficked urban areas are found, on average, to experience an estimated 6 to 30-fold increase

in contracting benzene induced leukemia during their life, the benzene levels encountered in these areas not expected

to produce chronic effects other than cancer, in particular haematological effects, nor acute sensory effects such as

odour perception (odour threshold: 1.2 mg/m

) and sensory irritation. Also, the contributions of specific indoor

sources in these areas are likely to be impaired due to increased background outdoor levels.

- People living in rural areas or poorly trafficked towns were found, on average, to experience an estimated 1 to 5-fold

increase in contracting benzene induced leukemia during their life, this factor depending principally on the presence

of indoor sources.

The INDEX project Final report

121

Naphthalene

Synonyms: antimite, naphthalin, naphthaline, naphthene, tar camphor

CAS Registry Numbers: 91-20-3

Molecular Formula: C

1. Compound identification

Naphthalene is white crystalline powder with aromatic odour (of mothballs). It is a two-ring hydrocarbon isolated from

coal tar. It is used as intermediate in chemical synthesis, as insect repellents, fungicides, lubricants, preservatives, and,

formerly, as topical antiseptics (EPA/Cal 2003, HSDB 2003). Gasoline and diesel fuels contain naphthalene.

Naphthalene is used indoors as a moth repellent, though this use is decreased. It has also been used in the manufacture

of phthalic anhydride, phthalic and anthranilic acids, naphthols, naphthylamines, 1-naphthyl-n-methylcarbamate

insecticide, beta-naphthol, naphthalene sulfonates, synthetic resins, celluloid, lampblack, smokeless powder,

anthraquinone, indigo, perylene, and hydronaphthalenes (EPA/Cal 2003, HSDB 2003).

Naphthalene emissions to atmosphere are mainly originated from fugitive emissions and motor vehicle exhaust. Spills

into land and water during the storage, transport and disposal of fuel oil and coal tar are lost and released to atmosphere

due to volatilisation, photolysis, adsorption, and biodegradation. Naphthalene has a relatively short half-life, 3-8 hr, in

the atmosphere. It is assessed that the primary route of exposure is inhalation, especially in vicinity of heavy traffic, gas

stations or refineries. Usual indoor sources of naphthalene are unvented kerosene heaters and tobacco smoke (EPA/Cal

2003, HSDB 2003).

2. Physical and Chemical properties

Molecular weight (g/mol) 128.18

Melting point (°C) 80.5

Boiling point (°C) 218

Density, (g/cm3, at 20 °C, 1 atm) 4.42

Solubility Soluble in alcohol, acetate

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 5.321 µg/m

1 µg/m

= 0.188 ppb

Source: CRC (1995), Verschueren 2001, EPA/Cal (2003)

3. Indoor Air Exposure assessment

Indoor air and exposure concentrations

Average outdoor naphthalene concentrations were low in Europe, ranging from 1 to 4 µg/m

(Table 4.0.1, Jantunen et al

1999). Also residential indoor concentrations were elsewhere low, typically average concentrations below 2 µg/m

, but

in Athens clearly higher indoor levels were measured, being on average 90 µg/m

. Personal exposures to naphthalene

ranged from 1 µg/m

to 3 µg/m

elsewhere (Jantunen et al 1999, Hoffman et al 2000), but in Athens average exposure

was 46 µg/m

. In general we can conclude that exposures to naphthalene were usually low in Europe, but in Athens

remarkable indoor sources of naphthalene were present, because personal exposure and workplace concentrations were

lower than indoor concentrations (Figure 3.1, Table 4.0.1). Figure 3.2 shows the same distributions as Figure 3.1, but

without Athens to be able to see distributions better focused on their scale. Similarly, exposure distributions with and

without Athens are presented in Figure 3.3 and Figure 3.4.

The INDEX project Final report

122

Maroni et al (1995) reported typical median and 90

percentile naphthalene concentrations in indoor air being 2 µg/m

and 5 µg/m

, respectively. Kostiainen et al (1995) detected slightly lower indoor concentrations in Helsinki, Finland,

having 0.44 µg/m

and 1.63 µg/m

as a mean and maximum concentrations.

Bituminous material, commonly used in UK for damp proofing floors emits naphthalene (Brown et al 1990).

Naphthalene concentrations up to 970 µg/m

were found in homes having an objectionable smell, where the damp proof

membrane had been applied, compared with <300 µg/m

for control homes (EU 2003).

In an Italian study average indoor naphthalene concentration was 11 µg/m

, maximum level being 70 µg/m

(DeBortoli

et al 1986). Similar concentrations were measured in Canada, showing average and maximum concentrations of 14

µg/m

and 77 µg/m

, respectively (Chan et al 1990).

100

0 50 100 150 200 250 300 350

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of naphthalene in Athens (Ath, n=42), Basel

(Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=38) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS 2002).

100

03691215

Indoor concentration (µg/m

)

Cumulative frequency (%)

Bas

Hel

Mil

Oxf

Pra

Figure 3.2. Cumulative frequency distributions of indoor air concentrations of naphthalene in Basel (Bas, n=47),

Helsinki (Hel, n=188), Milan (Mil, n=38) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS 2002).

The INDEX project Final report

123

100

0 50 100 150

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.3. Cumulative frequency distributions of 48-hour personal exposure concentrations of naphthalene in Athens

(Ath), Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean concentrations

of the German Survey GerES II (GeS) (Hoffman et al 2000).

100

024681012

Exposure (µg/m

)

Cumulative frequency (%)

Bas

Hel

GeS

Oxf

Pra

Figure 3.4. Cumulative frequency distributions of 48-hour personal exposure concentrations of naphthalene in Basel

(Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean concentrations of the German

Survey GerES II (GeS) (Hoffman et al 2000).

The INDEX project Final report

124

4. Toxicokinetics

Absorption

No empirical data that describe the rate or extent of naphthalene absorption following inhalation exposure were

identified. NTP (2000) developed a physiologically-based pharmacokinetic model to describe the uptake of naphthalene

in rats and mice following inhalation exposure. The model was calibrated using blood time course data for naphthalene

(parent compound). Results from this model suggest that inhaled naphthalene is absorbed rapidly into the blood

(Blood:air partition coefficient of 571). On the basis of estimates of naphthalene metabolism generated by the model,

approximately 22% to 31% of inhaled naphthalene is absorbed by rats and 65% to 73% of inhaled naphthalene is

absorbed by mice.

Assuming an average ambient concentration level of 5.19 µg/m

and an average inhalation rate of 15.2 m

/day (U.S.

EPA, 1996), an average daily dose of 1,127 ng/kgday can be calculated for a 70-kg adult. An estimated average daily

dose of 4,5 ng/kg-day can be calculated for a 10-kg child assuming an inhalation rate of 8.7 m

/day (U.S. EPA, 1996).

Distribution

There are limited data concerning the distribution of naphthalene in human tissues. Naphthalene was present in 40% of

the adipose tissue samples that were analyzed as part of the National Human Adipose Tissue Survey (EPA 1986). The

maximum concentration observed was 63 ng/g. Naphthalene was also detected in human milk samples (concentration

not reported) (Pellizzari et al. 1982). The sources of naphthalene in these milk and body fat samples are not known.

Three reports (Zinkham and Childs, 1958; Anziulewicz et al., 1959; Athanasiou et al., 1997) describe apparent

transplacental exposure of a fetus during pregnancy, which resulted in neonatal hemolysis. In the two older cases,

unspecified amounts of naphthalene had been ingested by the mother during pregnancy. The more recent report by

Athanasiou et al. (1997) documented the occurrence of hemolytic anemia in a neonate whose mother had inhaled

naphthalene during the 28

week of pregnancy.

Metabolism and elimination

The in vivo and in vitro metabolism of naphthalene in mammalian systems has been extensively studied (U.S. EPA

1998). As many as 21 metabolites, including oxidized derivatives and conjugates, have been identified in the urine of

animals exposed to naphthalene (Horning et al., 1980; Wells et al., 1989; Kanekal et al., 1990). Factors that potentially

influence the relative proportions of individual metabolites include species, tissue type, and tissue concentration of

naphthalene (U.S. EPA, 1998). The initial step in naphthalene metabolism is catalyzed by cytochrome P-450

monooxygenases, and results in the formation of the arene epoxide intermediate 1,2-naphthalene oxide (Figure 4.1).

1,2-Naphthalene oxide can undergo spontaneous rearrangement to form naphthols (predominately 1-naphthol). The

resulting intermediates may be further metabolized by oxidation reactions, resulting in the formation of di-, tri-, and

tetrahydroxylated intermediates (Horning et al., 1980). Some metabolites may undergo conjugation with glutathione,

glucuronic acid, or sulfate (ATSDR, 1995; U.S. EPA, 1998). Glutathione conjugates undergo additional reactions to

form cysteine derivatives (thioethers). These cysteine derivatives may be further metabolized to mercapturic acids and

may be excreted in the bile (U.S. EPA, 1998). Alternative pathway of naphthol metabolism involves enzymatic

hydration by epoxide hydrolase (U.S. EPA, 1998).

Bieniek (1994) analyzed the excretion patterns of 1-naphthol in three groups of workers occupationally exposed to

naphthalene. The mean excretion rate for these workers was 0.57 mg/hour, with a calculated excretion half-life of

approximately 4 hours. The highest urinary levels of 1-naphthol were reported for workers in a naphthalene oil

distribution plant. Peak 1-naphthol levels were detected in urine collected one hour after finishing the shift.

Excretion in the feces represents a minor pathway for naphthalene, and the possibility of excretion via exhalation of the

unmetabolized compound has not been examined in available studies.

The INDEX project Final report

125

Figure 4.1: Proposed Pathways For Naphthalene Metabolism

Source: U.S. EPA (1998)

5. Health effects

Information is scarce regarding dose-response relationships for health effects in humans with acute, subchronic, or

chronic exposure by any route. Reported health effects associated with indoor exposures to naphthalene are limited to

its principal indoor source: mothballs.

Effects of short-term exposure

Reports that establish associations between naphthalene exposure and health effects in humans are restricted to

numerous reports of hemolytic anemia or cataracts following acute exposure or occupational exposure to naphthalene,

either by ingestion or by inhalation of naphthalene vapors, but these reports have not identified exposure levels

associated with these effects. A relationship appears to exist between an inherited deficiency in the enzyme, glucose 6-

phosphate dehydrogenase (G6PD), and susceptibility to naphthalene-induced hemolysis. Newborn infants also appear to

The INDEX project Final report

126

be susceptible to naphthalene-induced hemolysis presumably due to a decreased ability to conjugate and excrete

naphthalene metabolites.

Individuals with a G6PD genetic defect are prone to hemolysis after exposure to a variety of chemical oxidizing agents

including nitrates, nitrites, aniline, phenols (Dean et al. 1992), and naphthalene. Valaes et al. (1963) reported adverse

effects in 21 Greek infants exposed to naphthalene from clothing, diapers, blankets, and other items that had been stored

in contact with mothballs. Durations of exposure ranged from 1 to 7 days. Inhalation was identified as the primary route

of exposure because 19 of the 21 infants did not have dermal contact with the naphthalene-contaminated materials. A

total of 21 infants developed hemolytic anemia and two infants died from kernicterus, a severe neurological condition

that was thought to be a consequence of massive hemolysis. Ten of the 21 anemic children and 1 of the 2 infants that

died from naphthalene exposure had a genetic polymorphism that resulted in a deficiency in glucose-6-phosphate

dehydrogenase (G6PD). This enzyme helps to protect red blood cells from oxidative damage, and G6PD deficiency

makes the cells more sensitive to a wide variety of toxicants, including naphthalene. Eight adults and one child reported

gastrointestinal (nausea, vomiting, abdominal pain) and neurological (headache, malaise, confusion) symptoms after

exposure to large numbers of mothballs in their homes (Linick, 1983). The duration of exposure was not specified.

Testing at one home following the incident indicated an airborne naphthalene concentration of 105 mg/m

(20 ppb).

Symptoms abated after removal of the mothballs.

There are no data available on skin or respiratory sensitisation in humans. Acute (4-hour) inhalation exposure to

naphthalene induced necrosis of Clara cells in the epithelium of the proximal airways of the lungs of mice at exposure

levels as low as 53 mg/m

(10 ppm), but did not affect lung tissue in rats at concentrations as high as 526 mg/m

(100

ppm) (West et al. 2001). These results, and those from the chronic inhalation studies, show that mice are more

susceptible than rats to lung damage from inhaled naphthalene. However, there are no studies that have examined nasal

tissues for the development of lesions following acute inhalation exposure. A change to mouth breathing occurred in

rats during exposure to 410 mg/m

naphthalene, but no other effects on respiration were noted (Fait and Nachreiner

1985).

Hemolytic anaemia (potentially sensitive populations)

Increased sensitivity to naphthalene-induced hemolysis has been associated with reduced levels of glucose-6-phosphate

dehydrogenase (G6PD). This enzyme helps to protect red blood cells from oxidative damage, and G6PD enzyme

deficiency makes the cells more sensitive to a wide variety of toxicants, including naphthalene. Higher rates of inherited

G6PD deficiencies are found more often in defined subpopulations of males from Asian, Arab, Caucasian (of Latin

ancestry), African, and African-American ancestry than in other groups (U.S. EPA, 1987). Multiple forms of G6PD

deficiency have been identified in these subpopulations. The mildest forms are totally asymptomatic, while moderate

forms are associated with an adverse response to chemical stressors, including naphthalene. The most severe forms of

G6PD deficiency are associated with hemolytic anemia, even in the absence of external stressors (Beutler, 1991). The

overall prevalence of G6PD-deficiency in the United States is reported to be 5.2 to 11.5% (Luzzatto and Mehta, 1989).

There is no evidence of haemolytic anaemia in rodents.

Prevalence of G6PD enzyme deficiency in the EU: Polymorphic G6PD variants are those that have achieved a high

frequency in some populations. They represent balanced polymorphisms in which the benefit of inheriting the mutation

(probably resistance to malaria) counterbalances the disadvantage (susceptibility to hemolysis and neonatal icterus).

Generally, each population has its own characteristics mutations, although, as noted below, there are occasional

exceptions to this rule.

- Mediterranean Variants: G6PD deficiency is very prevalent in some Mediterranean countries. A gene frequency of

about 0.7 has been documented among Kurdish Jews. This is the highest incidence known in any group. Among

Greeks, Turks, Sardinians, Sephardic Jews and Italians, G6PD deficiency is also quite prevalent, but more commonly

with gene frequencies that range from 0.02 to 0.20. The situation regarding Mediterranean variants is in one respect

the reverse of the situation of the African variants. Here several different variants (e.g., G6PD “Sassari” and

“Cagliari”) were believed to be different on the basis of biochemical characteristics, but all seem to share the same

mutation at nt 563. Variants from other parts the world, thought to be unique - G6PD Dallas, Birmingham, and

Panama - proved to be G6PD Mediterranean 563T. Other mutations are found in the Mediterranean region as well.

G6PD Seattle844T (also described as G6PD Lodi and as G6PD “Modena”) and, as noted above, G6PD A- are

relatively common.

- African Variants: G6PD deficiency among Africans is relatively mild; red cells contain 10-15% residual enzyme that

is electrophoretically rapid. Accordingly it is designated G6PD A- to distinguish its mobility from the normal

enzyme, which is designated B. Among Afro- Americans the gene frequency of G6PD A- is about 11%.(11) It is

considerably higher in some parts of Africa.(12) Some Africans have an enzyme with the same rapid mobility

encountered among deficient individuals, but the activity of the enzyme is normal. This African enzyme is designated

The INDEX project Final report

127

G6PD A+, and it is known to be due to an A®G transition at nt 376 that predicts substitution of a negatively charged

aspartic acid for asparagine. Its gene frequency is in the range of 20-30%

Polymorphic G6PD variants that have been characterized at the DNA level have been summarized in Beutler, 1994,

Vulliamy et al., 1993 and Beutler et al., 1995.

Summary of short-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

EXP

Respiratory: Clara cell necrosis and decreased

Clara cell mass

(volume/surface area) in proximal

airways

Mouse; 4h West et al. 2001

ATSDR 2003;

no MRL

ECB (2003): In relation to haemolytic anaemia, the available data do not allow the identification of a NOAEL.

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

ATSDR (2003)

Agency for Toxic Substances and Disease Registry

Data are inadequate for deriving an acute-duration inhalation MRL for naphthalene. Data are restricted to a 14-day (6

hours/day, 5 days/week) range-finding study in B6C3F1 mice (NTP 1992a), which only examined hematologic end

points and did not histologically examine expected critical toxicity targets (lung and nasal cavity epithelial tissue) (NTP

1992a), and a study (West et al. 2001) with Swiss Webster mice and Sprague-Dawley rats, which involved single 4-

hour exposure periods. The more recent study, however, only histologically examined the lung and did not examine

nasal tissue. A comprehensive inhalation study involving an acute repeated exposure scenario and examining the other

critical target (the nose, based on the findings from chronic mouse and rat bioassays) is not currently available. Results

from such a study may be useful for deriving an acute-duration inhalation MRL for naphthalene.

Hemolysis is the best documented effect of acute naphthalene exposures in humans, but it has not been observed in

studied strains of rats (F344) or mice (CD-1, B6C3F1). Dose-response data for hemolysis from a susceptible animal

species (such as dogs or the Jackson Laboratory hemolytic anemia mouse) may be useful to obtain data that could be

used for considering changes to the acute-duration oral MRL. Data from both inhalation and oral exposure protocols

would be useful.

ECB (2003)

European Chemical Bureau - Institute for Health and Consumer Protection - European Union Risk Assessment Report

In relation to hemolytic anaemia, the available data do not allow the identification of a NOAEL nor the dose-response

characteristics for this endpoint. In the absence of such information, any significant body burden is considered to give

rise to concerns for human health. Exposure of infants to textiles (clothing/bedding) which have been stored for long

periods with naphthalene moth repellent raises significant concern. There is documented evidence for the development

of severe hemolytic anaemia resulting from such use, although there is no quantitative information available on the

level or duration of exposure to naphthalene in these cases.

Effects of long-term exposure (noncancer)

There are no epidemiological studies on the human health effects of naphthalene, and the only human information

available derives from a limited number of early case-reports which provide no quantitative data on the levels or

duration of exposure. A principal human health effect is haemolytic anaemia (see previous chapters) which in some

cases has been of marked severity in humans exposed by inhalation to naphthalene vapour and by ingestion to solid

naphthalene. Dermal exposure to the solid and vapour was also likely in these cases.

Animal studies reveal species differences in response to naphthalene. Haemolytic anaemia was noted in a dog following

oral dosing of 220 mg/kg/day for 7 days but not in rodents even with high/prolonged exposures. Cataract formation was

the principal effect seen in rats and rabbits following oral exposure to 700 and 1000 mg/kg/day, respectively in studies

ranging from 10-180 days, but this effect was not seen in mice with similar exposures. The lack of reliable reports of

The INDEX project Final report

128

cataracts in humans despite the widespread use of naphthalene and high dose accidental exposure, suggests that cataract

formation is unlikely to be a significant health effect in humans (ECB, 2003).

The few chronic animal studies that are available for naphthalene were conducted primarily to characterize its

carcinogenic potential effects. Noncancer endpoints reported in these studies are summarized below.

Adkins et al. (1986) investigated the impacts of less-than-lifetime inhalation exposures to naphthalene on rats. Groups

of 30 female A/J strain rats were exposed to 0, 53, or 158 mg/m

(0, 10, or 30 ppm) naphthalene vapors for 6 hours per

day, 5 days per week, for 6 months. All animals were sacrificed at the end of the exposure period and their lungs

excised and examined for tumors. No adverse noncancer effects on the lung were reported (U.S. EPA, 1998). Other

organs were not examined in study.

A chronic inhalation study of naphthalene toxicity was conducted in B6C3F

1 mice by NTP (1992a). Naphthalene

exposure concentrations in air were 0, 53, or 158 mg/m

(0, 10 ppm, or 30 ppm). The concentration of 53 mg/m

was

chosen because it was equal to the ACGIH TLV® for naphthalene, while the 158 mg/m

concentration was chosen

because it was one-half the air saturation concentration. The control and low-exposure groups consisted of 75 mice of

each sex, while the high-exposure group consisted of 150 mice of each sex. Exposure was for 6 hours per day, 5 days

per week, for 2 years. Comprehensive histopathological evaluations were performed on all control and high-exposure

mice, and on all low-exposure mice that died or were sacrificed during the first 21 months of exposure. The original

study plan called for 50 animals per sex to be exposed for 2 years, and 5 animals per sex to be sacrificed for

hematological evaluations at 14 days, 3, 6, 12, and 18 months. However, as a result of excessive mortality in the control

males, only the 14-day hematological evaluation was conducted. All of the remaining animals were incorporated into

the two-year study.

A statistically significant decrease in survival was noted in the male control group. This phenomenon was attributed to

the frequent fighting that occurred among the control group mice. In contrast, the exposed groups tended to huddle

together during exposure periods, and fought less. Statistically significant increases were seen in several types of

noncancer respiratory tract lesions in both exposed groups. The observed responses included chronic lung

inflammation, chronic nasal irritation with hyperplasia of the nasal epithelium, and metaplasia of the olfactory

epithelium. The authors of the study described the lung lesions as a chronic inflammatory response with granuloma.

These lesions consisted of “focal intra-alveolar mixed inflammatory cell exudates and interstitial fibrosis.” The more

advanced lesions consisted primarily of “large foamy macrophages sometimes accompanied by giant cells.”

No changes in hematological parameters were seen among the exposed animals at 14 days. No cataract formation was

observed after 2 years of exposure. Histopathological examination did not reveal treatment-related effects on the liver,

gastrointestinal system, reproductive system, brain, or any other organs. The results of this study have been interpreted

by U.S. EPA (1998) to support a chronic LOAEL for nasal and respiratory irritation of 53 mg/m

NTP (2000) conducted a chronic inhalation study in F344/N rats. Male and female rats (49/sex/dose) were exposed to

naphthalene vapor concentrations of 0, 53, 158, and 316 mg/m

(0, 10, 30, and 60 ppm) for 6 hours per day plus T

(the

theoretical time to achieve 90% of the target concentration in the vapor chamber: 12 minutes), 5 days per week for 105

weeks. Additional groups of rats were similarly exposed for up to 18 months for evaluation of toxicokinetic parameters.

Dose calculations were based upon model estimates of the amount of naphthalene inhaled by rats at the exposure

concentrations used in the two-year study, the total amount of naphthalene metabolized following a six-hour exposure

period (21% to 31% of inhaled naphthalene), and average weights of 125 grams (male rats) and 100 grams (female

rats). Because essentially all of the naphthalene that is absorbed into the bloodstream is metabolized, the total amount of

naphthalene metabolized was assumed to represent the internalized dose to rats from the exposure concentrations used

in this two-year study. The estimated daily doses determined by this method were 0, 3.6, 10.7, 20.1 mg/kg-day for

males, and 0, 3.9, 11.4, and 20.6 mg/kg-day for females.

Rats were clinically examined twice daily and findings were recorded every four weeks beginning at week 4 and every

two weeks beginning at week 92. Body weights were recorded at study initiation, every four weeks beginning at week

4, and every two weeks beginning at week 92. Full necropsies and complete histopathologies were performed on all

core study animals.

There were no clinical findings related to naphthalene exposure from the two-year inhalation study. The mean body

weights all exposed groups of male and female rats were similar to those observed in the appropriate control chamber

group. No significant difference in survival rate was observed for any exposed group when compared to the chamber

control. The mean body weights of female rats were generally similar to the body weights of the control group, while

the mean body weights of naphthalene-exposed male rats were generally less than the chamber control for all exposed

groups.

The INDEX project Final report

129

Although naphthalene is a known cataractogen and ocular irritant, no naphthalene-related cataractogenic effects or

ocular abnormalities were observed in rats during this study. Treatment-related non-neoplastic lesions were observed in

the nose and lungs of male and female rats. The incidence and average severity of nasal lesions (glands, goblet cells,

respiratory epithelium and olfactory epithelium) are summarized in Table 5.1. The incidences of these lesions were

significantly greater than those in the chamber controls for all male and female exposed groups, with the exception of

squamous metaplasia of glands in male and female rats in the 53 mg/m

exposure groups (NTP, 2000). In general, the

severities of olfactory epithelial and glandular lesions increased with increasing exposure concentrations.

Two noteworthy type of lesions occurred in the lungs of exposed rats: alveolar epithelial hyperplasia and minimal

chronic inflammation. Female rats in all exposure groups had increased incidences of alveolar epithelial hyperplasia

when compared to the chamber control (chamber control: 4/49, 53 mg/m

10ppm: 11/49, 158 mg/m

: 11/49, 316 mg/m

9/49). This effect reached statistical significance in the 53 and 158 mg/m

exposure groups. The incidences of alveolar

epithelial hyperplasia in male rats (chamber control: 23/49, 53 mg/m

: 12/49, 158 mg/m

: 9/48, 316 mg/m

: 16/49) were

significantly decreased in the 53 and 158 mg/m

exposure groups. The incidences of minimal chronic inflammation of

the lung were increased in males and females exposed to naphthalene. This lesion is characterized by small focal

interstitial and intra-alveolar collections of macrophages, neutrophils, and lymphocytes and minimal interstitial fibrosis.

As noted by the NTP study authors, foci of minimal inflammation are common in chamber control rats (as evident in

this study). Therefore, this change could not be confidently related to naphthalene exposure.

The study conducted by NTP (2000) identified an estimated inhalation LOAEL of 3.6 mg/kg-day based on the

occurrence of nasal lesions in male rats in the 53 mg/m

exposure group. A NOAEL was not identified in this study.

The 53 mg/m

concentration associated with the LOAEL corresponds to the threshold limit value for naphthalene

(ACGIH, 2000).

Table 5.1: Nonneoplastic and Neoplastic Lesions of the Nose in Male and Female F344/N Rats

Exposed to Naphthalene 6 Hours/Day, 5 Days/Week for 105 Weeks (NTP, 2000)

Concentration 0 mg/m

53 mg/m

(10 ppm)

158 mg/m

(30 ppm)

316 mg/m

(60 ppm)

Lesion M F M F M F M F

Nonneoplastic lesions

Olfactory epithelium

Hyperplasia 0/49 0/49 48/49 48/49 45/48 48/49 46/48 43/49

Atrophy 0/49 0/49 49/49 49/49 48/48 49/49 47/48 47/49

Chronic inflammation 0/49 0/49 49/49 47/49 48/48 47/49 48/48 45/49

Hyaline degeneration 0/49 13/49 46/49 46/49 40/48 49/49 38/48 45/49

Respiratory epithelium

Hyperplasia 0/49 0/49 21/49 18/49 29/48 22/49 29/48 23/49

Squamous metaplasia 0/49 0/49 15/49 21/49 23/48 17/49 18/48 15/49

Hyaline degeneration 0/49 8/49 20/49 33/49 19/48 34/49 19/48 28/49

Goblet cell hyperplasia 0/49 0/49 25/49 16/49 29/48 29/49 26/48 20/49

Gland hyperplasia 0/49 0/49 49/49 48/49 48/48 48/49 48/48 42/49

Gland squamous metaplasia 0/49 0/49 3/49 2/49 14/48 20/49 26/48 20/49

Neoplastic lesions

Respiratory epithelial adenoma 0/49 0/49 6/49 0/49 8/48 4/49 15/48 2/49

Olfactory epithelial

neuroblastoma

0/49 0/49 0/49 2/49 4/48 3/49 3/48 12/49

Source: Adapted from NTP Technical Report on the Toxicology and Carcinogenesis Studies of Naphthalene in Rats (Inhalation Studies), Table 6

(NTP, 2000) (M=male, F=female)

Naphthalene administered by inhalation at concentrations of 10, 30, or 60 ppm for 6 hours per day, 5 days per week for

105 weeks caused nonneoplastic and neoplastic effects in nasal respiratory and olfactory regions of male and female

F344/N rats. Non-neoplastic nasal effects were characterized by an increase in the incidence and severity of a complex

group of lesions, including atypical hyperplasia, atrophy, chronic inflammation, and hyaline degeneration of olfactory

epithelium; hyperplasia, squamous metaplasia, hyaline degeneration, and goblet cell hyperplasia of the respiratory

epithelium; and hyperplasia and squamous metaplasia of mucosal glands. Neoplastic effects were characterized by the

induction of two types of rare primary nasal tumors, olfactory neuroblastomas and respiratory epithelial adenomas. The

incidences of olfactory neuroblastomas in males at 0 ppm, 10 ppm, 30 ppm, and 60 ppm were, respectively, 0%, 0%,

8%, and 6%, whereas in females they were 0%, 4%, 6%, and 24%. The incidences of respiratory epithelial adenomas in

The INDEX project Final report

130

males at 0 ppm, 10 ppm, 30 ppm, and 60 ppm were, respectively, 0%, 12%, 17%, and 31% and in females 0%, 0%, 8%,

and 4%. The olfactory neuroblastomas and respiratory epithelial adenomas were considered carcinogenic effects related

to naphthalene exposure based on their relatively high incidence in exposed rats, their absence in concurrent control rats

and NTP historical controls, and their rare spontaneous occurrence in rats of any strain (Abdo et al., 2001; Long et al.,

2003).

In a well conducted unpublished study, groups of 10 male and 10 female rats were exposed nose only for 6 hours/day, 5

days a week for 13 weeks to 0, 11, 53 or 305 mg/m

(0, 2, 10 or 58 ppm) vapourised naphthalene (Huntingdon Research

Centre, 1993a). A gross pathological examination was carried out on a wide range of tissues and a microscopic

examination was carried out on a range of tissues including the lungs, liver, kidneys, adrenals, testes, eyes and optic

nerve. Prior to terminal sacrifice, samples of blood were taken from all rats for haematological and clinical chemistry

evaluation. In high dose animals body weight gain was reduced by 43% and 34% in males and females, respectively and

was associated with reduced food consumption. There were no toxicologically significant haematological or clinical

chemistry findings observed. Similarly, no significant changes were noted in organ weight or gross pathology.

Microscopic analysis of the nasal epithelium revealed treatment-related effects at all dose levels. The severity of the

effects was dose-related. At the highest exposure level (305 mg/m

) changes included erosion of the olfactory

epithelium, hyperplasia of basal cells in the olfactory epithelium and loss of Bowmans' glands. At the lowest exposure

level (11 mg/m

) changes in olfactory epithelium were less marked but included slight disorganisation, mild erosion (in

one rat), minimal atrophy, rosette formation (an attempt at proliferative repair by the olfactory neuroepithelium),

occasional degenerate cells, loss of Bowmans' glands and minimal hyperplasia. There were no treatment related effects

observed in the lungs or nasal respiratory epithelium at this dose. There were no observed changes in the nasal passages

of control animals. In one low dose rat there was evidence of squamous metaplasia of the respiratory epithelium,

however as this lesion was not seen in the other rats at higher doses this lesion was not considered toxicologically

significant. The effects at 10 mg/m

were generally minimal in severity and seen in only small numbers of animals, and

therefore appear to represent the low end of the dose-response curve for nasal effects. Overall, signs of damage to the

olfactory epithelium were seen at all doses down to 11 mg/m

(2 ppm), and a NOAEL cannot be identified for local

effects.

In a well conducted unpublished study, groups of 5 male and 5 female rats were exposed nose only for 6 hours/day, 5

days a week for 4 weeks to 0, 5, 16, 53, 153 or 374 mg/m

(0, 1, 3, 10, 29 or 71 ppm) vapourised naphthalene

(Huntingdon Research Centre, 1993b). Investigations were similar to the 13-week study performed in the same

laboratory. Results were similar to those observed in the 13-week study. High dose animals showed approximately a

50% reduction in body weight gain associated with reduced food consumption. There was no evidence of systemic

toxicity. Local effects were observed with signs of proliferative repair in the nasal olfactory epithelium changes

observed at all doses down to 5 mg/m

(1 ppm), and therefore a NOAEL for local effects cannot be identified.

For both the 4 and 13-week studies the mechanism by which the observed effects in the olfactory nasal epithelium arise

is unclear, although the effects may be mediated by locally produced metabolite(s) of naphthalene. The relevance of

these effects to human health is uncertain, as there may be significant species differences in local metabolism. However,

there is no evidence to indicate that these effects are not relevant to human health.

Reproductive and developmental toxicity

No information is available on both the reproductive and developmental effects of naphthalene in humans, although the

occurrence of hemolytic anemia in the neonates of anemic, naphthalene-exposed mothers demonstrates that naphthalene

and/or its metabolites can cross the placental barrier (Anziulewicz et al. 1959; Zinkham and Childs 1957, 1958).

Animal studies involving naphthalene exposure during gestation reported no reproductive effects in rabbits

administered doses of up to 120 mg/kg/day by gavage or in rats given doses of up to 450 mg/kg/day, although doses of

150 mg/kg/day and greater were maternally toxic to rats. There was a decrease in the number of live mouse pups per

litter with a dose of 300 mg/kg/day given during gestation (Plasterer et al. 1985) and in vitro studies of naphthalene

embryotoxicity in the presence of liver microsomes support the concept that naphthalene metabolites may be harmful to

the developing embryo (Iyer et al. 1991).

No exposure-related lesions in reproductive tissues were found in intermediate-duration oral exposure studies in rats

(NTP 1980b) and mice (NTP 1980a) or in chronic inhalation studies in rats (Abdo et al. 2001; NTP 2000) or mice (NTP

1992a). One- or two-generation reproductive toxicity studies evaluating reproductive performance variables in male and

female animals exposed to naphthalene are not available.

The INDEX project Final report

131

Studies of the developmental effects of orally administered naphthalene in rats (NTP 1991), mice (Plasterer et al. 1985),

and rabbits (NTP 1992b; PRI 1985, 1986) have been negative, except for a slight nonsignificant increase in fused

sternebrae in female rabbit pups from a small number of litters at doses of 80 and 120 mg/kg/day (NTP 1992b). No

developmental toxicity studies involving inhalation or dermal exposure to naphthalene are available.

Summary of long-term exposure effect levels (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

EXP

9.4

ADJ

Respiratory; Nasal inflammation, olfactory

epithelial metaplasia, and respiratory epithelial

hyperplasia

Mice, 105w

NTP, 1992a OEHHA 2000;

REL: 0.009;

EPA-IRIS 1998 ;

RfC: 0.003

EXP

9.4

ADJ

1.2

HEC

Respiratory; nonneoplastic lesions in nasal

olfactory epithelium and respiratory epithelium

in two species

Combined studies:

Mice 2y - Rats 105w -

Rats 2y

NTP, 1992a;

NTP, 2000;

Abdo et al., 2001

ATSDR 2003;

MRL: 0.004

EXP

0.94

ADJ

Respiratory; proliferative repair in the nasal

olfactory epithelium

Rats, 4w Huntingdon Research

Centre, 1993b

(unpublished)

ECB 2003

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (2000)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Chronic Reference Exposure Level (REL):

Study NTP (1992a)

Study population B6C3F1 mice (75 or 150/group/sex)

Exposure method Discontinuous whole-body inhalation exposures to 0, 10, or 30 ppm

naphthalene vapor

Critical effects Nasal inflammation, olfactory epithelial metaplasia, and respiratory

epithelial hyperplasia

LOAEL 10 ppm (96% incidence for males and 100% incidence for females)

NOAEL Not observed

Exposure continuity 6 hours/day for 5 days/week

Average experimental exposure 1.8 ppm (10 ppm x 6/24 x 5/7) for LOAEL group

Exposure duration 104 weeks

Subchronic uncertainty factor 1

LOAEL uncertainty factor 10

Interspecies uncertainty factor 10 (see below)

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 1000

Inhalation reference exposure level 0.002 ppm (2 ppb, 0.009 mg/m

, 9 µg/m

)

The NTP study was chosen for the REL derivation since it is the only available lifetime animal inhalation bioassay and

because no adequate epidemiological studies of long-term human exposure are available. The study was judged to be of

adequate study design. The complete lack of nasal effects among control animals and the nearly total effect among

animals exposed at 2 different concentrations strongly indicates a causal relationship between naphthalene exposure and

nasal effects. The effects seen are consistent with those reported among exposed workers, who developed

rhinopharyngolaryngitis or laryngeal carcinoma (Wolf, 1978). However, the hematological effects observed in humans

have not been reported in laboratory animals, which raises the possibility that humans may be significantly more

sensitive to naphthalene.

The most important limitation of the study is that the lowest concentration tested caused adverse effects in most (≥96%)

of the animals tested. Thus the study amply demonstrates the risk of lifetime exposures to 53 mg/m

(10 ppm), but is

uninformative regarding the concentration-response relationship at lower concentrations. Only a general assumption can

be drawn on the magnitude of uncertainty factor needed to predict a concentration at which adverse effects would most

likely not be observed. Lacking specific guidance or relevant research for this situation, the default 10-fold factor was

applied. U.S. EPA also used the NTP study to develop its RfC of 3 mg/m

with slightly different assumptions and a

cumulative uncertainty factor of 3000 (U.S. EPA, 2000). OEHHA followed the U.S. EPA precedent in using an

The INDEX project Final report

132

intraspecies UF of 10 for naphthalene, rather than using the HEC/RGDR approach. According to U.S. EPA (2000),

because of its low water solubility and low reactivity, naphthalene-related effects on the nasal epithelium are expected

to result following absorption of naphthalene and its metabolism to reactive oxygenated metabolites, not from direct

contact. This is supported by data on naphthalene metabolism indicating that toxic effects on the respiratory tract are

due to a naphthalene metabolite that may be formed either in the liver or in the respiratory tract. Necrosis of bronchial

epithelial (Clara) cells in mice and necrosis of olfactory epithelium in mice, rats, and hamsters occur following

intraperitoneal injection of naphthalene. The nasal effects from inhalation exposure to naphthalene were considered to

be extra-respiratory effects of a category 3 gas (U.S. EPA, 1994). The assumption is made that nasal responses in mice

to inhaled naphthalene are relevant to humans; however, it is uncertain that the RfC for naphthalene based on nasal

effects will be protective for hemolytic anemia and cataracts, the more well-known effects from naphthalene exposure

in humans.

U.S.EPA - IRIS (1998)

Integrated Risk Information System

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC)

Critical effect Experimental

doses*

UF MF RfC

Nasal effects: hyperplasia and metaplasia in respiratory

and olfactory epithelium, respectively

Chronic mouse inhalation study

NTP, 1992a

NOAEL: None

LOAEL(HEC): 9.3

mg/m

3000 1 0.003 mg/m

*Conversion Factors and Assumptions -- Following the Category 3 guidance (U.S. EPA, 1994), experimental exposure concentrations of

0, 10, and 30 ppm were converted to 0, 52, and 157 mg/m

, respectively; adjusted to a continuous exposure basis in mg/m

(6/24 hr ×

5/7 days) equals mg/m

× 0.1786: 0, 9.3, and 28 mg/m

. Because the blood:gas (air) coefficients for naphthalene were not available, the

default ratio of 1 was used and the values for the LOAEL(HEC) were 0, 9.3, and 28 mg/m

. Scenario -- The LOAEL human equivalent

concentration (HEC) was calculated for an extrarespiratory effect for a category 3 gas. Since the b:a lambda for humans (h) is unknown,

a default value of 1.0 is used for this ratio. LOAEL(HEC) × [b:a lambda(animal)/b:a lambda(human)] = 9.3 mg/m

The NTP (1992a) study was given medium confidence because adequate numbers of animals were used, and the

severity of nasal effects increased at the higher exposure concentration. However, the study produced high mortality, (<

40% survival in the male control group due to wound trauma and secondary lesions resulting from increased fighting).

Also, hematological evaluation was not conducted beyond 14 days. The database was given a low-to-medium

confidence rating because there are no chronic or subchronic inhalation studies in other animal species, and there are no

reproductive or developmental studies for inhalation exposure. In the absence of human or primate toxicity data, the

assumption is made that nasal responses in mice to inhaled naphthalene are relevant to humans; however, it cannot be

said with certainty that this RfC for naphthalene based on nasal effects will be protective for hemolytic anemia

and cataracts, the more well-known human effect from naphthalene exposure. Medium confidence in the RfC

follows.

Dose conversion: Because of its low water solubility and low reactivity, naphthalene-related effects on the nasal

epithelium are expected to result following absorption of naphthalene and metabolism to reactive oxygenated

metabolites, rather than being a result of direct contact. This hypothesis is supported by data on naphthalene metabolism

indicating that toxic effects on the respiratory tract are due to a naphthalene metabolite that may be formed either in the

liver or in the respiratory tract. For example, necrosis of bronchial epithelial (Clara) cells in mice (O'Brien et al., 1985,

1989; Tong et al., 1981) and necrosis of olfactory epithelium in mice, rats, and hamsters (Plopper et al., 1992) occur

following intraperitoneal injection of naphthalene. The nasal effects from inhalation exposure to naphthalene were

considered to be extra-respiratory effects of a category 3 gas, as defined in the U.S. EPA guidance for deriving RfCs

(U.S. EPA, 1994). Following this guidance, experimental exposure concentrations were adjusted to a mg/m

basis (0,

52, and 157 mg/m

), adjusted to a continuous exposure basis (mg/m

× 6h/24h × 5d/7d = mg/m

× 0.1786: 0, 9.3, and 28

mg/m

), and converted to human equivalent concentrations (HECs) by multiplying the adjusted concentrations by the

ratio of mouse:human blood/gas partition coefficients. Because the blood/gas coefficients for naphthalene were not

available, the default ratio of 1 was used.

Dose-response modeling: Whereas the data from the NTP (1992a) study show nasal effects to be the most sensitive

effects from chronic inhalation exposure to naphthalene, they provide no indication of the shape of the dose-response

curve because the incidence of nasal lesions at the lowest exposure level was 100% in females and nearly 100% in

males. In this case, application of a BMD approach, in which quantal mathematical models are fit to the incidence data

for nasal effects, does not sensibly assist in extrapolating to a NOAEL, and a NOAEL/LOAEL approach was taken for

deriving an RfC for naphthalene.

The INDEX project Final report

133

ATSDR (2003)

Agency for Toxic Substances and Disease Registry

Derivation of the Minimal Risk Level (MRL):

Dose and end point used for MRL derivation: The lowest exposure level in both studies (NTP, 1992a; 2000), 53 mg/m

(10 ppm), was a LOAEL in both sexes of both species for nonneoplastic lesions in nasal olfactory epithelium and

respiratory epithelium. Applying EPA inhalation dosimetry, a human equivalent LOAEL of 1.2 mg/m

(0.2 ppm), based

on the rat LOAEL, was selected as the point of departure for the chronic inhalation MRL.

Modifying Factors used in MRL derivation:

10 for use of a LOAEL

3 for extrapolation from animals to humans with dosimetric adjustment

10 for human variability

Total Uncertainty Factor = 10x3x10=300

Determination of the human equivalent dose:

10 ppm x 6 hours/24 hours x 5 days/7 days x 128.18/24.45 = 9.4 mg/m

(duration-adjusted LOAEL for nasal effects in

rats or mice)

Following EPA (1994) Methods for Derivation of Inhalation Reference Concentrations and Application of Inhalation

Dosimetry, equations for a category 1 gas producing nasal effects were used to derive human equivalent concentrations:

HEC=Animal Concentration x RGDR

;

RGDR

= regional gas dose ratio in the extrathoracic (ET) region

= (Dose

)A/(Dose

)H = [minute volume/ETsurface area]A÷[minute volume/ETsurface area]H;

Reference minute volumes (L/min): 13.8 human, 0.137 rat, 0.0368 mouse;

Reference ET surface area (cm2 ): 200 human, 15 rat, 3 mouse;

RGDR

(Rat to Human)=[0.137/15]÷[13.8/200]=0.132; LOAEL

HEC

=duration-adjusted LOAEL x 0.132=9.4 mg/m

0.132=1.2 mg/m

RGDR

(Mouse to Human)=[0.0368/3]÷[13.8/200]=0.178; LOAEL

HEC

=duration-adjusted LOAEL x 0.132=9.4 mg/m

x 0.178=1.7 mg/m

Using public health protection reasoning, the LOAEL

HEC

based on the rat data was selected as the point of departure for

the chronic inhalation MRL.

ECB (2003)

European Chemical Bureau - Institute for Health and Consumer Protection - European Union Risk Assessment Report

The hazardous properties of naphthalene have been evaluated to the extent that the minimum data requirements

according to Article 9(2) of Council Regulation EEC No.793/93 have been met. The key health effects of haemolytic

anaemia, repeated inhalation toxicity and carcinogenicity have been identified. For haemolytic anaemia, it is not

possible to identify a NOAEL from the available data. For repeated inhalation toxicity, the key effect of concern is local

damage to the upper respiratory tract. The available data do not allow the identification of a NOAEL; in a 28-day study

in rats (Huntingdon Research Centre, 1993b)., damage to nasal olfactory tissue occurred at 5 mg/m

, the lowest

concentration level used. Therefore, the LOAEL of 5 mg/m

from this study will be used in the risk characterisation for

repeated inhalation toxicity, including carcinogenicity. This experimental exposure concentration has been further

adjusted by the authors to a continuous exposure basis: 5 mg/m

× 6h/24h × 5d/7d = 0.9 mg/m

Genotoxicity

Naphthalene has given reproducible negative results in bacterial mutation assays, and was negative in an in vitro UDS

assay. It was however found to be clastogenic in CHO cells in the presence but not the absence of S9. Two in vitro

studies using CHO cells and human peripheral lymphocytes were negative for induction of SCE. Naphthalene was

found to be negative in two in vivo bone-marrow micronucleus tests and an in vivo rat liver UDS study. Overall, the

balance of evidence indicates that naphthalene is not genotoxic (ECB, 2003).

The available data suggest that genotoxic action by the naphthalene metabolite, 1,2-naphthoquinone, is plausible and

that the mutagenic/genotoxic potential of naphthalene and its metabolites may be weak. Assays of possible genotoxic

The INDEX project Final report

134

action in sensitive target tissues of naphthalene in rodents (lung and nasal epithelial tissue), however, are not available

(ATSDR, 2003).

Carcinogenic potential

The only studies of cancer in humans exposed to naphthalene are two case series reports of cancer; one report of four

laryngeal cancer cases (all of whom were smokers) among workers in a naphthalene purification plant in East Germany,

and another report of 23 cases of colorectal carcinoma admitted to a hospital in Nigeria. NTP (2002), EPA (2002), and

IARC (2002) concur that these studies provide inadequate evidence of naphthalene carcinogenicity in humans. No

cohort mortality or morbidity studies or case-control studies examining possible associations between naphthalene

exposure and increased risk of cancer (or other health effects) are available.

There are two comprehensive chronic-duration inhalation toxicology and carcinogenicity studies of naphthalene in

animals, one in rats (Abdo et al. 2001; NTP 2000) and one in mice (NTP 1992a). These studies identify respiratory

tissues as the most sensitive toxicity targets of chronic-duration exposure to inhaled naphthalene in animals: in

particular neuroblastoma of the nasal olfactory epithelium was observed in rats and neoplastic lesions in the lungs of

mice. Exposure related lesions in other tissues were not found in these studies. NTP (2002) and IARC (2002) concurred

that these studies provide sufficient evidence of naphthalene carcinogenicity in animals. Overall, the proposed

mechanism of carcinogenic action IARC (2002) is that the higher rates of metabolism of naphthalene in mice lead to

cytotoxic metabolites in the lung, causing increased cell turnover and tumours. The absence of lung tumours in rats is

entirely consistent with this mechanism. The maximal rates of metabolism measured in human lung microsomes are

about 10–100 times lower than those in mice.

The Office of Environmental Health Hazard Assessment (OEHHA) on march 2004 adopted a unit risk value for

naphthalene of 3.4 x 10

-5

(µg/m

)

-1

and slope factor of 1.2 x 10

-1

(mg/kg-day)

-1

. These values are based on data for

incidence of nasal respiratory epithelial adenoma and nasal olfactory epithelial neuroblastoma in male rats.

Interactions with other chemicals

Schmeltz et al. (1978) suggested that it is likely that certain naphthalenes compete with benzo[a]pyrene (BaP) for the

same enzyme sites, resulting in alteration of the BaP metabolic pathway and decreased production of the active BaP

metabolite. This hypothesis is consistent with the observation that benzo(a)pyrene hydroxylase is inhibited by

naphthalene (Shopp et al. 1984). Dermal application of the naphthalene mixture did not induce tumors in the absence of

BaP. The results of these studies were not analyzed statistically.

Several studies have been conducted to assess factors that influence the toxicity of naphthalene. For the most part, these

studies have evaluated the effects of mixed function oxidase activity (MFO) and alterations in glutathione levels on

pulmonary and ocular toxicities. The effects of cyclooxygenase activity, antioxidants, and epoxide hydrolase inhibitors

on the cataractogenic effect of naphthalene have also been evaluated. The administration of MFO inhibitors (SKF-

525A, metyrapone) and antioxidants (caffeic acid and vitamin E) decreased ocular toxicity in mice (Wells et al. 1989).

Use of ALO1576, an inhibitor of the enzyme aldose reductase, prevented cataract formation in both in vivo and in vitro

studies (Xu et al. 1992a, 1992b). On the other hand, naphthalene-induced cataracts were enhanced by pretreatment with

a MFO inducer (phenobarbital) and a glutathione depletor (diethyl maleate) (Wells et al. 1989). Pulmonary damage was

decreased by prior treatment with a MFO inhibitor (piperonyl butoxide), but enhanced by prior treatment with a

glutathione depletor (diethyl maleate) (Warren et al. 1982). For the most part, these studies support the role for mixed

function oxidase activity and glutathione conjugation in naphthalene-induced pulmonary and ocular lesions.

Odour perception

Source: Amoore and Hautala (1983)

Odour threshold: 0.42 mg/m3 (0.08 ppm)

Source: Devos et al. (1990)

Odour threshold: 7.5 µg/m3

Source: Canadian Centre for Occupational Health and Safety - CCOHS

Odour is perceptible at 1.6 mg/m3 to 4.7 mg/m3 (0.3 to 0.9 ppm)

The INDEX project Final report

135

Source: New Jersey Department of health and senior services - Hazardous Substance Fact Sheets

Odour threshold: 0.2 mg/m3 (0.038 ppm).

Summary of Naphthalene Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Rapid dermal absorption after skin contact

Toxicokinetics: Rapidly absorbed into the blood. Distributes to adipose tissue, breast milk and fetus.

Health effect levels of short- and long-term exposure (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

EXP

Respiratory: Clara cell necrosis and decreased

Clara cell mass (volume/surface area) in

proximal airways

Mouse; 4h West et al. 2001

ATSDR 2003;

no MRL

Long-term exposure

EXP

9.4

ADJ

Respiratory; Nasal inflammation, olfactory

epithelial metaplasia, and respiratory epithelial

hyperplasia

Mice, 105w

NTP, 1992a OEHHA 2000;

REL: 0.009;

EPA-IRIS 1998 ;

RfC: 0.003

EXP

9.4

ADJ

1.2

HEC

Respiratory; nonneoplastic lesions in nasal

olfactory epithelium and respiratory epithelium

in two species

Combined studies:

Mice 2y - Rats 105w -

Rats 2y

NTP, 1992a;

NTP, 2000;

Abdo et al., 2001

ATSDR 2003;

MRL: 0.004

EXP

0.94

ADJ

Respiratory; proliferative repair in the nasal

olfactory epithelium

Rats, 4w Huntingdon Research

Centre, 1993b

(unpublished)

ECB 2003

Carcinogenicity: IARC: 2B ; U.S.EPA: C ; ACGIH: A4 ; Inadequate evidence in humans

Genotoxicity: plausible for 1,2-naphthoquinone (naphthalene metabolite)

Odour threshold: 0.0075 mg/m³ (Devos), 0.2 mg/m³ (HSFS), 0.42 mg/m³ (Amoore &Hautala)

Susceptible population:

- Newborn infants susceptible to naphthalene-induced anaemia due to poor metabolism.

- Inherited G6PD enzyme deficiency (with mild, moderate, and severe polymorphic variants). G6PD deficiency is very

prevalent in some Mediterranean countries. Among Greeks, Turks, Sardinians, Sephardic Jews and Italians, G6PD

deficiency with gene frequencies that range from 0.02 to 0.20.

Remarks: Acute inhalatory exposure seems to be typically associated with wear of naphthalene contaminated clothing.

The INDEX project Final report

136

6. Risk Characterization

Health hazard evaluation of short-term exposure

Case reports of individuals (primarily infants) exposed to naphthalene via inhalation, dermal contact, or a combination

of both exposure routes (e.g. textiles which have been stored with naphthalene moth repellent) point to hemolytic

anemia and its sequelae as the most commonly manifested effects in humans following exposure at concentrations that

exceed average environmental levels. Sensitive populations include individuals deficient in the enzyme glucose-6-

phosphate dehydrogenase (G6PD). Also, newborns and infants are thought to be more susceptible than older people

because hepatic enzymes involved in conjugation and excretion of naphthalene metabolites are not well developed after

birth. There are no studies that have specifically examined the influence of age on naphthalene toxicokinetic capabilities

in humans. Although the availability of such studies may increase the understanding of the specific physiological basis

for the apparent susceptibility of newborns, they are unlikely to be conducted. Experiments examining the most

sensitive targets in animals are likely surrogates. In addittion, there is no evidence of haemolytic anaemia in rodents.

Consequently, in relation to naphthalene induced anaemia, available data do not allow the identification of a NOAEL

nor the dose-response characteristics for this endpoint. Naphthalene has a distinct odour of mothbolls or coal tar, with a

reported minimum thresholds at 7.5 µg/m³.

Health hazard and cancer risk evaluation of long-term exposure

No reliable human toxicity data were found for subchronic or chronic exposure to naphthalen.

In the NTP study (1992, chosen by U.S.EPA, ATSDR and OEHHA as a key-study for the derivation of exposure limits)

mice were exposed discontinuously to whole-body naphthalene inhalation at 0, 53 and 158 mg/m

. The complete lack of

nasal effects among control animals and the nearly total effect among animals exposed at 2 different concentrations

strongly indicates a causal relationship between naphthalene exposure and nasal effects. The effects seen (Toxicological

endpoint: Respiratory; Nasal inflammation, olfactory epithelial metaplasia, and respiratory epithelial hyperplasia) are

consistent with those reported among exposed workers, who developed rhinopharyngolaryngitis or laryngeal carcinoma

and the assumption is made that nasal responses in mice to inhaled naphthalene are relevant to humans.

Taking the lower concentration as a LOAEL and adjusting this value in order to allow for conversion from

discontinuous to a continuous pattern of exposure a concentration of 10 mg/m

is obtained. Incorporating factors of 10

for interspecies variability, 10 for interindividual variation and 10 for use of a LOAEL rather than a NOAEL this results

in an Exposure Limit of 0.01 mg/m

(see also Table 6.1). However, it cannot be said with certainty that this EL for

naphthalene based on nasal effects will be protective for hemolytic anemia, the more well-known human effect from

naphthalene exposure.

As stated in the ECB risk assessment report, in relation to carcinogenicity, naphthalene is not genotoxic in vivo and thus

the tumours are considered to arise via a non-genotoxic mechanism. The tumours develop only at the sites where non-

neoplastic inflammatory changes also occur. Thus, it is considered that the development of the nasal tumours in the rat

is a consequence of chronic tissue injury, for which an identifiable threshold of effect will exist, although currently not

identified. Given that the underlying mechanism for the development of nasal tumours in the rat is considered to be the

chronic inflammatory damage seen at this site, it follows that prevention of local tissue damage would prevent

subsequent development of tumours.

Result

With regard to the general population a long-term exposure limit has been set at 10 µg/m

, according to the assumption

that nasal effects observed in mice are consistent with the health effects reported among exposed workers. Available

exposure data indicate that, on average, the European population is exposed at naphthalene levels 10 times lower than

this EL, although an important exception resulted from a survey held in Athens, were levels exceeding the EL were

measured in nearly all residences. It is assumed that increased residential exposures originate from the use of

naphthalene based moth-repellents, a widespread use occuring in certain countries of the Mediterranean area.

An important source of uncertainty in establishing safe exposure limits is the potentially greater sensitivity of certain

subpopulations to naphthalene toxicity, including infants and children, neonates, fetuses, and individuals deficient in

glucose-6-phosphate dehydrogenase (G6PD), the prevalence of this inherited deficiency reported to be 2 to 20% in

defined Mediterranean subpopulations. In these latter cases manifested effects are hemolytic anemia and its sequelae.

The INDEX project Final report

137

In relation to carcinogenicity, naphthalene is not genototoxic in vivo and thus tumor development, observed in rodents,

is considered to arise via a non-genotoxic mechanism. Also, the underlying mechanism for the development of nasal

tumours in the rat is considered to be the chronic inflammatory damage seen at this site. It follows that prevention of

local tissue damage would prevent subsequent development of tumours.

Relevance of EU-population exposure to naphthalene

Table 6.1: Percentage of population exposed beyond the established exposure limit (EL) and margins of safety

Population based studies

LOAEL

100

LOAEL

1000

= EL

Margin of

safety

(MOS)

Description (Study, Year) N 10 mg/m

1 mg/m

0.1 mg/m

0.01 mg/m

(90

)

Athens (Expolis, 96-98) 42 < < 20% 80% no MOS

Basel (Expolis, 96-98) 47 < < < < 16 (9)

Helsinki (Expolis, 96-98) 188 < < < < 22 (9)

Milan (Expolis, 96-98) 41 < < < 5 % 5 (1)

Oxford (Expolis, 98-00) 40 < < < < 12 (4)

Prague (Expolis, 96-98) 46 < < < < 7 (3)

Avg. MOS:

excluding Athens

10 (4)

Animal LOAEL ;

not considering a NOAEL (10);

interspecies variability (10) ;

intraspecies variability (10);

< out of the evaluation range (i.e. <5% of the environments investigated)

The INDEX project Final report

138

Acetaldehyde

Synonyms: ethanal, acetic aldehyde, ethyl aldehyde, methyl formaldehyde

CAS Registry Number: 75-07-0

Molecular Formula: C

1. Compound identification

Acetaldehyde is a colourless, volatile liquid and at dilute concentrations has a pungent odour. Acetaldehyde is a highly

flammable and reactive compound that is miscible in water and most common organic solvents like alcohol, acetone,

gasoline, toluene, xylene, benzene, ether, paraldehyde. It is used in the manufacture of acetic acid, perfumes, and

flavours. As a liquid, it is lighter than water but the vapours are heavier than air. It is volatile at ambient temperature

and pressure. It is highly reactive and a strong reducing agent. Acetaldehyde can react violently with acid anhydrides,

alcohols, ketones, phenols, ammonia, hydrogen cyanide, hydrogen sulfide, halogens, phosphorus, isocyanates, strong

alkalis, and amines. It is also an intermediate in the metabolism of alcohol (SIS 2003).

2. Physical and Chemical properties

Molecular weight (g/mol) 44.1

Melting point (°C) -123.5

Boiling point (°C) 20.2

Density (g/l at 20 °C, 1 atm)

Relative density (air =1) 1.52

Explosion limits of mixtures (air vol-% acetaldehyde) 4.5-60.5

Solubility: miscible in water and most common solvents

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 1.829 µg/m

1 µg/m

= 0.547 ppb

Sources: Hagemeyer (1978), IPCS/CEC (1990), Verschueren 2001.

3. Indoor Air Exposure assessment

Contribution of inhalation exposure to total exposure

Acetaldehyde is a metabolic product of sugars and ethanol. On the basis of the assumptions that a standard drink

contains 10 g of ethanol and that about 90% of imbibed alcohol is metabolized to acetaldehyde, alcoholic beverages are

generally by far the most significant source of exposure to acetaldehyde for the general population (WHO, 1995).

On the basis of the average dietary intake of food groups in different regions of the world and the contents of

acetaldehyde in foodstuffs and non-alcoholic beverages in the Netherlands (Maarse & Visscher, 1992), food

(particularly fruit juices and vinegar) may be one of the principal sources of exposure to acetaldehyde in the general

population. More representative data on mean concentrations in foodstuffs have not been identified, but, on the basis of

the ranges of concentrations determined in the Dutch survey, intake in food is estimated to range from just less than 10

to several hundred µg/kg body weight per day.

On the basis of a daily inhalation volume for adults of 22 m

, a mean body weight for males and females of 64 kg, and

the assumption that mean concentrations are approximately 5 µg/m

(range: 2 to 8.6 µg/m

; Guicheret & Schulting,

The INDEX project Final report

139

1985; Watanabe, 1987; Grosjean, 1991), the mean intake of acetaldehyde from ambient air for the general population is

estimated to be 1.7 µg/kg body weight per day (WHO, 1995).

Emission sources

Typically, concentrations are higher indoors than outdoors due to combustion sources such as cigarettes, fireplaces, and

woodstoves. Acetaldehyde can also be emitted from cooking hamburgers, coffee roasting and from some building

materials such as rigid polyurethane foams, and some consumer products such as adhesives, coatings, lubricants, inks,

and nail polish remover. It is also used as a fruit and fish preservative, flavoring agent, a denaturant for alcohol, for

hardening gelatin fibers, and as a solvent in the synthetic rubber, paraldehyde, tanning and paper industries, in the

manufacture of perfumes, butanol, aniline dyes, plastics, and silvering mirrors (EPA/Cal 2003). Since acetaldehyde is a

product of human metabolism and it is found in exhaled air, it is probable that increased indoor concentrations may be,

in addition to other sources, related to emissions from humans (Jurvelin 2001).

Residences with smokers have two to eight times higher acetaldehyde concentrations than the outdoor mean

concentration (WHO 1995). Based on limited data collected in the US suggest an estimate of an average acetaldehyde

concentration inside residences of 5.4 to 27.0 µg/m

(3.0 to 15 ppb). The acetaldehyde concentrations in offices and

public buildings were similar in magnitude to those inside residences. Average and maximum in-vehicle acetaldehyde

concentrations measured in southern California were similar in magnitude to those inside residences (EPA/Cal 2003).

Acetaldehyde is present in vehicle exhaust and in wastes from various industries. Also open burning and incineration of

gas, fuel oil, and coal produce acetaldehyde. Similarly, degradation of hydrocarbons, sewage, and solid biological

wastes produce acetaldehyde (WHO 1995, EPA/Cal 2003).

Indoor air and exposure concentrations

There is a lack of European population based data for acetaldehyde. In Helsinki average indoor air concentrations of

acetaldehyde are clearly higher than ambient concentrations, 18.2 µg/m

and 2.7 µg/m

respectively. Similarly, indoor

concentrations were higher than personal exposures suggesting the presence of considerably strong indoor sources.

Cumulative distributions of the indoor and 48-hour personal exposure concentrations in Helsinki are presented in Figure

3.1 and Figure 3.2.

In Paris (Figure 3.1) in recently refurbished ﬂats (a non representative study) indoor concentrations measured in bed

rooms were at the same level than in Helsinki (n= 61, GM = 10.2 µg/m

, GSD = 1.8 (Clarisse et al 2003).

The concentration of acetaldehyde in 14 homes and a small office building in northern Italy ranged from 1 to 48 µg/m

with a mean value of 17 µg/m

(DeBortoli M et al, 1986). The ratio of the minimum, maximum, mean, and median

concentration in indoor versus outdoor air was 0.5, 24, 6.0, and 3.6, respectively. In Sweden Gustafsson et al (2003)

reported mean values ranging from 12 µg/m

to 14 µg/m

in non-smoking homes. In 1986 - 1987, the mean yearly

exposure to acetaldehyde from air pollution in Sweden was 1.0 µg/m

(Bostrom CE et al 1994).

Lee and Wang (2004) studied emissions of burning ten types of incense in chamber (18 m

) tests. Acetaldehyde

concentrations of the nine incense types ranged from 50 to 150 µg/m

, one type exceeded a level of 450 µg/m

. From

these results we can conclude that incense burning may increase dramatically indoor air concentrations of acetaldehyde

compared to the ‘normal’ residential levels. Few short time peak concentrations are presented in Table 3.1.

In Denmark (Granby and Kristensen 1997) and Finland (Viskari et al 2000), ambient acetaldehyde showed considerably

low average (3-8 months periods) ambient concentrations being 1.8 µg/m

and 1.1-3.2 µg/m

, respectively. In Italy and

Greece, much higher average concentrations were measured, being 8.2 µg/m

and 15.1 µg/m

, respectively (Possanzini

et al 1996, Bakeas et al 2003).

In the USA ambient concentrations ranged from 2.0 to 5.7 µg/m

(Grosejan et al 1993, CARB 1999). The concentration

of acetaldehyde measured in EPA headquarters building in Washington, DC ranged from 3.8 to 11.1 µg/m

with a

median value of 5.2 µg/m

(EPA 1990). High concentrations were found in Brazil (Grosejan et al 2002), Japan,

(Satsumabayashi et al 1995), and in Mexico (Baez et al 1995), being 10.4 µg/m

, 2.3-12 µg/m

, and 28.6 µg/m

respectively. In Chinese studies ambient levels ranged from 1.7 µg/m

to 7.6 µg/m

. The mean indoor concentration of

acetaldehyde in hotels was tenfold compared to outdoor levels, identified as a consequence of smoking indoors (Ho et al

2002, Feng et al 2004).

The INDEX project Final report

140

100

0 1020304050

Indoor concentration (µg/m

)

Cumulative frequency (%)

Hel

Par

Fre

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of acetaldehyde in Helsinki (Hel, n= 15,

Jurvelin et al 2001), in the French Survey (Fre, n= 201, Kirchner 2004) and in a non-representative study in Paris (Par)

(n= 61, Clarisse et al 2003).

100

010203040

Exposure (µg/m

)

Cumulative frequency (%)

Hel

Figure 3.2. Cumulative frequency distribution of 48-hour personal exposure concentrations of acetaldehyde in Helsinki

(Hel) (n= 15, Jurvelin et al 2001).

The INDEX project Final report

141

Table 3.1. Short time acetaldehyde concentrations related to specific microenvironments or emission sources.

Averaging time Reference

MGMmax

Insence burning during burning 50-450 Lee and Wang 2004

Tavern tobacco smoke 203 EPA 2003

Houses in US manufactured 30-min 5 - 34 18 Hodgson et al 2000

site-built 30-min 21 - 103 36

Exposure to ETS non-smoker living with a smoke

1-da

7.6 Nazaroff and Singer (2002)

AM = arithmetic mean, GM = geometric mean

Environment or

emission source

Concentration (µg/m

)

4. Toxicokinetics

Absorption

Consistent with effects being observed primarily at the initial site of exposure following inhalation (i.e., in the

respiratory tract), available data indicate that the greatest proportion of inhaled acetaldehyde is retained at the site of

contact, becoming rapidly and irreversibly bound to free protein and non-protein sulphydryl groups (notably, cysteine

and glutathione). The results of pharmacokinetic studies conducted in humans (Dalhamn et al., 1968; Egle, 1970) and

rodents (David and Heck, 1983; Morris, 1997) indicate that the absorption of acetaldehyde into the systemic circulation

is likely not extensive following inhalation.

The percentage of acetaldehyde retained by 8 volunteers inhaling acetaldehyde vapour (100-800 mg/m

) from a

recording respirometer ranged from 45 to 70%, at different respiratory rates. Total respiratory tract retention was the

same whether the vapour was inhaled through the nose or the mouth. A direct relationship was found between the

contact time and uptake, independent of rate. Thus, the critical factor in determining acetaldehyde uptake is the duration

of the ventilatory cycle (Egle, 1970).

Distribution

Following inhalation by rats, acetaldehyde is distributed to the blood, liver, kidney, spleen, heart, and other muscle

tissues (Hobara et al., 1985; Watanabe et al., 1986). Low levels were detected in embryos after maternal intraperitoneal

injection of acetaldehyde (mouse) and following maternal exposure to ethanol (mouse and rat) (Blakley and Scott,

1984).

Distribution of acetaldehyde to brain interstitial fluid, but not to brain cells, has been demonstrated following

intraperitoneal injection of ethanol in rats. A high affinity, low Michaelis constant ALDH (aldehyde dehydrogenase)

may be important in maintaining low levels of acetaldehyde in the brain during the metabolism of ethanol (WHO,

1995)

. Acetaldehyde is taken up by red blood cells and, following ethanol consumption in humans and in baboons, in

vivo, intracellular levels can be 10 times higher than plasma levels (Baraona et al., 1987)

Small amounts of acetaldehyde are produced endogenously during the normal intermediary catabolism of deoxyribose

phosphate and various amino acids (Nicholls et al., 1992; Jones, 1995). Consumption of alcoholic beverages is also an

important source of acetaldehyde in the body, formed through the metabolism of ethanol by alcohol dehydrogenase.

Studies using the perfused human placental cotyledon indicated that the human placenta has the potential to produce

acetaldehyde, which can enter the fetal circulation. Furthermore, partial transfer of acetaldehyde from maternal to fetal

blood may occur (Karl et al., 1988).

Metabolism and elimination

Based on the high degree of retention of acetaldehyde in the respiratory tract following inhalation in humans, it is likely

that the predominant pathway for the metabolism of acetaldehyde involves conjugation to thiols (i.e., cysteine and

glutathione) at the site of exposure, subsequent formation of hemimercaptal or thiazolidine intermediates, and

elimination of thioethers and disulphides in the urine (Sprince et al., 1974; Cederbaum and Rubin, 1976; Hemminiki,

1982; Brien and Loomis, 1983; Nicholls et al., 1992).

The INDEX project Final report

142

Inhaled acetaldehyde is also rapidly oxidized to acetate by aldehyde dehydrogenase (ALDH) in human nasal and lung

epithelia (Bogdanffy et al., 1986; Yin et al., 1992; Morris et al., 1996). Acetate enters the citric acid cycle as acetyl-

CoA. ALDH activity has been localized in the respiratory tract epithelium (excluding olfactory epithelium) in rats

(Bogdanffy et al., 1986), in the renal cortex and tubules in the dog, rat, guinea-pig, and baboon (Michoudet and Baverel,

1987a; 1987b), and in the testes in the mouse (Anderson et al.,1985).

There are several isoenzymes of ALDH with different kinetic and binding parameters that influence acetaldehyde

oxidation rates (Marjanen, 1973; Parrilla et al 1974, Teschke et al., 1977). Human liver ALDH consists of at least 4

main isoenzymes, which are also present in many other tissues (Koivula, 1975; Goedde et al., 1979). Subjects

homozygous or heterozygous for a point mutation in the mitochondrial ALDH corresponding gene have low activity of

this enzyme, metabolize acetaldehyde slowly, and are intolerant of ethanol (Yoshida et al., 1984; Goedde and Agarwal,

1986; Hsu et al., 1978).

Acetaldehyde is metabolized by mouse and rat embryonic tissue in vitro (Priscott and Ford, 1985). Though

acetaldehyde is metabolized in a dose-dependent manner in human renal tubules (Michoudet and Baverel, 1987b), the

liver is the most important metabolic site. The metabolism of acetaldehyde can be inhibited by crotonaldehyde,

dimethylmaleate, phorone, disulfiram, and calcium carbamide (WHO, 1995; IARC 1985).

5. Health effects

Effects of short-term exposure

Humans exposed acutely to moderate concentrations of acetaldehyde experience irritation of the eyes and respiratory

tract and altered respiratory function. Animals exposed to moderate to high concentrations exhibit skin and eye irritation

and notable cellular alterations in the respiratory epithelium and hyper- keratosis of the forestomach.

A group of 12 human subjects were exposed to acetalclehycle vapour for 15 min while being shown a movie "to divert

their attention". Most of the subjects developed eye irritation at 90 mg/m

(50 ppm), but it took more than 360 mg/m

(200 ppm) to cause nose or throat irritation in the majority of the subjects (Silverman et al., 1946). Several subjects

strenuously objected to the vapor at as low as 45 mg/m

(25 ppm). Even those who reported no eye irritation at 90

mg/m

showed erythematous eyelids and bloodshot eyes when exposed to 360 mg/m

of acetalctehyde (Silverman et al.,

1946).

A group of 14 "healthy male" volunteers, aged 18-45 years, were exposed in a 100 m

chamber to a measured

concentration of acetaldehyde vapour of 240 mg/m

(134 ppm) for 30 min. This concentration was said to be mildly

irritating to the upper respiratory tract. No other clinical signs were reported (Sim and Pattle, 1957).

Intravenous infusion of human subjects with 5% acetaldehyde at a rate of 20.6-82.4 mg/min for up to 36 minutes (the

lowest dose converts to 10.6 mg/kg over 36 minutes) resulted in an increased heart rate, increased ventilation rates and

respiratory dead space, and a decreased alveolar carbon dioxide level (Asmussen et al., 1948; IARC 1985).

Summary of short-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Study

Eye irritation Human volunteers,

15min; poor quality of

data

Silverman et al., 1946 WHO 1995;

TC: 2

Study

11.5

1h-ADJ

Eye irritation Human volunteers,

15min; poor quality of

data

Silverman et al., 1946 OEHHA 1999;

HPC: 0.115

Final statement (UNIMI) : Human LOAEL: 5 mg/m³, Animal NOAEL: 50 mg/m³, no evidence in considering acute and chronic effect levels separately

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

WHO (1995)

International Programme on Chemical Safety - Environmental Health Criteria 167

Approach to risk assessment: The following guidance is provided as a potential basis for derivation of limits of

exposure by relevant authorities. Though the principal sources of exposure to acetaldehyde in the general population are

The INDEX project Final report

143

through the metabolism of alcohol, in cigarette smoke, and food, air is believed to be the main route of exposure in the

occupational environment.

On the basis of data on irritancy in humans, a tolerable concentration can be derived as follows:

Tolerable concentration = 45 mg/m

: 20 = 2 mg/m

(2000 µg/m

) where no effects were observed in a limited study on

human volunteers at 45 mg/m

(Silverman et al., 1946) and 20 is the uncertainty factor (×10 for intraspecies variation

and × 2 for the poor quality of the data).

OEHHA (1999b)

Office of Environmental Health Hazard Assessment

Derivation of Health Protective Concentration (draft)

Since adopted health assessment values suitable for assessing potential health impacts from short-term inhalation

exposures are not available for acetaldehyde, OEHHA calculated a draft health protective concentration (HPC). In

calculating the HPC, OEHHA followed the risk assessment methodology used for developing the acute reference

exposure levels (RELs) under the Air Toxics Hot Spots Program risk assessment guidelines process (OEHHA, 1999a).

As mandated by state legislation, these guidelines underwent scientific and public peer review, prior to approval by the

Scientific Review Panel (Senate Bill 1731, Statutes of 1992, Ch. 1162 of the California Health and Safety Code) and

adoption by OEHHA.

Silverman et al. (1946) found that a 15-minute exposure to acetaldehyde induced eye irritation in male and female

human volunteers at a concentration of 45 mg/m3 (25 ppm). This concentration was identified as a lowest-observed-

adverse-effect level (LOAEL); a no-observed-adverse-effect level (NOAEL) was not observed in this study.

Application of “Haber’s Law” to extrapolate from a 15 minute exposure to 1 hour results in an adjusted LOAEL of 11.5

mg/m3 (6.5 ppm).

An acute 1-hour draft HPC can be calculated for acetaldehyde using the formula:

Draft HPC = LOAEL / UF = 11.5 mg/m3 / 100 = 115 µg/m3 (65 ppb).

The uncertainty factor (UF) for this calculation is 100, which incorporates uncertainty contributions for extrapolation

from a LOAEL to a NOAEL (10) and for potentially sensitive human subpopulations (10).

Effects of long-term exposure

The critical health effects arising from exposure to airborne acetaldehyde are eye and upper respiratory tract irritation

with the possibility of chronic tissue damage and inflammation in the respiratory tract following long-term exposure.

Such long-term inflammatory changes have been associated with tumour in the nasal passages of experimental animals.

The underlying mechanism for tumour formation is considered to be sustained irritation/inflammation accompanied by

cellular proliferation. Such a mechanism is consistent with a threshold for tumour formation and thus control to prevent

local tissue damage and inflammation would be predicted to control for any risk of tumour formation.

No information is available in humans regarding chronic tissue inflammation, the only data available relate to sensory

irritation. The only objective evidence of eye irritation in humans was with a single exposure to 360 mg/m

. There are

reports of subjective irritation symptoms at lower concentrations (down to 45 mg/m

), although there is uncertainty

regarding the reliability of these observations.

In short-term inhalation studies (> 4 weeks) conducted in rodents, degenerative changes (including inflammation,

hyperplasia, thinning and disarrangement of epithelial cells, and loss of microvilli and sensory cells) and associated

functional effects were observed in the nasal olfactory epithelium in rats exposed (by inhalation) to ≥437 mg/m

acetaldehyde, while degenerative changes in the nasal respiratory epithelium, larynx, trachea and lungs were observed

at higher concentrations (i.e. ≥1800 mg/m

) (Appelman et al., 1982, 1986; Saldiva et al., 1985; Cassee et al., 1996).

In the only subchronic inhalation study identified, in which hamsters were exposed to acetaldehyde for 13 weeks

(Kruysse et al., 1975), non-neoplastic lesions in the tracheal epithelium (including stratification, keratinization,

inflammation, metaplasia and granulation) were observed at ≥2412 mg/m

acetaldehyde (considered to be the LOAEL),

while histopathological lesions in the nasal cavities, larynx, bronchi and lungs were observed only at 8208 mg/m

acetaldehyde; the NOEL in this study was considered to be 702 mg/m

acetaldehyde (Kruysse et al., 1975).

In chronic inhalation studies in which rats were exposed to acetaldehyde for up to 28 months, focal basal cell

hyperplasia of the nasal olfactory epithelium was observed at ≥1350 mg/m

acetaldehyde (considered to be the

The INDEX project Final report

144

LOAEL), while non-neoplastic lesions in the nasal respiratory epithelium (squamous metaplasia, papillomatous

hyperplasia, and focal or pseudoepitheliomatous hyperplasia) and larynx (squamous metaplasia and hyperplasia) were

observed at concentrations ≥2700 mg/m

acetaldehyde (Woutersen et al., 1984, 1986; Feron et al., 1985; Woutersen and

Feron, 1987). Similarly, nonneoplastic lesions in the nasal epithelia, larynx and trachea have been observed in hamsters

exposed by inhalation to ≥2700 mg/m

acetaldehyde for 52 weeks (Feron, 1979; Feron et al., 1982).

The studies that provide best characterization of concentration–response for the critical effects in the most sensitive

species (i.e., rats) are the short-term studies of Appelman et al. (1982; 1986). While these studies are short-term in

duration, together they establish a concentration-response for lesions after only 4 weeks of exposure. These same types

of lesions appear at longer exposure times and higher exposure levels in chronic studies (Wouterson et al., 1986;

Wouterson and Feron, 1987; Kruysse et al., 1975). Under other circumstances, studies of short duration may not be

considered appropriate, but for acetaldehyde the observed effects are consistent with pathology seen in long-term

studies.

Appleman et al. (1986) conducted two inhalation studies on male Wistar rats (10/group) exposing them 6 hours/day, 5

days/week for 4 weeks to 0, 273 and 910 mg/m

(0, 150, and 500 ppm respectively). Duration-adjusted concentrations

are 0, 48.75, and 162.5 mg/m

, respectively. One group was exposed without interruption, a second group was

interrupted for 1.5 hours between the first and second 3-hour period, and a third group was interrupted as described with

a superimposed peak exposure profile of 4 peaks at 6-fold the basic concentration per 3-hour period. The purpose was

to test intermittent and peak exposure effects. Urine samples were collected from all rats and lung lavage performed on

4-5 per group at the end of the experiment. Cell density, viability, number of phagocytosing cells, and phagocytic index

were determined on the lavage fluid. Microscopic examination was performed on the nasal cavity, larynx, trachea with

bifurcation and pulmonary lobes of all rats of all groups.

Continuous and interrupted exposure to 910 mg/m

did not induce any visible effect on general condition or behavior,

but peak exposures at this level caused irritation. No behavioral differences were noted in the other groups. Mean body

weights of the group exposed to 910 mg/m

with interruption and with peak exposures were statistically significantly

lower than those of the controls. Body weights were similar to controls in the other exposure groups. Mean cell density

and cell viability were significantly decreased in the group exposed to 910 mg/m

with or without peak exposures. The

mean percentage of phagocytosing cells and the phagocytic index were significantly lower than controls in all groups

exposed to 910 mg/m

, especially the group exposed to superimposed peaks. Histopathological changes attributable to

exposure were found only in the nasal cavity. Degeneration of the olfactory epithelium was observed in rats exposed to

910 mg/m

. Interruption of the exposure or interruption combined with peak exposure did not visibly influence this

adverse effect. No compound-related effects were observed in rats interruptedly or uninterruptedly exposed to 273

mg/m

during the 4-week exposure period; therefore, the NOAEL is 273 mg/m

Appelman et al. (1982) exposed Wistar rats (10/sex/group) for 6 hours/day, 5 days/week for 4 weeks to 0, 728, 1820,

4004 and 9100 mg/m

acetaldehyde (0, 400, 1000, 2200, or 5000 ppm, respectively). Duration-adjusted concentrations

are 0, 130, 325, 715 and 1625 mg/m

, respectively. The general condition and behavior of the rats were checked daily.

Blood picture (Hb, Hct, RBC, total and differential WBC, and plasma protein) and chemistry were examined at the end

of the treatment period. Activities of plasma glutamic-oxalacetic transaminase, glutamic-pyruvic transaminase, and

alkaline phosphatase were also determined. Urine was analyzed for density, volume, pH, protein, glucose, occult blood,

ketones, and appearance. The kidneys, lungs, liver, and spleen were weighed. Microscopic examination was performed

on the lungs, trachea, larynx, and nasal cavity (3 transverse sections) of all animals and on the kidneys, liver, and spleen

of all control and high- concentration groups.

During the first 30 minutes of each exposure at the 9100 mg/m

level, rats exhibited severe dyspnea that gradually

became less severe during the subsequent exposure period. Two animals died at this level (1 female, 1 male) and one

male died at the 4004 mg/m

level, but the cause of death could not be determined due to autolysis or cannibalism.

Growth was retarded in males at the three highest exposure concentrations and in females at the 9100 mg/m

level. The

percentage of lymphocytes in the blood was lower and the percentage of neutrophilic leukocytes higher in males and

females of the 9100 mg/m

group than in controls. There were a few statistically significant differences in several blood

chemistry parameters between the exposure groups and the control group but none of them were concentration-related.

Statistically significant changes in organ-to-body weight ratios included decreased liver weights in both sexes and

increased lung weights in males at the 9100 mg/m

level. Males in the 9100 mg/m

level produced less urine, but it was

of higher density. Compound-related histopathological changes were observed only in the respiratory system. The nasal

cavity was most severely affected and exhibited a concentration-response relationship. At the 728 mg/m

level,

compound-related changes included: slight to severe degeneration of the nasal olfactory epithelium, without hyper- and

metaplasia, and disarrangement of epithelial cells. At the 1820- and 4004 mg/m

levels, more severe degenerative

changes occurred, with hyperplastic and metaplastic changes in the olfactory and respiratory epithelium of the nasal

cavity. Degeneration with hyperplasia/metaplasia also occurred in the laryngeal and tracheal epithelium at these levels.

The INDEX project Final report

145

At 9100 mg/m

changes included severe degenerative hyperplastic and metaplastic changes of the nasal, laryngeal, and

tracheal epithelium. Based on the degenerative changes observed in the olfactory epithelium, the 728 mg/m

level is

designated as a LOAEL.

Available data are inadequate to assess the potential reproductive, developmental, neurological or immunological

effects of direct exposure to acetaldehyde. Based on the limited number of investigations conducted to date, however,

reproductive, developmental, neurological and immunological effects have not been observed at concentrations below

those that induce damage in the upper respiratory tract (Ortiz et al., 1974; Kruysse et al., 1975; Shiohara et al., 1985;

Aranyi et al., 1986; Roumec et al., 1988).

Summary of long-term exposure effect levels

Health Canada and U.S. EPA have evaluated the noncancer inhalation toxicity data for acetaldehyde. Health Canada

derived a tolerable concentration of 0.39 mg/m

. EPA derived a reference concentration (RfC) of 0.009 mg/m

, which is

40-fold lower than the Health Canada value. Both agencies used the same studies, but differed in their estimation. While

EPA used the NOAEL of 273 mg/m

from Appleman et al. (1982), Health Canada used the lower 95% confidence limit

on a benchmark dose (BMD) for the concentration associated with a 5% increase in the incidence of nasal olfactory

epithelial lesions for male rats [from data in the studies by Appleman et al. (1982, 1986)]. Both agencies adjusted for

the intermittent exposure in the laboratory, but EPA also adjusted to a Human Equivalent Concentration (HEC) by

calculating for a gas:respiratory effect in the extrathoracic region. The agencies also differed in choice of uncertainty

factor, with EPA utilizing three factors of ten (intraspecies, use of a subchronic study, and interspecies and database

combined for a third ten). Health Canada did not use a factor for limitations in the database due to the fact that a TC that

is based on critical effects at the site of entry is likely to be protective for systemic effects. A factor for use of a shorter-

term study was not deemed appropriate because there was no indication that severity of the critical effects increases

with duration of exposure. Click on the green circle(s) for more information.

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

273

EXP

ADJ

8.7

HEC

728

EXP

130

ADJ

16.9

HEC

Degeneration of

olfactory epithelium

Rat, 4w Appleman et al.,

1986;1982

EPA-IRIS 1991;

RfC: 0.003

218

BC05

ADJ

Degeneration of

olfactory epithelium

Rat, 4w Appleman et al.,

1986;1982

Health Canada

2000; TC: 0.39

- - Respiratory system not specified OEHHA 1995;

REL: 0.009

Although the studies cited establish a concentration-response for lesions after only 4 weeks of exposure, same types of lesions appear at longer exposure

times and higher exposure levels, and are consistent with pathology seen in chronic studies.

Final statement (UNIMI) : Human LOAEL: 5 mg/m³, Animal NOAEL: 50 mg/m³, no evidence in considering acute and chronic effect levels separately

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

U.S.EPA - IRIS (1991)

Integrated Risk Information System

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC)

Critical effect Exposures* UF MF RfC

Degenration of

olfactory epithelium

Short-term Rat

Inhalation Studies

Appleman et al.,

1986;1982

NOAEL: 273 mg/m

(150 ppm)

NOAEL:(ADJ): 48.75 mg/m

NOAEL(HEC): 8.7 mg/m

LOAEL: 728 mg/m

(400 ppm)

LOAEL(ADJ): 130 mg/m

LOAEL(HEC): 16.9 mg/m

1000 1

0.003 mg/m

*Conversion Factors -- MW = 44.5. Appleman et al., 1986: Assuming 25C and 760 mmHg, NOAEL(mg/cu.m) = 150 ppm x 44.5/24.45 = 273.

NOAEL(ADJ) = 273 mg/cu.m x 6 hours/day x 5 days/7 days = 48.75 mg/cu.m. The NOAEL(HEC) was calculated for a gas:respiratory effect in the

ExtraThoracic region. MVa = 0.23 cu.m/day, MVh = 20 cu.m/day, Sa(ET) = 11.6 sq. cm, Sh(ET) = 177 sq. cm. RGDR(ET) = (MVa/Sa) / (MVh/Sh) =

0.18. NOAEL(HEC) = NOAEL(ADJ) x RGDR = 8.7 mg/cu.m.

Appleman et al., 1982: Assuming 25C and 760 mmHg, LOAEL(mg/cu.m) = 400 ppm x 44.5/24.45 = 130. LOAEL(ADJ) = 728 mg/cu. m x 6 hours/day

x 5 days/7days = 130 mg/cu.m. The LOAEL(HEC) was calculated for a gas:respiratory effect in the ExtraThoracic region. MVa = 0.17 cu.m/day,

The INDEX project Final report

146

MVh = 20 cu.m/day, Sa(ET) = 11.6 sq. cm., Sh(ET) = 177 sq.cm. RGDR(ET) = (MVa/Sa) / (MVh/Sh) = 0.13. LOAEL(HEC) = LOAEL(ADJ) x RGDR

= 16.9 mg/cu.m.

Two short-term studies conducted by the same research group are the principal studies used. While these studies are

short-term in duration, together they establish a concentration-response for lesions after only 4 weeks of exposure.

These same types of lesions appear at longer exposure times and higher exposure levels in chronic studies (Wouterson

et al., 1986; Wouterson and Feron, 1987; Kruysse et al., 1975). Under other circumstances, studies of short duration

may not be considered appropriate, but for this chemical the observed effects are consistent with pathology seen in

long-term studies. The 150-ppm exposure level was therefore established as the NOAEL from the Appleman et al.

(1986) study and the LOAEL from the Appleman et al. (1982) study.

Uncertainty factors (UF) for the determination of the Inhalation RfC: An uncertainty factor of 10 was applied to account

for sensitive human populations. A factor of 10 was applied for both uncertainty in the interspecies extrapolation using

dosimetric adjustments and to account for the incompleteness of the data base. A factor of 10 was applied to account for

subchronic to chronic extrapolation.

Confidence in the Inhalation RfC: Confidence in the principal studies (Appleman et al., 1982; 1986) is medium since

appropriate histopathology was performed on an adequate number of animals and a NOAEL and LOAEL were

identified, but the duration was short and only one species was tested. Confidence in the data base is low due to the lack

of chronic data establishing NOAELs and due to the lack of reproductive and developmental toxicity data. Low

confidence in the RfC results.

Health Canada (2000)

Canadian Environmental Protection Act (CEPA).

Derivation of a Tolerable Concentration (TC):

A tolerable concentration (TC) has been developed on the basis of Benchmark Concentration (BMC) for degeneration

in the nasal olfactory epithelium of Wistar rats exposed to acetaldehyde (by inhalation) for four weeks, using combined

data from Appelman et al. (1982, 1986).

For many types of effects, studies of short duration are not preferred as the basis for development of a TC. While the

data were derived from short-term studies, the incidence of degenerative changes in the olfactory epithelium was not

dissimilar to that observed in the same strain of rats in the long-term carcinogenesis bioassay at similar concentrations,

conducted by Woutersen et al. (1986). Although group sizes were larger in the long-term bioassay, concentration-

response for these lesions was not well characterized because of the small number of dose groups exposed to higher

concentrations compared with the short-term study and early mortality among animals at the highest concentration.

Data in the Woutersen et al. (1986) study are insufficient to serve as a basis for development of a meaningful BMC for

acetaldehyde, even simply for purposes of comparison.

Using the THRESH program (Howe, 1995), the BMC

(the concentration associated with a 5% increase in the

incidence of nasal olfactory epithelial lesions) for male Wistar rats is 357 mg/m

; the lower 95% confidence limit for

this value (BMCL

) is 218 mg/m

. For female Wistar rats, the BMC

and BMCL

are 445 mg/m

and 17 mg/m

respectively.

The BMCL

of 218 mg/m

was adjusted from an intermittent to a continuous exposure by multiplying the value by 6/24

and 5/7. There are no data that provide direct evidence and few related data to serve as a basis for whether or not such

an adjustment is appropriate for acetaldehyde. In short-term studies in the same strain of rats, effects seemed slightly

more severe following exposure for eight hours per day (Saldiva et al., 1985) versus six hours per day for four weeks

(Appelman et al., 1986). Interruption of daily exposure by 1.5-hour exposure-free periods or by the superimposition of

eight five-minute peak exposure periods did not appreciably influence the cytotoxic potency of acetaldehyde in short-

term studies in rats compared with uninterrupted exposure to a fixed concentration (Appelman et al., 1986).

An uncertainty factor of 100 was used (10 for interspecies variation, 10 for intraspecies variation). Available data are

inadequate to further address toxicokinetic and toxicodynamic aspects of components of uncertainty with data-derived

values. The value for interspecies variation is considered to be conservative since, due to greater penetration of inhaled

gases into the lower airways of rodents versus humans, the compound is distributed over a larger surface area for the

latter; available data are inadequate, however, to quantitatively account for this variation. No additional quantitative

element has been included to address limitations of the database such as lack of adequate developmental or reproductive

studies by a relevant route of exposure, due to the fact that a TC that is based on critical effects at the site of entry is

likely to be protective for systemic effects. Also, in view of the fact that there is no indication that severity of the critical

The INDEX project Final report

147

effects increases with duration of exposure, an additional quantitative element to address the use of a shorter-term study

as the basis for the TC is considered inappropriate.

The resulting Tolerable Concentration (TC) is 0.39 mg/m

. This TC is similar to that derived from the NOEL for

irritation in the study of Appelman et al. (1986) (TC of 0.49 mg/m

). On the basis of limited available data in human

studies, the TCs derived above (0.39 and 0.49 mg/m

) are two orders of magnitude lower than the threshold for sensory

irritation (i.e., 45 mg/m

[25 ppm]) (Silverman et al., 1946).

The degree of confidence in the database on toxicity that serves as the basis for the development of the tolerable

concentration (TC) for inhalation is moderate, although there is a relatively high degree of confidence that critical

effects occur at the initial site of exposure. Despite differences in the anatomy and physiology of the respiratory tract in

rats and humans, respiratory tract defense mechanisms are similar. Thus, it is reasonable to assume that the response of

the human respiratory tract mucosa to acetaldehyde will be qualitatively similar to that of experimental species,

although the likely site of development of lesions may vary due to oro-nasal breathing patterns in humans, which result

in greater potential to deliver acetaldehyde to the lower respiratory tract.

WHO (1995)

Environmental Health Criteria 167

Data suggest that acetaldehyde causes genetic damage to somatic cells in vivo. The irritancy of acetaldehyde may also

play an important role in the development of tumours in the nose and larynx of rats and hamsters, respectively, exposed

by inhalation, though all concentrations of acetaldehyde administered in carcinogenesis bioassays induced both irritancy

and nasal tumours. Therefore, two approaches were adopted for the provision of guidance with respect to the potential

carcinogenicity of acetaldehyde.

1.) In the first, a tolerable concentration (TC) was derived on the basis of division of an effect level for irritancy in the

respiratory tract of rodents by an uncertainty factor, based on the principles outlined in WHO and the assumption that

there is a threshold for acetaldehyde-induced cancer of the respiratory tract in rodents exposed via inhalation. There is

some support for this approach on the basis of relevant data on the analogues of acetaldehyde, i.e., formaldehyde and

glutaraldehyde, which have similar spectra of in vitro mutagenic effects, but are clearly not mutagenic in vivo (IARC,

1985; WHO, 1989). Thus,

Tolerable concentration = 275 mg/m

: 1000 = 0.3 mg/m

(300 µg/m3)

where: 275 mg/m

was the NOEL for irritation in rats in a 4-week study (Appelman et al., 1986) and 1000 is the

uncertainty factor (×10 for interspecies variation, ×10 for intraspecies variation and ×10 for a less than long-term study

and severity of effect, i.e., carcinogenicity associated with irritation).

2.) Since the mechanism of induction of tumours by acetaldehyde has not been well studied, lifetime cancer risk has

also been estimated on the basis of a default model (i.e., linearized multistage). However, it is very likely that, since

estimated risk is based on tumour incidence at concentrations that induce irritancy in the respiratory tract and no-

observed-effect levels for irritancy are well below these concentrations, the true cancer risk is most likely much lower at

concentrations generally present in the environment and may, indeed, be zero.

Concentrations associated with a 10

-5

excess lifetime risk (lower 95% confidence limits) for nasal tumours

(adenocarcinomas, squamous cell carcinomas, and carcinomas in situ) in male and female rats, in the only

carcinogenicity study in which animals were exposed via inhalation to acetaldehyde over the lifetime (Woutersen et al.,

1986), calculated on the basis of the linearized multistage model (Global 82), are 11-65 µg/m

. The high-dose animal

groups were excluded in the derivation of these estimates because of early mortality. A body surface area correction

was not incorporated.

OEHHA (1993)

Office of Environmental Health Hazard Assessment

A chronic non-cancer Reference Exposure Level (REL) of 9.0 µg/m is listed for acetaldehyde in the California Air

Pollution Control Officers Association Air Toxics “Hot Spots” Program, Revised 1992 Risk Assessment Guidelines.

The toxicological endpoint considered for chronic toxicity is the respiratory system.

The INDEX project Final report

148

Carcinogenic potential

Acetaldehyde is reasonably anticipated to be a human carcinogen based on sufficient evidence of carcinogenicity in

experimental animals (IARC 1985, 1987, 1999). When administered by inhalation, acetaldehyde increased the

incidence of squamous cell carcinomas and adenocarcinomas in the nasal mucosa in rats of both sexes and laryngeal

carcinomas in hamsters of both sexes. In another inhalation study using a lower exposure level and in an intratracheal

instillation study, no increased incidence of tumors in hamsters was observed. When administered by inhalation,

acetaldehyde enhanced the incidence of respiratory tract tumors as induced by intratracheal instillation of

benzo[a]pyrene in hamsters of both sexes.

There is inadequate evidence for the carcinogenicity of acetaldehyde in humans (IARC 1985, 1987, 1999). A single

study of workers in an aldehyde plant reported nine cases of cancer, including five cases of bronchial tumors and two

cases of carcinomas of the oral cavity. This study was considered to be inadequate for evaluation because of mixed

exposure, the small number of cases, and the poorly defined population. Three case control studies investigated the risk

of oral, throat, and esophageal cancers following heavy alcohol intake. These studies consistently showed an increased

risk of these cancers in people with genetic polymorphisms; these polymorphisms resulted in higher blood acetaldehyde

concentrations after drinking alcohol (IARC 1999). Overall evaluation (IARC 1999): Acetaldehyde is possibly

carcinogenic to humans (Group 2B; there is inadequate evidence in humans for the carcinogenicity of acetaldehyde.

There is sufficient evidence in experimental animals for the carcinogenicity of acetaldehyde).

Acetaldehyde is considered a probable human carcinogen (Classification: B2; EPA 1998), based on increased incidence

of nasal tumors in male and female rats and laryngeal tumors in male and female hamsters after inhalation exposure.

Health Canada (2000) and EPA (1998) each used the same study (Woutersen et al., 1986; Woutersen and Appelman,

1984) and tumors (nasal adenocarcinomas and squamous cell carcinomas) to calculate cancer potency, but their

approaches differ in that Health Canada does not do low dose extrapolation. Health Canada calculated a tumorigenic

concentration with 5% response (TC

) of 86 mg/m

, with a lower 95% confidence limit (TCL

) of 28 mg/m

. EPA

estimated an inhalation unit risk of 2.2 E-6 per µg/m

Summary of principal study: The carcinogenicity of acetaldehyde was studied in 420 male and 420 female albino SPF

Wistar rats (Woutersen and Appelman, 1984; Woutersen et al., 1985). After an acclimatization period of 3 weeks, these

animals were randomly assigned to four groups of 105 males and 105 females each. The animals were then exposed by

inhalation to atmospheres containing 0, 1351, 2702, or 5403 mg/m

(0, 750, 1500, or 3000 ppm) acetaldehydefor 6

hours/day, 5 days/week, for 27 months. The concentration in the highest dose group was gradually reduced from 5403

to 1801 mg/m

because of severe growth retardation, occasional loss of body weight and early mortality in this group.

Interim sacrifices were carried out at 13, 26, and 52 weeks. One tumor was observed in the 52 week sacrifice group and

none at earlier times. Exposure toacetaldehyde increased the incidence of tumors in an exposure-related manner in both

male and female rats. In addition, there were exposure-related increases in the incidences of multiple respiratory tract

tumors. Adenocarcinomas were increased significantly in both male and female rats at all exposure levels, whereas

squamous cell carcinomas were increased significantly in male rats at middle and high doses and in female rats only at

the high dose. The squamous cell carcinoma incidences showed a clear dose-response relationship. The incidence of

adenocarcinoma was highest in the mid-exposure group (2702 mg/m

) in both male and female rats, but this was

probably due to the high mortality and competing squamous cell carcinomas at the highest exposure level. In the low-

exposure group, the adenocarcinoma incidence was higher in males than in females.

Genotoxicity

Acetaldehyde is genotoxic in vitro, inducing gene mutations, clastogenic effects, and sister-chromatid exchanges

(SCEs) in mammalian cells in the absence of exogenous metabolic activation (He and Lambert, 1990; Badr and

Hussain, 1977; Obe et al., 1978; 1979; 1985; Veghelyi and Osztovics, 1978; Boehlke, 1983).

The results of in vivo studies suggest that acetaldehyde can react directly with DNA and proteins to form stable adducts.

Acetaldehyde produced a concentration-related reduction in the extractability of DNA (suggestive of increased

formation of DNA–protein cross-links) from the respiratory nasal mucosa of Fischer 344 rats exposed (whole body) to

1800 or 5400 mg/m

acetaldehyde for six hours or to 1800 mg/m

acetaldehyde for six hours per day for five days.

Significant reduction in the extractability of DNA in the nasal olfactory epithelia was observed only following exposure

to 1800 mg/m

acetaldehyde for six hours per day for five days (Lam et al., 1986). There is indirect evidence from in

vitro and in vivo studies to suggest that acetaldehyde can induce protein-DNA and DNA-DNA cross-links (Lam et al.,

1986).

Many of the toxicological effects of acetaldehyde may be due to the saturation of protective cellular mechanisms at the

initial site of exposure. As with formaldehyde, the potential for acetaldehyde to react with epithelial DNA (and other

cellular components) in the upper respiratory tract may be dependent upon the levels of intracellular thiols (notably

The INDEX project Final report

149

glutathione and cysteine), which prevent binding of acetaldehyde with critical sulphydryl groups in proteins, peptides

and DNA (Cederbaum and Rubin, 1976; U.S. EPA, 1987; von Wartburg, 1987).

In addition, regional deficiencies in aldehyde dehydrogenase activity in rats correlate with the distribution of nasal

lesions in another strain of rats exposed to acetaldehyde in inhalation studies (Bogdanffy et al., 1986). Observed

decreases in uptake of acetaldehyde at high concentrations >180 mg/m

in a range of species may be a function of

exceedance of the metabolic capacity of nasal aldehyde dehydrogenase (Morris, 1997).

The pattern of observed irritancy of acetaldehyde at the site of contact and the results of studies indicating that it can

react directly with DNA and proteins to form stable adducts is similar to that for other aldehydes (such as

formaldehyde) that have been carcinogenic to the respiratory system in sensitive inhalation bioassays. Although the

exact mechanism is unknown, induction of tumours by these aldehydes is considered to be a function of both

regenerative proliferative response and DNA–protein cross-linking at the site of contact.

Similarly, it has been proposed that the genotoxicity of acetaldehyde is based principally upon its ability to interact with

single-stranded DNA during cell division (Feron et al., 1982, 1984; Woutersen et al., 1986; Roe and Wood, 1992;

DECOS, 1993). Thus, a crucial determinant in the carcinogenicity of acetaldehyde in the nasal passages may be the

cytotoxicity of this substance at high concentrations (Feron et al., 1982, 1984; Woutersen et al., 1986; Roe and Wood,

1992); cytotoxic concentrations of acetaldehyde cause recurrent tissue damage (and the presence of single-stranded

DNA) and possess significant initiating activity. Moreover, the increased cell turnover may strongly enhance the

fixation of relevant DNA damage and subsequently increase the progression of pre-cancer (initiated) cells to cancer.

However, the limited available data indicate that the pattern of DNA–protein cross-linking and proliferative response

induced by acetaldehyde varies from that of other aldehydes, such as formaldehyde. For acetaldehyde, at concentrations

at which tumours are observed 1350 mg/m

, there are increases in DNA–protein cross-links in the respiratory and

olfactory mucosa of rats but no increase in proliferation (Cassee et al., 1996a).

For formaldehyde, at the lower concentrations at which tumours are observed 7.2 mg/m

, there are increases in DNA–

protein cross-links and proliferation in the nasal respiratory (but not olfactory) epithelium (Casanova et al., 1994).

While acetaldehyde is genotoxic in vitro and in vivo, information concerning the potential roles of cytotoxicity, cell

proliferation and DNA–protein cross-links in tumour formation is lacking.

Interactions with other chemicals

Acetaldehyde is a highly reactive molecule that can react with many other large or small molecules by addition,

condensation, or polymerization. These pathways may have little quantitative significance in acetaldehyde metabolism,

but the by-products may have biological significance.

Acetaldehyde can react with various macromolecules in the body, which can lead to marked alterations in the biological

function of these molecules, as evidenced by inhibition of enzyme activity, impaired histone-DNA binding, and

inhibition of polymerization of tubulin. The best characterized nucleophiles able to form adducts with acetaldehyde are

amino groups, notably the alpha-amino terminus of peptides and proteins and the epsilon-amino group on the side-chain

of lysine residues.

The toxicity studies with mixtures of aldehydes showed that histopathological changes and cell proliferation of the nasal

epithelium induced by mixtures of formaldehyde, acetaldehyde and/or acrolein appeared to be more severe and more

extensive, both in the respiratory and the olfactory part of the nose, than those observed after exposure to the individual

aldehydes at comparable exposure levels. However the combined effect of the mixtures was at most the sum of the

individual effects. Neither dose addition nor potentiating interactions occurred upon exposure to combinations of these

aldehydes at exposure levels slightly below or around the minimal-observed-effect level (MOEL) (Flemming, 1995;

Cassee, 1996).

Metronidazole or cotrimoxazole (antimicrobial agents) may enhance accumulation of acetaldehyde in blood induced by

antialcohol drugs like disulfiram or nitrefazole (Heelon and White, 1998; Cina et al., 1996; Suokas et al., 1985).

Odour perception

Acetaldehyde is a volatile liquid with a pungent, suffocating odour that is fruity in dilute concentrations.

The INDEX project Final report

150

Source: Devos et al.(1990):

Odour threshold: 0.025 mg/m

Source: Amoore and Hautala (1983)

The odour threshold for acetaldehyde is reported to be 0.09 mg/m

(0.05 ppm), a geometric average of all available

literature data

Source: AIHA (1989)

A wide range of values has been reported: 0.005 to 1800 mg/m

. An acceptable, critiqued value is 0.12 mg/m

(detection)

Source: New Jersey Department of Health and Senior Services

Odour threshold: 0.12 mg/m

The INDEX project Final report

151

Summary of Acetaldehyde Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Acetaldehyde is a metabolic product of ethanol and sugars. Estimate of diatary

intake: 10 to several hundred µg/kg bw., one std alc.drink: 100 mg/kg bw., ambient air: 2 µg/kg bw.. Consequently,

most toxicological studies in the literature consider systemic effects of acetaldehyde.

Toxicokinetics: The greatest proportion (45-70%) of inhaled acetaldehyde is retained at the site of contact, with

irreversible binding to sulphhydryl groups (e.g.cisteine) or oxidation to acetate (acetyl-CoA). No extensive absorption

into the systemic circulation. Critical factor determining the uptake is the duration of the ventilation cycle, no difference

between nose and mouth breathing.

Health effect levels of short- and long-term exposure (noncancer):

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

Study

Eye irritation Human volunteers,

15min; poor quality of

data

Silverman et al., 1946 WHO 1995;

TC: 2

Study

11.5

1h-ADJ

Eye irritation Human volunteers,

15min; poor quality of

data

Silverman et al., 1946 OEHHA 1999;

HPC: 0.115

Long-term exposure

273

EXP

ADJ

8.7

HEC

728

EXP

130

ADJ

16.9

HEC

Degeneration of

olfactory epithelium

Rat, 4w Appleman et al.,

1986;1982

EPA-IRIS 1991;

RfC: 0.003

218

BC05

ADJ

Degeneration of

olfactory epithelium

Rat, 4w Appleman et al.,

1986;1982

Health Canada

2000; TC: 0.39

- - Respiratory system not specified OEHHA 1995;

REL: 0.009

Remark on Appleman studies: Although the studies cited establish a concentration-response for lesions after only 4 weeks of exposure, same types of

lesions appear at longer exposure times and higher exposure levels, and are consistent with pathology seen in chronic studies.

Carcinogenicity: IARC: 2B; U.S.EPA: B2, Unit risk (EPA-IRIS): 2.2E-06 (µg/m

)

-1

; The underlying mechanism for

tumour formation is considered to be sustained irritation/inflammation accompanied by cellular proliferation, a

mechanism consistent with a threshold for acetaldehyde induced cancer.

Genotoxicity: Genotoxic in vitro and in vivo, though information concerning the potential roles of cytotoxicity, cell

proliferation and DNA-protein cross-links in tumour formation is lacking.

Odour threshold: Pungent, suffocating odour that is fruity at dilute concentrations; Detection: 0.12 mg/m

(AIHA);

geomean: 0.09 mg/m

(Amoore and Hautala); 0.025 mg/m

(Devos)

Susceptible population: Systemic: Low activities of the aldehyde dehydrogenase enzyme (ALDH), occuring in

subjects with point mutations in the corresponding gene, reduce acetaldehyde oxidation rate and conduce to intolerance

for ethanol. Acetaldehyde levels in the blood are induced by antialcohol drugs like disulfiram or nitrefazole.

Remarks: RD

: 7 g/m

; neither dose addition nor potentiating interactions occured upon exposure to aldehyde

mixtures (formaldehyde, acetaldehyde, acrolein) at exposure levels around the MOEL.

The INDEX project Final report

152

6. Risk Characterization

Cancer risk and health hazard evaluation

Based on short-term and long-term inhalation studies conducted in experimental animals, the upper respiratory tract is

the principal target site for effects of inhaled acetaldehyde. In short-term studies acetaldehyde caused degenerative non-

neoplastic effects. Although it is genotoxic both in vitro and in vivo, tumours have been observed following inhalation

only at concentrations that have produced significant cytotoxicity and it is likely that both the genotoxicity and irritancy

of acetaldehyde play a role in its carcinogenicity. At acetaldehyde concentrations for which the prevalence of sensory

irritation is low (i.e., < 45 mg/m

), risks of respiratory-tract cancers for the general population are considered

exceedingly low.

An Limit of Exposure (EL) has been here derived for short- and long term effects, based on acute eye irritation

observed in human volunteers and on chronic degeneration of the olfactory epithelium observed in rats. No distinction

has been done between short- and long-term ELs since the observed acute effects and endpoints are consistent with the

pathology seen in long-term studies. Also, two studies were considered for the EL derivation, in order to compensate for

the poor quality of the study on volunteers (Silverman et al., 1946). In this study, among 12 volunteers “several“

subjects after 15 min of exposure strenuously objected to acetaldehyde vapor at as low as 45 mg/m

with this supposed

threshold for sensory irritation set here as a LOAEL (eye irritation was developed by most of the subjects at 90 mg/m

This level has been devided by an assessment factor of 200, which incorporates uncertainty contributions for

extrapolation from a LOAEL to a NOAEL (10), intraspecies variation (10) and the poor quality of the data (2), resulting

in an Exposure Limit (EL) of 0.2 mg/m

. Also, a NOAEL was derived from a subchronic (4 weeks) animal study at 50

mg/m

, with an exposure limit of 0.5 mg/m

(AF=100) resulting from the extrapolation from an animal study (10) and

intraspecies variation (10).

In addition to the EL derivation, the present assessment takes into account the possible contribution to inhalatory

exposure of the endogenous generation of acetaldehyde in exhaled air. Acetaldehyde is an endogenous metabolite of

sugars and could also be locally formed from the microflora inhabiting the upper airways and mouth. Concentrations

expelled in breath span from 0.009 to 0.026 mg/m

, with higher levels observed in smokers and abstinent alcoholics. It

is the first metabolite of ethanol with reported breath concentration increasing to 0.22 – 2.2 mg/m

in European subjects

who drank moderate doses of ethanol (0.4-0.8 g/kg bw) (Jones, 1995).

Percentage of population exposed beyond given threshold levels

human LOAEL

(acute)

LOAEL

100

LOAEL

200

45 mg/m

4.5 mg/m

0.45 mg/m

0.2 mg/m

Studies available

animal NOAEL

(subchronic)

NOAEL

100

Description (Study, Year) N 50 mg/m

5 mg/m

0.5 mg/m

Paris – refurbished flats (Clarisse, 03) 61 < < < <

Helsinki (Expolis, 97) 15 < < < <

French National survey 7-d TWA (IAQ

observatory, 2003-04)

201 < < < <

< out of the evaluation range (i.e. <5% of the environments investigated)

Comments

Only three exposure studies were found, enabling the construction of cumulative frequency distributions of indoor

exposure to acetaldehyde. Median (90th percentile) indoor concentrations were 0.010 mg/m

(0.022) in Paris, 0.016

mg/m

(0.030) in Helsinki, and 0.014 mg/m

(0.040) in the French National Survey, respectively. Ambient air

concentrations over Europe range from 1-3 mg/m

in northern countries, up to 8 and 15 mg/m

recently measured in

Italy and Greece, respectively.

The INDEX project Final report

153

Result

The results from only indoor air monitoring surveys allow a crude estimate of average acetaldehyde concentrations in

European residences. Median concentrations (10-20 µg/m

) are one order of magnitude lower than the Exposure Limit

set here at 200 µg/m

and are within the same range of concentrations occuring in exhaled breath following its

endogenous production in the general populaton, not taking into account increases resulting from the consumption of

alcoholic beverages. Considering that exogenous acetaldehyde peak exposures are mainly associated with tobacco

smoke, concentrations in the order of the Exposure Limit could be expected following intense sigarette consumption.

Assuming that the available exposure data are indicative of the population residential exposure it is concluded that

people in Europe do not experience increased health hazards associated with acetaldehyde levels in their homes,

although additional work should be warranted for a better characterization of exposure and dose response. Also,

measured indoor levels are lower than a presumed threshold for cytotoxic damage to the nasal mucosa, and hence

considered low enough to avoid any significant risk of upper respiratory tract cancer in humans.

The INDEX project Final report

154

Toluene

Synonyms: Methyl benzene, methyl benzol, phenyl methane, toluol

CAS Registry Numbers: 108-88-3

Molecular Formula: C

1. Compound identification

Toluene is a clear, odorous, colourless, readily volatile liquid at ambient conditions. It is flammable and explosive in

air. The technical toluene may contain small amounts of benzene. Toluene will not react with dilute acids or bases and it

is not corrosive. It is removed from air by reacting with hydroxyl radicals. It is one of the most prevalent hydrocarbons

in the troposphere. The lifetime of toluene range from several days in summer to several months in winter. Toluene is

produced in high volume in industrial processes such as catalytic conversion of petroleum and aromatisation of aliphatic

hydrocarbons, and coke oven operations. It is used in blending gasoline to enhance octane ratings, in leather tanning and

as a common solvent. Occupational exposure to toluene might be high during production and use of toluene-containing

products. Toluene is common in many products such as paints, household aerosols, thinners, cleaning agents, coatings,

rubber, nail polish and other cosmetics, adhesives, resin and printing products (IARC 1999, WHO 1986, WHO 2000,

HSDB 2003)

2. Physical and Chemical properties

Molecular weight (g/mol) 92.14

Melting point (°C) -94.9

Boiling point (°C) 110.6

Density (g/l at 20 °C, 1 atm) 0.86

Relative density (air =1) 3.2

Solubility: Soluble in ethanol, benzene, diethyl ether, acetone,

chloroform, glacial acetic acid and

carbon disulfide, insoluble in water

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 3.824 µg/m

1 µg/m

= 0.261 ppb

Sources: WHO (1986), Verschueren 2001, EPA/Cal (2003), HSDB (2003)

3. Indoor Air Exposure assessment

Emission sources

Toluene emissions to the atmosphere result from industrial point sources and gas stations, and from mobile sources such

as traffic. Exposure to toluene in indoor environment occurs due to emissions from a variety of toluene-based household

products, uses of paints and thinners, together with tobacco smoke (WHO 2000).

Indoor air and exposure concentrations

Average indoor concentrations of toluene were clearly higher than respective outdoor concentrations in all latitudes in

Europe (Annex 3). Toluene was the most abundant compound in indoor air accounting for 33-54% of the total sum of

The INDEX project Final report

155

aromatics measured in the European EXPOLIS study (Saarela et al 2003). Average outdoor concentrations ranged from

about 6 µg/m

in Helsinki and Basel to 43 µg/m

in Milan (Jantunen et al 1999). Average indoor concentrations ranged

from 20 µg/m

in Helsinki to 74 µg/m

in Prague. Personal exposures were clearly higher than indoor concentrations.

The highest average personal exposures were measured in German study, being 130 µg/m

(Hoffman et al 2000), and

lowest in Helsinki, 25 µg/m

. In a German population based study of children and teenagers (GerES IV), weekly mean

exposure to toluene was higher than the mean residential indoor concentration being 25 µg/m

and 15 µg/m

respectively. The highest exposure and indoor concentrations were 280 µg/m

and 87 µg/m

respectively. (Ullrich et al

2002). Italian study carried out by Carrer et al (2000) determining 24-hour exposures of office workers in Milan,

showed an average of 35 µg/m

. Based on these results, it is obvious that at least in Central and Southern Europe the

guideline value of 260 µg/m

set by WHO (2000) will be exceeded in certain subpopulations.

Median and 90

percentile indoor concentrations, 10 µg/m

and 49 µg/m

, in Arizona, USA (Gordon et al 1999) were

clearly lower than the respective concentrations measured in EXPOLIS in European cities.

Ambient air concentrations of toluene in Europe have been measured in several studies (Saeger et al 1995, Ballesta et al

1997, Ballesta et al 1998, Kemp et al., 1998). Typical ambient air concentrations of toluene were in the order of 10 to

30 µg/m

. The concentrations of toluene near petrol stations were generally around an order of magnitude higher than

the urban air concentrations (EPA 2003).

Ambient toluene concentrations in urban environments in US have been in the same range as in Europe ranging about

17-20 µg/m

, the highest level being 62 µg/m

in a metropolitan area. In rural sites levels were typically below 1 µg/m

(Helmig and Arey 1992, WHO 2000).

Residential indoor air concentrations of toluene in European urban populations measured in the EXPOLIS study and the

National Survey in England are presented in Figure 3.1 and Figure 3.2. Respective exposure concentrations are

presented in Figure 3.3.

Source related short time concentrations of toluene are presented in Figure 3.4. Even higher concentration of toluene,

2.1 mg/m

was measured in a kindergarten in Athens in the AIRMEX project (Kotzias 2004). The source of toluene

identified in the kindergarten was a spray that was used when preparing Christmas decorations. Lee and Wang (2004)

studied emissions of incense burning in chamber (18 m

) tests. Concentrations ranged from 5 µg/m

to 99 µg/m

Measured concentrations did not exceed the Recommended Indoor Air Quality Objectives for Office Buildings and

Public Places in Hong Kong (HKIAQO) of < 1092 µg/m

. Brown (2002) studied short time (30-50 min) concentrations

of toluene in established and new buildings, reporting mean concentrations of 14 µg/m

and 250 µg/m

respectively.

Toluene concentration in a new building decreased to 18 µg/m

and 6.9 µg/m

in 72 and 246 days respectively.

Emission sources of toluene were not identified, but VOC emissions in general were related to water-based paints,

adhesives and wood-based panels. Elke at al (1998) measured short time concentrations of toluene inside buildings (60-

min average), inside a train and a car (30-min average), being at the same level, 55 µg/m

, in a train and a car, and 21

µg/m

in a smoker day room.

Short time exposure scenarios for potential consumer activities were analysed in European Union Risk Assessment

Report (EU 2003). The highest toluene concentration, 1000 mg/m

(note the unit, mg/m

) were expected during

spraying painting. High peak concentrations were also assessed during carpet laying, car polishing and gluing, 195

mg/m

, 10 mg/m

and 7.1 mg/m

respectively.

The INDEX project Final report

156

100

0 100 200 300 400 500

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of 30-hour indoor air concentrations of toluene in Athens (Ath, n=42),

Basel (Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=41) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS

2002).

100

0 1020304050607080

Indoor concentration (µg/m

)

Cumulative frequency (%)

Toluene 28-day

Figure 3.2. Cumulative frequency distribution of 28-day indoor air concentrations of toluene in England (GM= 15.1

µg/m

, max 1784 µg/m

n=796, Brown et al 2002).

The INDEX project Final report

157

100

0 100 200 300 400 500

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.3. Cumulative frequency distributions of 48-hour personal exposure concentrations of toluene in Athens (Ath),

Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean concentrations of the

German Survey GerES II (GeS) (Hoffman et al 2000).

100

150

200

250

300

Incence

burning

New

buildings

In a car Room with

ETS

Train with

ETS

Concentration (µg/m

)

Toluene

Figure 3.4. Short time toluene concentrations related to specific microenvironments or emission sources. Sampling

times of ‘New buildings’ (Brown 2002), ‘In a car’, ‘Room with ETS’, and ‘Train with ETS’ (Elke at al 1998) were 30-

50 min, 30-min, 60-min, 30-min respectively. During incense burning grab samples were obtained (Lee and Wang

2004).

The INDEX project Final report

158

4. Toxicokinetics

Absorption

The major uptake of toluene vapour is through the respiratory system. A number of investigations in humans (Carlsson,

1982; Carlsson and Lindqvist, 1977; Nomiyama and Nomiyama, 1974; Åstrand 1975) have shown that at rest a three-

hour exposure to toluene vapour will result in an uptake amounting to approximately 50% of the inhaled toluene.

The blood/air partition coefficient for toluene is 11,2-15,6 at 37°C (Lindqvist, 1977; Sato et al., 1972; Sato and

Nakajima, 1977; Sherwood, 1976; Ulfvarson and Övrum, 1976).

The concentration of toluene in alveolar air and in arterial and venous blood rises quickly during the first 10-15 minutes

of exposure (Carlsson, 1982; Åstrand et al., 1972). After only 10 seconds of exposure toluene can be detected in blood

from brachial arteries (Åstrand et al., 1972).

Data from experimental exposure of voluntary study subjects show that physical work results in increased toluene

uptake (Veulemans and Masschelein, 1978b; Carlsson, 1982). Using a 50 W workload, exposure to 300 mg/m

(80

ppm) toluene for 2 hours did not result in steady state of the blood concentration of toluene in 12 study subjects. The

toluene uptake was 2,4 times higher than the uptake at rest. During the work, lung ventilation was increased 2,8 times.

Concentrations of toluene in alveolar air and blood increased with increasing work loads (0-150 W in periods of 30

minutes) (Carlsson, 1982). The amount of toluene absorbed increased with greater amounts of body fat (Carlsson and

Ljungquist, 1982).

In nine male volunteers exposed to 200 mg/m

(53 ppm) toluene for 2 hours during a workload of 50 W, the total uptake

of toluene was 50% of that inhaled (Löf et al., 1993).

Distribution

The distribution of toluene in the body is among other factors dependent on the tissue/blood partition coefficients and

the metabolism. In rabbits the following partition coefficients have been found: brain, heart, liver, and intestine: 2,3,

muscle tissue: 1,6, adipose tissue: 74,3, bone, connective tissue, and lung tissue: 1,9 (Sato et al., 1974). In rats

brain/blood ratios of 1,2 (Kishi et al., 1988) and 1,7 (Zahlsen et al., 1992) have been determined. In humans the adipose

tissue/blood partition coefficient for toluene is determined to be 81-83 (Sato et al., 1974; Sherwood, 1976). This high

partition coefficient suggests that toluene can be accumulated in adipose tissue.

In mice, the distribution of toluene and its metabolites was investigated using whole body autoradiography after acute

inhalation of side chain marked

C toluene (Bergman, 1979; Bergman, 1983). In adipose tissue, bone marrow, spinal

nerves, spinal cord, and in the white parts of the brain, high concentrations of radioactivity occurred as judged from the

photographs. In blood, liver, and kidneys, radioactivity was also found. One hour after exposure nerve tissue showed no

radioactivity. In adipose tissue nearly all radioactivity had disappeared four hours after exposure, and only traces of

non-volatile radioactivity could be found in the liver. After 24 hours all radioactivity had disappeared from the body.

In rats, subcutaneous injection of toluene (100 or 500 mg/kg) resulted in maximum concentrations of blood toluene

after 2 hours (Benignus et al., 1984).

Toluene passes the placenta. Two hours following exposure of rats via inhalation to 1375 or 2700 mg/ m

(367 or 720

ppm) for 24 hours, foetal blood had a toluene concentration of 74% of that found in the dam’s blood. The amniotic

liquid contained a toluene concentration of 5% of that in the dam’s blood. Four and six hours after exposure, similar

relative toluene concentrations were found (Ungváry, 1984; quoted from IPCS 1985).

Groups of four 11, 14 or 17 day pregnant mice were killed 0, 30, 60 and 240 minutes after having inhaled 7500 mg/m

(2000 ppm)

C-toluene for ten minutes. Radioactivity as volatile and non-volatile was measured in lung, liver, kidney,

brain, cerebellum, fat, plasma, amniotic fluid, placenta, and foetus. It was shown that toluene immediately after

inhalation was taken up in the foetal tissue at a concentration of about 10% of that found in the maternal lungs. At four

hours after exposure the toluene radioactivity was decreased to 2% of the original value (Ghantout and Danielsson,

1986).

Toluene has been found in human breast milk. In 12 pooled samples from four urban areas in the United States, toluene

was identified qualitatively in at least 7 samples (Pellizzari et al., 1982; quoted from Jensen and Slorach, 1991).

The INDEX project Final report

159

Metabolism and elimination

The major pathway of toluene metabolism in both humans and laboratory animals involves sidechain oxidation by

sequential action of cytochrome P-450, alcohol dehydrogenase and aldehyde dehydrogenase leading to benzoic acid

which, upon conjugation with glycine, results in hippuric acid, the major urinary metabolite. Minor metabolites include

ortho- and para-cresol. A minor metabolite that is specific for toluene exposure is S-benzyl-N-acetyl-L-cysteine

(Takahashi et al., 1994). In vitro evidence suggests that the rate of metabolism in humans is greater than that in rats

(Chapman, et al., 1990).

Analysis of blood and urine samples from workers and voluntary study subjects exposed to toluene via inhalation in

concentrations ranging from 375 to 2250 mg/m

(100-600 ppm) indicate that of the biotransformed toluene, appr. 99%

is oxidised via benzyl alcohol and benzaldehyde to benzoic acid. The remaining 1% is oxidised in the aromatic ring,

forming ortho-, meta- and para-cresol (Woiwode et al., 1979; Woiwode and Drysch, 1981).

Water solubility of the biooxidation products is achieved through linkage with suitable substances (phase 2 reaction).

Benzoic acid is linked to either glycine or glucuronic acid forming either hippuric acid or benzoylglucuronide. Cresols

and benzyl alcohol are linked to glucuronic acid or sulphate (IPCS, 1985).

The elimination of toluene from adipose tissue is prolonged according to the findings of Periago et al. (1992), who

examined a worker population. The half-time for elimination from adipose tissue was reported to range between 0,5 and

2,7 days, depending on the amount of body fat. Elimination from bone marrow also is prolonged (Chemical and

engineering news, 1995; WHO, 1987).

Toluene or its metabolites may be eliminated via the lungs, the kidneys, or the liver.

Data from experimental inhalation exposure of voluntary subjects show that the toluene concentration in expired air

decreases rapidly during the first 10 to 20 minutes after cessation of exposure to toluene via inhalation (Veulemans and

Masschelein, 1978a; Carlsson, 1982; Echeverria et al., 1989). Two to four hours later, very low toluene concentrations

are found in expired air (Carlsson, 1982). Of the toluene absorbed, 15-20% is exhaled during the first few hours after

exposure has stopped (Nomiyama and Nomiyama, 1974). The cumulative elimination of toluene via the lungs amounts

to 4-8% and 7-14% after 2 and 20 hours, respectively (Carlsson, 1982).

The cumulative elimination (in per cent) of toluene via the lungs appears to increase with increasing amounts of toluene

taken up (Carlsson, 1982).

The majority (80-90%) of absorbed toluene is biotransformed and excreted from the body via the kidneys. At an

exposure level of 750 mg/m

(200 ppm), the excretion is mainly as hippuric acid. About 1% of the biotransformed

toluene is excreted as glucuronides or sulphates of o-, m-, or pcresol (IPCS, 1985).

A very small proportion, approximately 0,06% of the toluene absorbed via inhalation, is excreted unchanged in the

urine in humans (Williams, 1959).

A good correlation was found between toluene exposure (air concentration multiplied by time) and concentration of

hippuric acid in post exposure urine. However, a background level of hippuric acid is present in human urine, as a

product of endogenous metabolism, and of metabolism of substances present in food. In the Western part of the world,

at exposure levels below 375 mg/m

(100 ppm) hippuric acid in post exposure urine cannot be used to separate an

exposed person from an unexposed one because the difference between the background level and the toluene-generated

level is too small (Lauwerys, 1983). However, hippuric acid background levels in urine vary geographically. In some

Third World countries a low urinary hippuric acid background level is found. Thus, in these parts of the world it is

possible to use this metabolite as a biological marker for toluene exposure even at exposure levels lower than 375

mg/m

(Chang et al., 1996; Vrca et al., 1997a; 1997b).

Data from rats exposed to 3750, 6675, or 11250 mg/m

(1000, 1780, or 3000 ppm) of toluene via inhalation for 2 hours

showed that elimination could be described by a bi-exponential function with average half-times of 6 and 90 minutes

(Rees et al., 1985).

In rats a small proportion, less than 2%, of the absorbed toluene is excreted via the bile to the intestine. The substances

excreted are reabsorbed in the intestine. Thus very small amounts are excreted in faeces (Abou-El-Makarem et al.,

1967).

The INDEX project Final report

160

5. Health effects

Effects of short-term exposure

Dysfunction of the central nervous system and narcosis are the major effects of acute exposure to toluene (ATSDR,

1989). Irritation of the skin, eye, and respiratory tract can also result. Inhalational abuse of toluene with high level

exposure for long periods of time has produced progressive and irreversible changes in brain structure and function

(Spencer and Schaumberg, 1985).

Reaction time and perceptual speed were studied in 12 young male subjects exposed by inhalation to toluene

concentrations ranging from 400 to 3000 mg/m³ (100 to 700 ppm), each for a 20-minute interval (Gamberale and

Hultengren, 1972). Statistically significant impaired reaction time was apparent following exposure to 1100 mg/m³ (300

ppm) toluene. A statistically significant impairment in perceptual speed was observed at 3000 mg/m³ toluene. No

effects were observed at 400 mg/m³.

Two groups of middle aged workers, one with previous occupational exposure to solvents and one without, were

exposed once to 400 mg/m³ (100 ppm) of toluene for 6,5 hours (Baelum et al., 1985). Fatigue, sleepiness, a feeling of

intoxication, and eye, nose and throat irritation were reported. Decrements in manual dexterity, color discrimination,

and accuracy in visual perception were also observed. Greater sensitivity to toluene was noted for those subjects with

previous solvent exposure.

A random population sample of 32 male and 39 female subjects were allocated into three groups, one exposed to clean

air, one exposed to constant toluene at a concentration of 377 mg/m

(100 ppm), and one exposed to varying

concentrations of toluene (Fourteen 30-minute episodes during the exposure period, each episode starting with an

increasing concentration reaching a peak of 1125 mg/m

(300 ppm) after 5 min and then decreasing to a stable period of

about 15 min at 188 mg/m

(50 ppm), giving a TWA of 375 mg/m

for the whole exposure period). An exposure period

comprised a single day (7 hours). Toluene caused throat and respiratory irritation, headache and dizziness. In

performance tests only minimal effects were found, with a tendency towards lower score and more errors but fewer

false reactions in the primary task of the vigilance test. The effects were not statistically significant (p<0,1). There was

no difference between constant exposure and peak exposure (Bælum et al., 1990).

Nasal mucus flow, lung function, psychometric performance, and subjective responses were studied in 16 young

healthy males exposed to toluene concentrations ranging from 38 mg/m³ to 377 mg/m³ (10 to 100 ppm) for 6 hours

(Andersen et al, 1983). Headaches, dizziness, a feeling of intoxication, and slight eye and upper respiratory irritation

were reported at 377 mg/m³. The subjects also reported that it became more difficult to participate in the battery of

psychometric tests and that their reaction time felt impaired at 377 mg/m³. No significant objective changes compared

to control exposures were observed in the performance test results. No symptoms were reported at 38 and 151 mg/m³.

A battery of neurobehavioral and performance tests was conducted among 42 young men and women exposed by

inhalation for 7 hours to 0, 283, and 566 mg/m

(0, 75, and 150 ppm) toluene (Echeverria et al., 1989). Statistically

significant decrements in visual short term memory, visual perception, and psychomotor skills were observed at 560

mg/m³ compared to control exposures. A dose-dependent increase in subjective symptoms of headache and eye

irritation was also observed.

Wilson (1943) reported that workers exposed to concentrations of commercial toluene ranging from 200 to 750 mg/m³

(50 to 200 ppm) for periods of 1 to 3 weeks experienced headaches, lassitude, and loss of appetite. At 750 to 2000

mg/m³ (200 to 500 ppm), symptoms of nausea, bad taste in the mouth, slightly impaired coordination and reaction time,

and temporary memory loss were also observed. Exposure to 2000 to 5600 mg/m³ (500 to 1500 ppm) resulted in

palpitations, extreme weakness, pronounced loss of coordination, and impaired reaction time. Red blood cell counts

were decreased and there were 2 cases of aplastic anemia. The hematologic effects were most likely caused by benzene

impurities (ACGIH, 1986).

Three volunteer subjects exposed by inhalation to toluene concentrations ranging from 200 to 400 mg/m³ (50 to 100

ppm), 8 hours per day, 2 times per week over 8 weeks experienced fatigue, drowsiness, and headaches (von Oettingen

et al., 1942). At 750 to 3000 mg/m³ (200 to 800 ppm), symptoms of muscular weakness, confusion, impaired

coordination, paresthesia, and nausea were also reported. After exposure to 3000 mg/m

, all 3 subjects reported

considerable aftereffects (severe nervousness, muscular fatigue, and insomnia) lasting several days.

The INDEX project Final report

161

Summary of short-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

151

EXP

369

1h-ADJ

377

EXP

923

1h-ADJ

CNS-mild, eyes, respiratory;

impaired reaction time, headache, dizziness,

feeling of intoxication, slight eye and nose

irritation

volunteers, 6h

Andersen et al. 1983 OEHHA (1999);

REL: 37

ATSDR (2003);

MRL: 3.8

151

EXP

283

EXP

CNS; NOAEL: see above

LOAEL: visual short term memory, visual

perception and psychomotor skills

volunteers, 6-7h

NOAEL: Andersen et

al. 1983

LOAEL: Echeverria et

al. 1989

ECB (2003)

Final statement (UNIMI) : Human NOAEL = 150 mg/m³

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Acute Reference Exposure Level (protective against mild adverse effects):

Study Andersen et al., 1983

Study population 16 young, healthy males

Exposure method inhalation

Critical effects impaired reaction time and symptoms of headache, dizziness, a feeling of

intoxication and slight eye and nose irritation

LOAEL 100 ppm

NOAEL 40 ppm

Exposure duration 6 hours

Extrapolated 1 hour concentration 98 ppm (40

ppm* 6 h = C

* 1 h )

LOAEL uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 10

Reference Exposure Level 9,8 ppm (37 mg/m³; 37.000 µg/m³)

ECB (2003)

European Chemical Bureau - Institute for Health and Consumer Protection - European Union Risk Assessment Report

In the exposure chamber studies: Echeverria et al., (1989) and Andersen et al., (1983), headache, dizziness, feeling of

intoxication, irritation and sleepiness were recorded to occur with significantly increased frequency at exposure levels

from 562 mg/m

(150 ppm) down to 283 mg/m

(75 ppm). At 151 mg/m

(40 ppm) and below the effects have not been

recorded to occur with increased frequency. For these subjective symptoms a LOAL of 283 mg/m

(75 ppm) and a

NOAEL of 151 mg/m

(40 ppm) can be established.

With respect to function in performance tests, inhalation of 281 mg/m

(75 ppm) and 562 mg/m

(150 ppm) for 7 hours

have resulted in significantly worse results in a number of performance tests, indicating a LOAEL of 281 mg/m

(75

ppm) for function in performance tests while a NOAEL cannot be established.

ATSDR (2003)

Agency for Toxic Substances and Disease Registry

Derivation of the acute Minimal Risk Level (MRL).

There are several human studies for which the central nervous system is the major end point and could have been used

to derive an acute inhalation MRL. However, the Andersen et al. (1983) study was chosen as the basis for the MRL

because this was the only human study which reported a NOAEL. Baelum et al.(1985) also reported a LOAEL of 377

mg/m

for neurological effects in humans. In this study, 43 occupationally-exposed subjects and 43 controls were

exposed to either clean air or air containing 377 mg/m

toluene for 6.5 hours in a climate chamber. A battery of ten tests

of visuomotor coordination, visual performance, and cortical function were administered during the 6.5 hour period. For

The INDEX project Final report

162

both the controls and toluene exposed subjects, there were complaints of air quality, irritation of the nasal passages, and

increased feelings of fatigue and sleepiness. Subjects also complained of headaches and dizziness. Toluene exposure

decreased performance on four of the neurobehavioral tests; three of these were tests of visual perseverance. The fourth

test affected was the simple peg board test of visuomotor function, where the effect was noted in toluene-exposed

workers to a much greater extent than controls. Escheverria et al. (1991) reported a LOAEL of 283 mg/m

for

neurological effects in humans. In this study, two groups of 42 students were exposed to 0, 75, and 150 ppm toluene for

a 7 hour period. A complete battery of 12 tests was administered before and at the end of each exposure. Toluene

caused a dose-related impairment of function on digit span pattern recognition, the one hole test, and pattern memory.

Rahill et al. (1996) reported a LOAEL of 377 mg/m

for neurological effects in humans. In this study, six volunteers

were exposed for 6 hours a day to either 377 mg/m

toluene or clean air. Three repetitions of two computerized

neuropsychological tests were performed, with the composite score on the multitasking test being significantly lower

with toluene exposure than with clean air.

Dose endpoint used for MRL derivation: 151 mg/m

for neurological effects

Uncertainty factors used in MRL derivation:10 (for human variability)

MRL = 151 mg/m

x 5 days/7 days x 8 hours/24 hours ÷10 = 3.8 mg/m

Effects of long-term exposure

Most human studies reporting adverse effects due to chronic toluene exposures involve either toluene-containing

solvent abuse or occupational exposure to toluene. Solvent abusers are generally exposed to higher levels of toluene

than are workers. A continuum of neurotoxic effects ranging from frank brain damage to degraded performance on

psychometric tests which roughly track exposure levels has been observed.

An extensive database on human exposure to toluene indicates that dysfunction of the CNS is of primary concern.

Deficits in neurobehavioural functioning have been viewed as precursors of more serious indications of CNS toxicity

(Chemical and engineering news, 1995; WHO, 1981). Their measurement is generally held to be more reliable than

subjective symptoms, which may involve situational factors unrelated to a cause–effect relationship with exposure.

Potential confounders such as age, alcohol, drugs and education need to be assessed to ensure correlations of toxicity

with exposure.

Solvent workers exposed to 160 mg/m

toluene (estimated as a time-weighted average) for an average duration of 6,8

years reported a significantly greater incidence of sore throat, dizziness and headache than controls; the sore throat and

headache incidence demonstrated a rough dose-response (Yin et al., 1987).

Orbaek and Nise (1989) examined the neurological effects of toluene on 30 rotogravure printers, 33-61 years of age

(mean 50), employed at two Swedish printing shops for 4-43 years (median 29) in 1985. Mean exposure levels at the

two printing shops were 43 and 157 mg/m

of toluene, respectively; however, before 1980 the exposure levels had

exceeded 300 mg/m

in both shops. The authors noted that rotogravure printing provides an occupational setting with

practically pure toluene exposure. Comparisons were made to a reference group of 72 men aged 27-69 (mean 47). The

alcohol consumption of both the workers and referents was also determined (< 200 g/week or > 200 g/week).

Neurological function in the workers and referents was evaluated using interviews and psychometric testing; the results

from each of the two printing shops were pooled. The printers reported statistically significantly higher occurrences of

fatigue (60%), recent short-term memory problems (60%), concentration difficulties (40%), mood lability (27%), and

other neurasthenic symptoms. The printers also scored significantly worse than referents in a number of psychometric

tests, including synonym, Benton revised visual retention and digit symbol tests, even after adjustment for age. For all

comparisons, tests of interaction between the effects of toluene exposure and alcohol consumption were not statistically

significant.

A battery of neurobehavioral tests was performed in 30 female workers exposed to toluene vapors in an electronic

assembly plant (Foo et al., 1990). The average number of years worked was 5,7 ± 3,2 for the exposed group and 2,5 ±

2,7 years for the controls. Study subjects did not smoke tobacco or drink alcohol, were not taking any medications, and

had no prior history of central or peripheral nervous system illness or psychiatric disorders. The exposed group of

workers inhaled a time-weighted average (TWA) of 332 mg/m

(88 ppm) toluene while the control workers inhaled 49

mg/m

(13 ppm). A significant decrease in neurobehavioral performance was observed in the exposed workers in 6 out

of 8 tests. Irritant effects were not examined, and concurrent exposures to other chemicals were not addressed. In this

study, 332 mg/m

was considered a LOAEL for central nervous system effects. However, the workers designated by the

authors to be controls did not comprise a true control group, since they were exposed to 49 mg/m

toluene. This may

have resulted in an underestimation of the effects of exposure to 332 mg/m

toluene. Similar effects were noted in a

follow-up study by Boey et al. (1997).

The INDEX project Final report

163

Abbate et al. (1993) evaluated alterations induced in the auditory nervous system by exposure to toluene in a group of

rotogravure workers. A sample of 40 workers of normal hearing ability was selected from a group of 300 workers who

were apparently in good health but were professionally exposed to toluene (12 – 14 years exposure, 343 mg/m

average

exposure, exposure assessment not described). They were subjected to an adaptation test utilizing a BAER technique

with 11 and 90 stimulus repetitions a second. The results were compared with an age and sex-matched control group not

professionally exposed to solvents. A statistically significant alteration in the BAER results was noted in the toluene-

exposed workers with both 11 and 90 stimuli repetitions. The authors suggested that these results can be explained as a

tolueneinduced effect on physiologic stimulus conduction mechanisms, even in the absence of any clinical sign of

neuropathy. Furthermore, this effect could be observed in the responses of the entire auditory system, from peripheral

receptors to brainstem nuclei.

A group of 49 printing-press workers occupationally exposed to toluene for approximately 21,6 years was studied by

Vrca et al. (1997). Toluene exposure levels were determined from blood toluene and urinary hippuric acid levels, and

were estimated to range from 151-226 mg/m

. No control group was used. Brain evoked auditory potential (BEAP;

similar to BAER) and visual evoked potential (VEP) measurements were performed on a Monday morning after a

nonworking weekend. There was a significant increase in the latencies of all the BEAP waves examined, except for P2

waves, as well as in the interpeak latency (IPL) P3-P4, while IPL P4-P5 decreased significantly with the length of

exposure. No correlation was noted between the amplitude of BEAP waves and the length of exposure. The amplitude

but not the latency of all the VEPs examined decreased significantly with the length of exposure.

The effects of acute and chronic toluene exposure on color vision were studied in a group of eight rotogravure printing

workers (Muttray et al., 1999). The workers had been employed as printers for an average of 9,8 years. The color vision

acuity of the workers before and after an acute toluene exposure (28– 41 minutes in duration, concentration 1115 –

1358 mg/m

) was evaluated using the Farnsworth panel D-15 test, the Lanthony desaturated panel D-15 test, and the

Standard Pseudoisochromatic Plates part 2. A control group of 8 unexposed workers was also tested. Acute toluene

exposure had no effect on color vision. Print worker performance prior to acute toluene exposure (chronic effects) was

similar to controls on the Farnsworth panel D-15 and Standard Pseudoisochromatic Plates part 2 tests. Print worker

performance on the Lanthony desaturated panel D-15 test was worse than that of controls (median scores of 1,18 and

1,05 for exposed and controls (higher number indicates degraded performance), respectively, but not significantly (p =

0,06). The authors noted that the small number of subjects limited the statistical power of the study.

Three groups of Croatian workers were examined by means of interviews, medical examination, and color vision testing

using the Lanthony 15 Hue desaturated panel in standard conditions (Zavalic et al., 1998a). Workers were excluded

from the study if they met any of the following criteria: less than 6 months employment, congenital color vision loss, a

medical condition which can affect color vision, visual acuity below 6/10, use of medications which can affect color

vision or a hobby that involved solvent exposure. Alcohol intake and smoking were also assessed for each individual.

The first group consisted of 46 workers (43 women and 3 men) employed in manually glueing shoe soles and exposed

to median levels of 121 mg/m

and geometric mean levels of 132 mg/m

toluene. The second group consisted of 37

workers (34 men and 3 women) employed in a rotogravure printing press and exposed to median levels of 498 mg/m

and geometric mean levels of 588 mg/m

toluene. The third group consisted of 90 workers (61 men and 29 women) not

occupationally exposed to any solvents or known neurotoxic agents. The study demonstrated a statistically significant

impairment of color vision in workers chronically exposed to 588 mg/m

toluene compared with controls. When the

data were adjusted to allow for the confounding effects of alcohol consumption and age, a significant difference due to

toluene exposure was also reported for workers exposed to 132 mg/m

toluene compared with controls.

Zavalic et al. (1998b) examined the effects of chronic occupational toluene exposure on color vision using a further

group of 45 exposed workers (mean toluene exposure concentration = 452 mg/m

) and 53 controls. Color vision was

evaluated using the Lanthony desaturated panel D-15 test; test scores were age and alcohol consumption-adjusted.

Color vision was significantly impaired in toluene-exposed workers (p < 0,0001) compared to controls. It was also

observed that there was no significant difference between test scores on Monday morning (prework) and Wednesday

morning. The authors stated that the effect of toluene on color vision can be chronic and that the possible recovery

period is longer than 64 hours.

Effects on kidney and liver

Among 24 toluene abusers examined on the day of admission to hospital, alkaline

phosphatase was elevated in 13 patients, while SGOT was elevated in 7. The elevated enzyme levels returned to normal

after 2 weeks of abstinence (Fornazzari et al., 1983).

No increase in levels of the enzymes serum aspartate aminotransferase and alanine aminotransferase was found, in 59

men with occupational exposure to toluene (recorded level 375 mg/m

) for more than one year (1-5 years, 22 men; 6-10

The INDEX project Final report

164

years, 18 men; more than 10 years, 19 men) when compared to an unexposed control group of equal size (Waldron et

al., 1982).

In 47 toluene-exposed workers a significant increase in S-ALP (20% relative to referents) compared with a referent

group of 46 non-exposed workers was found (Svensson et al., 1992b). The association was still significant when heavy

alcohol consumers were excluded from the analysis. The exposure levels were generally below 300 mg/m

(80 ppm).

Other liver functionrelated enzyme levels were unaffected. There was no association with cumulative exposure.

Sniffing of toluene resulted in reversible kidney damage (O'Brian et al., 1971), haematuria (Massengale et al., 1963),

reversible type 1 renal tubular acidosis (Bennett and Forman, 1980; Fischman and Oster, 1979; Kroeger et al., 1980;

Moss et al., 1980; Patel and Benjamin, 1986; Reisin et al., 1975; Streicher et al., 1981; Taher et al., 1974; Weinstein et

al., 1985; Will and McLaren, 1981) and hypokalaemia (Kelly, 1975; Taher et al., 1974). In some cases sniffing resulted

in irreversible damage of the kidneys (Russ et al., 1981).

A workplace accident with massive toluene exposure for 18 hours resulted in renal failure with oligouria probably

caused by dehydration and myoglobinuria (Reisin et al., 1975).

Inhalation of 382 mg/m

(100 ppm) toluene for 6,5 hours in an exposure chamber resulted in unchanged excretion of

albumin and beta-2-microglobulin for 43 printers with occupational exposure to toluene as compared to 43 age-matched

controls without occupational exposure to toluene (Nielsen et al., 1985).

No signs of renal damage in 118 painters were found compared with a control group. The painters had an average of 9

years occupational exposure to toluene and xylenes. At the time of investigation, the exposure was approximately 94

mg/m

(25 ppm) as determined from metabolites in urine (Franchini et al., 1983).

In 42 printers with an occupational toluene exposure averaging 300 mg/m

(80 ppm) (range 100-900 mg/m

(30-240

ppm)) compared with 48 unexposed controls, no changes in glomerular filtration rate, renal concentrating ability, beta-

2-microglobulin excretion, and excretion of erythrocytes and leukocytes were found (Askergren, 1982).

In conclusion, massive toluene exposure through abuse or workplace accidents has been associated with kidney damage

and renal failure. Three occupational studies did not show a relation between toluene exposure and kidney damage.

Developmental, reproductive and teratogenic effects

Abuse of toluene by pregnant women through deliberate inhalation of products such as paint thinners, glues and paints

has been associated with a number of developmental and congenital anomalies in infants (Donald et al., 1991). In such

reports it is difficult to determine the degree to which other substances may play a role in the development of adverse

effects. Effects commonly noted postpartum include low birth weight, growth retardation, microencephaly, CNS

dysfunction, renal tubule acidosis, and minor craniofacial and limb abnormalities. Perinatal death has also been

reported.

Two studies suggest an increased risk of spontaneous abortions associated with exposure to toluene in the workplace.

One of the studies provides no data on exposure levels, while the levels were around 332 mg/m

(range 189-566 mg/m

)

in the other study (Ng et al., 1992b). The Ng et al. (1992) study cannot be used to establish definitively a causal

relationship between late spontaneous abortions and toluene exposure or the magnitude of the LOAEL To establish a

definite relationship, a prospective study including pregnant women exposed to toluene at similar exposure levels with

individually monitored data on toluene exposure and fetal loss would be needed. However, based on the current

evidence suggesting an increased risk for late spontaneous abortions, exposure of pregnant women to such exposure

levels would raise serious ethical concerns. Consequently, the results of the Ng study are used as a basis for the risk

characterisation of developmental toxicity in humans.

Menstrual disorders in workers were reported, but the possible presence of other chemicals in the exposure

environments and unmatched characteristics of the exposed and control groups make these findings difficult to interpret

(WHO, 1987). Possible exposure-related effects upon follicle stimulating hormone and testosterone (Svensson et al.,

1992a), but not serum prolactin levels (Svensson et al., 1992b), have been observed in printers without overt effects of

toluene.

Developmental toxicity of toluene has mainly been studied in rats. Rat inhalation studies provide strong evidence of

developmental toxicity (lower birth weight and long-lasting developmental neurotoxicity) in the absence of maternal

toxicity. The effective dose levels are around or more than 3.77 g/m

. The NOAEL for lower birth weight and delayed

The INDEX project Final report

165

postnatal development is 2.26 g/m

(Thiel and Chahoud, 1997). A NOAEL for developmental neurotoxicity cannot be

determined from the available studies. The LOAEL for this effect is 4.52 g/m

(Hass et al., 1999).

There is no indication that toluene cause malformations in rats, mice or rabbits.

Summary of long-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

151

EXP

ADJ

301

EXP

ADJ

CNS; decreased brain weight, dopamine receptor

binding

Supporting study: Neurobehavioral deficits

rats, 4w

(supported by

occupational, 5.7y)

Hillefors-Berglund et

al. 1995; supported by

Orbaek and Nise

(1989), Foo et al.

(1990)

OEHHA (1999);

REL: 0.4

332

EXP

118

ADJ

CNS;

neurobehavioural deficits

Occupational, 5.7y

Foo et al. 1990 OEHHA (1999);

WHO (2001);

GV: 0.26;

EPA-IRIS (1992)

; RfC: 0.4

2261

EXP

437

ADJ

HEC

Respiratory system;

nasal epithelium degeneration

rats, 2y

NTP, 1990 EPA-IRIS (1992)

132

Study

ADJ

CNS;

color vision impairment

Occupational, >6m Zavalic et al. (1998a) ATSDR (2003);

MRL: 0.3

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Chronic Reference Exposure Level (REL):

Study Hillefors-Berglund et al. (1995); supported by Orbaek and

Nise (1989), Foo et al. (1990)

Study population Male Sprague-Dawley rats

Exposure method Inhalation

Critical effects Decreased brain (subcortical limbic area) weight, Altered dopamine

receptor (caudate-putamen) binding

LOAEL 80 ppm

NOAEL 40 ppm

Exposure continuity 6 hours/day, 5 days/week

Exposure duration 4 weeks, followed by 29-40 days recovery

Average experimental exposure 7 ppm (40 × 6/24 hours × 5/7 days)

Human equivalent concentration 7 ppm (gas with systemic effects, based on RGDR = 1,0

using default assumption that λa = λh

Subchronic uncertainty factor 10

Interspecies uncertainty factor 1 (see below)

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 100

Inhalation reference exposure level 0,07 ppm (70 ppb; 0,3 mg/m

; 300 µg/m

)

Supportive human study Foo et al., 1990

Study population 30 female workers in an electronic assembly plant

Exposure method Occupational inhalation

Critical effects Neurobehavioral deficits in 6 out of 8 tests

LOAEL 88 ppm

NOAEL Not observed

Exposure continuity 10 m

/day occupational inhalation rate, 5 days/week

Average occupational exposure 31,4 ppm (88 ppm x 10/20 x 5/7)

Exposure duration 5,7 + 3,2 years (exposed group); 2,5 + 2,7 years (controls)

LOAEL uncertainty factor 10

Subchronic uncertainty factor 3

The INDEX project Final report

166

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 300

Inhalation reference exposure level 0,1 ppm (100 ppb; 0,4 mg/m

; 400 µg/m

)

The critical animal study (Hillefors-Berglund et al., 1995) used to derive a REL for toluene describes adverse

neurological effects in rats after a well characterized inhalation exposure to toluene. The study results contain both a

LOAEL and a NOAEL. Decreased brain (subcortical limbic area) weight and altered dopamine receptor binding

compared to controls were noted at the NOAEL, but the changes were not statistically significant; this suggests that if a

threshold for adverse neurological effects exists in this study, it would be at or below the observed NOAEL. The study

LOAEL for altered dopamine receptor binding agrees qualitatively with results from similar studies (von Euler et al.,

1994).

Additionally, toluene induced neurotoxicity has been described in many studies by a variety of endpoints in both

animals and humans (ATSDR, 1999). The adverse neurotoxic effects associated with toluene exposure in the rat study

by Hillefors-Berglund et al. (1995), decreased brain (subcortical limbic area) weight and altered dopamine receptor

binding, occur in areas of the rat brain that are structurally and functionally similar to brain areas (basal ganglia,

thalami) of some human toluene abusers that demonstrate MRI alterations (T2 hypointensity). The altered MRI

parameters may be the result of the partitioning of toluene into the lipid membranes of brain cells (Unger et al., 1994).

If both human and animal adverse effect data on a chemical are available, OEHHA prefers to use the human data to

develop a REL when possible. However, the study by Hillefors-Berglund et al. (1995) provides data (decreased brain

[subcortical limbic area] weight and altered brain dopamine receptor binding) which are specific and sensitive measures

of neurotoxicity that would not be obtainable in human studies. In contrast, the psychometric tests used to generate the

neurotoxicity data in the human occupational exposure studies described above tend to be less sensitive and suffer from

greater measurement uncertainty. Additionally, the Hillefors-Berglund et al. (1995) study has better exposure

characterization than the human occupational exposure studies. Nonetheless, the human studies are useful in supporting

the derivation of the REL for toluene. Ordinarily, an interspecies uncertainty factor of 3 would be applied, in addition to

the human equivalent concentration calculation, to reflect the uncertainty associated with extrapolating from animals to

humans. However, in this case the uncertainty in the interspecies extrapolation is reduced by the availability of human

epidemiological data with generally consistent effect levels, after appropriate duration corrections. Based on

comparison of the data in both animals and humans, it appears that a REL of 271 µg/m

(rounded to 300 µg/m

in the

final derivation) would protect exposed humans from experiencing chronic neurotoxic effects.

Further studies considered by OEHHA for the derivation of the Reference Exposure are listed in Table 5.1.

Table 5.1: Supporting Neurotoxicity Studies

Study Duration Effect

NOAEL

mg/m

NOAEL

(TWA)

mg/m

LOAEL

mg/m

LOAEL

(TWA)

mg/m

VonEuler et al. (1994) 4 weeks rat: altered brain dopamine

receptor binding

n.o. n.o. 302 54

Orbaek and Nise

(1989) 29 years human: impairment on

neuropsychometric tests

n.o. n.o. 42-155 15-55

Korsak (1992) 6 months rat: impaired motor function n.o. n.o. 377 67

WHO (2001)

Air Quality Guidelines for Europe 2000

The lowest-observed-adverse-effect level for effects on the CNS from occupational studies, is approximately 332

mg/m

(88 ppm) (Foo et al., 1990). A guideline value of 0.26 mg/m

is established from these data adjusting for

continuous exposure (dividing by a factor of 4,2) and dividing by an uncertainty factor of 300 (10 for interindividual

variation, 10 for use of a lowest-observed-adverse-effect level rather than a no-observed-adverse-effect level, and an

additional factor of 3, given the potential effects on the developing CNS). This guideline value should be applied as a

weekly average. This guideline value should also be protective for reproductive effects (spontaneous abortions).

The air quality guideline could also be based on the odour threshold. In this case, the peak concentrations of toluene in

air should be kept below the odour detection threshold level of 1 mg/m

as a 30 minutes average.

The INDEX project Final report

167

U.S.EPA - IRIS (1992)

Integrated Risk Information System

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC)

Critical effect Exposures* UF MF RfC

Neurological effects

Occupational Study

Foo et al., 1990

NOAEL: None

LOAEL: 332 mg/m

(88 ppm)

LOAEL(ADJ): 119 mg/m

LOAEL(HEC): 119 mg/m

300 1

0,4 mg/m

Degeneration of

nasal epithelium

2-Year Rat Chronic

Inhalation Study

NTP, 1990

NOAEL: None

LOAEL: 2261 mg/m

(600 ppm)

LOAEL(ADJ): 437 mg/m

LOAEL(HEC): 79 mg/m

*Conversion Factors: MW = 92.15.

Foo et al., 1990: Assuming 25 C and 760 mmHg, LOAEL (mg/m

) = 88 ppm x 92.15/24.45 = 332 mg/m

. This is an extrarespiratory effect of a

soluble vapor. The LOAEL is based on an 8-hour TWA occupational exposure. MVho = 10 m

/day, MVh = 20 m

/day. LOAEL(HEC) =

LOAEL(ADJ) = 332 x MVho/MVh x 5 days/7 days = 119 mg/m

NTP, 1990: Assuming 25 C and 760 mmHg, LOAEL (mg/m

) = 600 ppm x 92.15/24.45 = 2261 mg/m

. LOAEL(ADJ) = LOAEL (mg/m

) x 6.5

hours/24 hours x 5 days/7 days = 437 mg/m

. The LOAEL(HEC) was calculated for a gas:respiratory effect in the extrathoracic region. MVa = 0.24

/day, MVh = 20 m

/day, Sa (ET) = 11.6 sq.cm, Sh (ET) = 177 sq.cm. RGDR = (MVa/Sa) / (MVh/Sh) = 0.18. LOAEL(HEC) = 437 x RGDR = 79

mg/m

In humans, toluene is a known respiratory irritant with central nervous system (CNS) effects. Because available studies

could not provide subthreshold (NOAEL) concentrations for either of these effects, the LOAELs for both effects need

to be considered in developing the RfC. Consequently, the study of Foo et al. (1990) was used for the CNS effects, and

that of the National Toxicology Program (NTP, 1990) for the irritant effects. Because the CNS effect was judged to be a

more severe and relevant endpoint, the LOAEL for this effect was used for deriving the RfC. Further, this effect is

supported by a number of other occupational studies that show effects around 377 mg/m

(100 ppm).

Foo et al. (1990) conducted a cross-sectional study involving 30 exposed female workers employed at an electronic

assembly plant where toluene was emitted from glue. Toluene levels reported in the study were from personal sample

monitoring and reported as an 8-hour TWA, although the number of samples taken and the actual sampling period were

not given. No historical exposure values were given. Co-exposure to other solvents was not addressed in the study. The

exposed and control cohorts were matched for age, ethnicity, and use of medications. Members of these cohorts did not

use alcohol and were nonsmokers. Medical histories were taken to eliminate any histories of central or peripheral

nervous system disorders. The average number of years (+/- SD) worked by the exposed population was 5,7 +/- 3,2 and

by the controls was 2,5 +/- 2,7. Exposed workers breathed toluene air levels of 332 mg/m

(88 ppm) as a TWA and

control workers 49 mg/m

(13 ppm) (TWA); both of which are averages of the individual personal samples. A battery

of eight neurobehavioral tests were administered to all exposed and control workers. The tests were performed

midweek, before the workers reported to their stations for the day. Group means revealed statistically significant

differences in 6/8 tests; all tests showed that the exposed workers performed poorly compared with the control cohort.

When individual test results were linearly regressed against personal exposure concentrations, poor concentration-

response relationships resulted for the six tests, with correlation coefficients ranging from 0,44 to 0,30. Irritation effects

were not evaluated in this study, and no clinical signs or symptoms were reported. The paucity of exposure information,

coupled with the small size of the cohort, limits the interpretation of this study, although the results were essentially

confirmed in a clinical study in which the toluene concentrations were carefully controlled (Echeverria et al., 1989) at

levels bracketing 332 mg/m

(88 ppm). Although the data in Echeverria et al. (1989) were generated from short- term

exposures (3-7 hours over a period of 142 days), the results may be considered relevant to longer-term exposures as

several studies indicate the absence of a duration-response relationship in toluene-induced symptomatology. Fornazzari

et al. (1983) noted the absence of a duration-effect relationship among toluene abusers when they were segregated into

neurologically impaired vs. unimpaired (p = 0,65). The human studies of Iregren (1982), Cherry et al. (1985), Baelum et

al. (1985), and the principal study of Foo et al. (1990) all report this lack of a duration-response relationship and

confirm the occurrence of CNS effects. Foo et al. (1990) indicate a LOAEL of 332 mg/m

(88 ppm) toluene for

neurobehavioral changes from chronic exposure to toluene.

In a 2-year bioassay, Fischer 344 rats (60/sex/group) were exposed to 0, 2261, or 4523 mg/m

(0, 600, or 1200 ppm,

respectively) toluene vapors, 6,5 hours/day, 5 days/week (duration-adjusted to 0, 437, and 875 mg/m

, respectively) for

103 weeks (NTP, 1990). To generate toluene vapor, the liquid material was heated, and the vapor diluted with nitrogen

and mixed with the chamber ventilation air. An interim sacrifice was carried out at 15 months on control and 4524

The INDEX project Final report

168

mg/m

(1200 ppm) groups (10/sex/group) to conduct hematology and histopathology of the brain, liver, and kidney.

Body weights were measured throughout the study. Gross necropsy and micropathology examinations were performed

at the end of the study on all major organs including the nasal passage tissues (three sections), lungs, and mainstem

bronchi. Mean body weights in both exposed groups were not different from controls for either sex. No exposure-

related clinical signs were reported, and survival rate was similar for all groups. At the interim sacrifice, there was a

mild-to- moderate degeneration in the olfactory and respiratory epithelium of the nasal cavity in 39/40 rats of the 600-

and 1200-ppm groups compared with 7/20 controls. At the end of 2 years, there was a significant (p<0,05) increase in

the incidence of erosion of the olfactory epithelium (males: 0/50, 3/50, and 8/49; females: 2/49, 11/50, and 10/50; at 0,

600, and 1200 ppm, respectively) and of degeneration of the respiratory epithelium (males: 15/50, 37/50, and 31/49;

females: 29/49, 45/50, and 39/50; at 0, 600, and 1200 ppm, respectively) in the exposed animals. The females exposed

to 600 and 1200 ppm also exhibited a significant increase in inflammation of the nasal mucosa (27/49, 42/50, and 41/50

at 0, 600, and 1200 ppm, respectively) and respiratory metaplasia of the olfactory epithelium (0/49, 2/50, and 6/50 at 0,

600, and 1200 ppm, respectively). A LOAEL of 2261 mg/m

toluene was determined for the concentration-dependent

increase in erosion of the olfactory epithelium in male rats and the degeneration of the respiratory epithelium in both

sexes. No NOAEL could be derived from this study.

ATSDR (2003)

Agency for Toxic Substances and Disease Registry

Derivation of a Minimal Risk Level (MRL):

Studies (Zavalic et al., 1998a; 1988c) demonstrated a statistically significant impairment of color vision in workers

chronically exposed to 588 mg/m

(156 ppm) toluene compared with controls. When the data were adjusted to allow for

the confounding effects of alcohol consumption and age, a significant difference due to toluene exposure was also

reported for workers exposed to 132 mg/m

(35 ppm) toluene compared with controls. Uncertainty factors used for

MRL derivation: 10 for use of a minimal LOAEL and 10 for human variability

MRL = 132 mg/m

x 5 days/7 days x 8 hours/24 hours ÷ 100 = 0.3 mg/m

Carcinogenic potential

There is no information from human studies that suggests that toluene has carcinogenic potential. Svensson et al. (1990)

examined a cohort of 1020 toluene-exposed workers who had been employed for a minimum period of three months

during the period 1925–1985. There were no significant increases in tumours and no cumulative dose–response

relationship in workers with an exposure period of at least five years and a latency period of 10 years. Exposure to

benzene had taken place up to the 1960s. Toluene has been shown, however, to hyperphosphorylate rat liver p53, a

tumour suppressor gene (Dees and Travis, 1994). This observation may be of concern, since a reduced ability of p53 to

suppress genetic errors may result in tumour formation.

Genotoxicity

An unequivocal evaluation of the genetic effects of occupational toluene exposure cannot be made because of the small

numbers of individuals analysed and insufficient information on possible exposure to other chromosome-damaging

agents (WHO, 1985). Recent data indicate that toluene induces clastogenic effects in pokeweed-mitogen-stimulated

peripheral blood lymphocytes of printers (Nise et al., 1991). However, variations in exposure history preclude

identification of an exposure–response relationship. Toluene exposure of printers was also highly correlated with an

excess of chromatid breaks in peripheral lymphocytes compared to controls (Pelclova et al., 1990), although concurrent

exposure to printing dyes as a factor cannot be excluded. No effects on sister chromatid exchange, cell cycle delay or

cell mortality were observed in peripheral blood lymphocytes in volunteers exposed for three consecutive days to 189

mg/m

(50 ppm) (Richer et al., 1993). Previously, Bauchinger et al. (1982) found a significant increase in sister

chromatid exchanges and chromosome aberrations in printers (smokers and nonsmokers relative to controls) exposed to

toluene for more than 16 years. Even after two years of exposure cessation a higher incidence of aberrations was

observed in exposed individuals compared to controls (Dudek et al., 1990).

Interactions with other chemicals

Exposure of volunteers to toluene 189 mg/m

and xylene 189 mg/m

, both being often found together in mixtures

such as paint thinners,

resulted in decrements in reaction time in one of a battery of psychomotor and cognitive tests

The INDEX project Final report

169

administered (Dudek et al., 1990). Toluene alone did not result in deficits in any of the tests. Similar

levels of both

solvents (189

mg/m

xylene, 151 mg/m

toluene) did not modify the conversion of either substance to its

urinary metabolites (Kawai et al. 1992b; Tardif et al. 1991). At higher concentrations (302 or 566

mg/m

xylene, 358 or 566 mg/m

toluene), the blood and exhaled air concentrations of both solvents were increased

compared to the controls exposed to either solvent alone, indicating that metabolism of both solvents was

decreased by the coexposure paradigm (Tardif et al. 1991, 1992).

Ethanol ingestion during exposure of volunteers to a toluene level of 302 mg/m

(80 ppm) for 4,5 hours was found not

to alter the occurrence or severity of the subjective symptoms associated with toluene alone (Iregren et al., 1986).

A number of studies with laboratory animals have shown that toluene interferes with the metabolism and toxicity of

several chemicals, including ethanol, benzene, xylene, hexane and styrene (Chemical and engineering news, 1995;

Plappert et al., 1994; Nylen et al., 1995)

Odour perception

Source: WHO, 1986

Odour threshold: 9,4 mg/m3

Source: American Industrial Hygiene Association, 1989

Characteristic : sour, burnt.

Odour Thresholds:

Detection: 6 mg/m3 ; range 0,6-140 mg/m3 (1,6 ppm; range 0,16 - 37 ppm)

Recognition: 41 mg/m3 ; range 7-260 mg/m3 (11 ppm; range 1,9 - 69 ppm)

Source: (Hoshika et al., 1993)

Barely perceptible concentration level of toluene: 3.5/4.2 mg/m

(in Japan/The Nederlands, respectively)

International comparison of odor threshold values of several odorants in Japan and in The Netherlands, given as the

barely perceptible concentration level revealed striking similarities for hydrogen sulfide (in Japan 0.0005 ppm/in The

Netherlands 0.0003 ppm), phenol (0.012/0.010), styrene (0.033/0.016), toluene (0.92/0.99), and tetrachloroethylene

(1.8/1.2) but not for m-xylene (0.012/0.12). Such a similarity was not found with any other literature sources.

The INDEX project Final report

170

Summary of Toluene Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Not relevant

Toxicokinetics: ~50% uptake of the inhaled amount

Health effect levels of short- and long-term exposure

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

151

EXP

369

1h-ADJ

377

EXP

923

1h-ADJ

CNS-mild, eyes, respiratory;

impaired reaction time, headache, dizziness,

feeling of intoxication, slight eye and nose

irritation

volunteers, 6h

Andersen et al. 1983 OEHHA (1999);

REL: 37

ATSDR (2003);

MRL: 3.8

151

EXP

283

EXP

CNS; NOAEL: see above

LOAEL: visual short term memory, visual

perception and psychomotor skills

volunteers, 6-7h

NOAEL: Andersen et

al. 1983

LOAEL: Echeverria et

al. 1989

ECB (2003)

Long-term exposure

151

EXP

ADJ

301

EXP

ADJ

CNS; decreased brain weight, dopamine receptor

binding

Supporting study: Neurobehavioral deficits

rats, 4w

(supported by

occupational, 5.7y)

Hillefors-Berglund et

al. 1995; supported by

Orbaek and Nise

(1989), Foo et al.

(1990)

OEHHA (1999);

REL: 0.4

332

EXP

118

ADJ

CNS;

neurobehavioural deficits

Occupational, 5.7y

Foo et al. 1990 OEHHA (1999);

WHO (2001);

GV: 0.26;

EPA-IRIS (1992)

; RfC: 0.4

2261

EXP

437

ADJ

HEC

Respiratory system;

nasal epithelium degeneration

rats, 2y

NTP, 1990 EPA-IRIS (1992)

132

Study

ADJ

CNS;

color vision impairment

Occupational, >6m Zavalic et al. (1998a) ATSDR (2003);

MRL: 0.3

Carcinogenicity: IARC: 3 ; U.S.EPA: D ; No indications of carcinogenicity

Genotoxicity: No clear indication of genotoxicity

Odour threshold: Range (detection): 0.6 - 140 mg/m³ , GM: 6 mg/m³ (AIHA); 9.4 mg/m³ (WHO);

Barely perceptible concentration level of toluene: 3.5/4.2 mg/m

(in Japan/The Nederlands, respectively;Hoshika et al.,

1993)

Susceptible population: No evidence

Remarks: RD

: 19 g/m

The INDEX project Final report

171

6. Risk Characterization

Health hazard evaluation of short- and long-term exposure

Toluene has low acute toxicity. In humans experimentally exposed to toluene, concentrations of 283 mg/m

and above

caused headache, dizziness, and feeling of intoxication, irritation and sleepiness. Furthermore, toluene causes impaired

neuropsychological function, as demonstrated in performance tests (LOAEL: 283 mg/m

). For acute effects a NOAEL

of 150 mg/m

has been identified (Andersen et al. 1983) and will be taken forward to the risk characterisation. Dividing

by an assessment factor of 10 for interindividual variation, the limit of exposure results in 15 mg/m

(see also Table

6.1), applied as a hourly average.

The LOAEL for long-term effects on the CNS from occupational studies is approximately 30 mg/m

after adjustment

from intermittent to continuous exposure. Dividing by an assessment factor of 100 (10 for interindividual variation, 10

for use of a LOAEL rather than a NOAEL) a limit of exposure of 0.3 mg/m

is obtained (to be applied as a weekly

average; see also Table 6.1), which should also be protective for reproductive effects (spontaneous abortions).

Limited data in humans indicate an increased risk for late spontaneous abortions at dose levels around 332 mg/m

Human data as well as studies in rats and limited data in mice provide evidence of similar developmental effects, i.e.

lower birth weight, delayed postnatal development and developmental neurotoxicity. Only very high exposure levels

were investigated in humans. In animals, the NOAEL for lower birth weight and delayed postnatal development is 2.3

g/m

. A NOAEL for developmental neurotoxicity cannot be determined from the available studies. The LOAEL for this

effect is 4.5 g/m

There has been no indication that toluene is carcinogenic in bioassays conducted to date and the weight of available

evidence indicates that it is not genotoxic.

Table 6.1: Derivation of toluene short- and long-term limits of exposure (EL)

Effect level - mg/m

Assessment

factor

mg/m

Toxicological endpoint

Short-term

Exposure Limit

Human

NOAEL

Volunteers

150 10

Impaired reaction time, headache,

dizziness, feeling of intoxication, slight eye

and nose irritation

Long-term

Exposure Limit

Human

LOAEL

Occupation

30 100

0.3 CNS: color vision impairment

not considering a NOAEL (10);

intraspecies variability (10)

The INDEX project Final report

172

Relevance for the EU-population exposure

Table 6.2: Percentage of population exposed beyond the derived EL and margins of safety

Available exposure data

EL derived

Margins of

Safety (MOS)

Description (Study, Year) N 0.3 mg/m

(90

)

Athens 30-h TWA (Expolis, 96-98) 42 5% 7 (1)

Basel 30-h TWA (Expolis, 96-98) 47 < 19 (9)

Helsinki 30-h TWA (Expolis, 96-98) 188 < 23 (7)

Milan 30-h TWA (Expolis, 96-98) 38 < 5 (2)

Oxford 30-h TWA (Expolis, 98-00) 40 < 26 (6)

Prague 30-h TWA (Expolis, 96-98) 46 < 5 (2)

England 28d-TWA (BRE, 97-99) 796 < 20 (6)

German Survey 48-h PEM (GerEs II,

1990-92)

113 5 % 4 (1)

Germany 4w-TWA (GerES IV pilot

study, spring+summer 2001)

44 < 27 (9)

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

110 < 20 (5)

Avg. MOS: 16 (5)

< out of the evaluation range (i.e. <5% of the environments investigated)

Result

Human effects on the central nervous system are considered as the most sensitive effects in both short- and long-term

inhalatory exposure to toluene. Available exposure data indicate that the European population is not experiencing health

effects of concern resulting from the exposure to toluene in their homes. Results from ten monitoring surveys show that

toluene levels in the order of the established exposure limit of 300 µg/m

could be reached under worse-case conditions

and in a limited number of urban residences. On average, median concentrations (90th percentile) were found to be 16

(5) times lower than the EL. Also, short-term exposures associated with human indoor activities are not expected to

exceed the acute EL set here at 15.000 µg/m

The INDEX project Final report

173

Xylenes (ortho-, meta- and para-)

Synonyms: o-xylene (1,2-dimethylbenzene or 2-xylene),

m-xylene (1,3-dimethylbenzene or 3-xylene),

p-xylene (1,4-dimethylbenzene or 4-xylene, also noted as

methyltoluene, benzene-dimethyl, dimethylbenzene)

CAS Registry Numbers: 95-47-6 o-xylene,

108-38-3 m-xylene,

106-42-3 p-xylene

Molecular Formula: C

1. Compound identification

Xylene is an aromatic hydrocarbon, which exists, in three isomeric forms: meta (m-), para (p-) and ortho (o-). Technical

grade xylene contains a mixture of the three isomers. Xylenes are widely used in the chemical industry as solvents for

products such as paints, inks, dyes, adhesives, pharmaceuticals, and detergents (HSDB, 2003). Approximately 92% of

mixed xylenes is blended into gasoline as antiknock agents. It is also used in a variety of solvent applications,

particularly in the paint and printing ink industries. Xylene is a colourless liquid at room temperature with an aromatic

odour. All three isomers evaporate easily to the air from water. In soil and water, the meta and para isomers are

biodegradable, but the ortho isomer is more persistent. p-Xylene is produced in the highest quantities in the U.S. for use

in manufacture of plastics and polymer fibers including mylar and dacron (WHO 1997).

Acute (short-term) inhalation exposure to mixed xylenes in humans results in irritation of the eyes, nose, and throat,

gastrointestinal effects, eye irritation, and neurological effects. Chronic (long-term) inhalation exposure results

primarily in central nervous system (CNS) effects, such as headache, dizziness, fatigue, tremors, and incoordination;

respiratory, cardiovascular, and kidney effects have also been reported. EPA has classified mixed xylenes as a Group D,

not classifiable as to human carcinogenicity. (EPA 2003)

2. Physical and Chemical properties

o-Xylene m-Xylene p-Xylene

Molecular weight (g/mol) 106.16 106.16 106.16

Melting point (°C; 101.3 kPa) -25.2 -47.9 13.3

Boiling point (°C; 101.3 kPa) 144.4 139.1 138.3

Relative density (25°/4°C) 0.876 0.860 0.857

Solubility in water (mg/litre) 142 146 185

Conversion factors at 20 °C and 760 mm Hg:

1 ppm = 4.406 mg/m

1 mg/m

= 0.227 ppm

Sources: WHO 1997, Verschueren 2001, EPA/Cal. 2003.

3. Indoor Air Exposure assessment

Emission sources

The products that contain and may emit xylenes to indoor air are such as perfumes, pesticide formulations,

pharmaceuticals, adhesives, paints, printed materials, rubber, plastics, leather, polyester fibres, film and fabricated items

(IARC, 1989, ECETOC, 1986; Fishbein, 1988).

The INDEX project Final report

174

Outdoor sources of xylenes are facilities that produce or use xylenes, and exhaust gases from motor vehicles and

through volatilisation from their use as solvents (EPA/Cal. 2003).

Indoor air and exposure concentrations

Mean indoor concentrations of m&p-Xylenes were more than twofold in Northern Europe (Helsinki), almost threefold

in Central Europe (Basel, Praque) and almost twofold in Southern Europe (Milan), compared to respective ambient

concentrations (Annex 3). Xylenes were the second most abundant compounds in indoor air in the EXPOLIS study,

accounting for 18-27% of the target aromatics (Saarela et al 2003). The mean indoor concentrations tend to increase

from north to south being lowest in Helsinki, 7.8 µg/m

, and the highest in Milan, 37 µg/m

. Lowest mean exposures to

m&p-Xylenes were found in Helsinki, 25 µg/m

and the highest in Germany and Athens, 51 µg/m

and 55 µg/m

respectively. Personal exposures were clearly higher than indoor concentrations. The highest personal/indoor ratios

were found in Basel and Helsinki, being over 5 and 3, respectively. Also in all other cities the personal/indoor ratio

exceeded 1, suggesting the presence of important personal sources for m&p-xylenes.

Similar patterns were found for o-Xylene including the highest concentrations in exposures and also higher indoor

concentrations compared to ambient levels. In general, concentrations of o-Xylene were lower than those of m&p-

Xylenes.

TEAM studies carried out in the USA showed maximum 12-hour day time exposures to m&p-Xylenes as high as 1800

µg/m

and 460 µg/m

in New Jersey in 1981 (Wallace et al 1985) and in Los Angeles in 1987 (Wallace et al 1991),

respectively. Simultaneous ambient air concentrations were typically 10 to 100 times lower. In California, also

residential indoor air concentrations were measured showing a maximum 12-hour concentration of 170 µg/m

. Personal

exposure concentration in California showed similar values compared to European distributions.

Maximum 12-hour exposure, 160 µg/m

, measured in the TEAM study in California (Wallace et al 1991) agreed also

with European results (Figure ). Instead maximum residential indoor concentration, 68 µg/m

, was clearly higher than in

European homes. In New Jersey, maximum daytime 12-hour exposure was as high as 830 µg/m

(Wallace et al 1985).

The cumulative distributions of the indoor and 48-hour personal exposure concentrations of m&p- xylenes are presented

in Figure 3.1 and Figure 3.2, and source related short time concentrations are presented in Figure 3.3. m&p -Xylene

concentrations caused by emissions from incense burning were studied by Lee and Wang (2004). Usually,

concentrations caused by incense burning in an 18-m

chamber ranged from 0.5 µg/m

to 3.8 µg/m

, but one type of

incense caused a concentration of 21 µg/m

. Brown (2002) studied short time (30-50 min) concentrations of m&p -

Xylene in established and new buildings, reporting mean concentrations of 6.9 µg/m

and 30 µg/m

respectively. m&p -

Xylene concentration in a new building decreased to 25 µg/m

and 2.8 µg/m

in 72 and 246 days respectively. Expected

sources of m&p -Xylene in new buildings were building materials. Elke at al (1998) measured short time concentrations

of m&p -Xylene inside buildings (60-min average), inside a train and a car (30-min average). The highest

concentrations were found in a car, the second highest in a train and the lowest in a smoker day room.

Source related short time concentrations of o-Xylene are presented in Figure 3.4, and indoor air and personal exposure

distributions in Figure 3.5 and Figure 3.6. o-Xylene concentrations caused by incense burning were typically less than 5

µg/m

, but one type of incense caused a concentration of 96 µg/m

. Short time (30-50 min) concentrations of o-Xylene

in established and new buildings were 8.9 µg/m

and 32 µg/m

respectively. Concentration in a car was about the same

level with a new building, but lower in a train and in a smoker day room. Due to the similar health outcomes, combined

(m-, p-, and o-) indoor air Xylene concentration distributions are presented in Figure 3.7 to be used in the risk

characterisation.

The INDEX project Final report

175

100

0 50 100 150 200 250

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of m&p – xylenes in Athens (Ath), Basel

(Bas), Helsinki (Hel), Milan (Mil) Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002).

100

0 50 100 150 200 250

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.2. Cumulative frequency distributions of 48-hour personal exposure concentrations of m&p – xylenes in

Athens (Ath), Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean

exposures of the German Survey GerES II (GeS) (Hoffman et al 2000).

The INDEX project Final report

176

Incence

burning

New buildings In a car Room with

ETS

Train with

ETS

Concentration (µg/m

)

m,p-Xylene

Figure 1. Short time m&p -Xylene concentrations related to specific microenvironments or emission sources.

100

120

Incence

burning

New

buildings

In a car Room with

ETS

Train with

ETS

Concentration (µg/m

)

o-Xylene

Figure 2. Short time o-Xylene concentrations related to specific microenvironments or emission sources.

The INDEX project Final report

177

100

02040

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.5. Cumulative frequency distributions of indoor air concentrations of o –Xylenes in Athens (Ath, n=42), Basel

(Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=38) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS 2002).

100

0 10203040506070

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.6. Cumulative frequency distributions of 48-hour personal exposure concentrations of o –xylenes in Athens

(Ath), Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean exposures of the

German Survey GerES II (GeS) (Hoffman et al 2000).

The INDEX project Final report

178

100

0 50 100 150

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.7. Cumulative frequency distributions of indoor air concentrations of m,p and o –Xylenes in Athens (Ath,

n=42), Basel (Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=38) Oxford (Oxf, n=40) and Prague (Pra, n=46)

(EXPOLIS 2002).

The INDEX project Final report

179

4. Toxicokinetics

Absorption

The blood/air partition coefficient for the three isomers range from 26.4 to 37.6 (Sato and Nakajima, 1979), indicating

that xylenes entering the body would be readily absorbed into the blood.

Sedivec and Flek (1976a) made direct measurements of the level of absorption of xylenes by subjecting volunteers to

200 or 400 mg/m

of either individual isomers or mixtures of the three isomers of xylenes vapor for 8 hours without

interruption and measuring the difference in the concentration of xylenes in the inspired air relative to the amount

expired. The amount of the individual isomers absorbed over time was consistent for all three isomers and ranged from

62.4 to 64.2% of the inhaled volume, reflecting a high solubility of xylenes in blood.

Riihimäki and Savolainen (1980) conducted studies on human subjects both at rest and during exercise to measure the

kinetics resulting from exposure to mixed xylenes. Healthy male subjects were exposed to xylene for 5 days, 6 hours

per day with a 1-hour break at midday and then for an additional 1 to 3 days after a 2-day weekend break. The exposure

scenarios included either constant exposure to 434 or 868 mg/m

(100 or 200 ppm) or fluctuating exposure with peaks

of 868 or 1736 mg/m

(200 or 400 ppm) that lasted for 10 minutes. The subjects were either sedentary or exercising on

a stationary bike for short periods of time. Regardless of the exposure scenario (constant or fluctuating or with different

xylene concentrations), retention consistently remained around 60% (i.e., 60% of the inhaled xylene was retained in the

blood and 40% was expired). The results indicate that partitioning of xylene between the tissues and the air occurs, but

it is limited by the solubility in the tissue lipids and the rate of passive diffusion through the matrix. Overall, the lowest

uptake rate was noted with 434 mg/m

exposure during sedentary conditions (22 µmol/min) and the highest uptake was

seen with the fluctuating concentrations in which the peaks reached 1.7 mg/m

during exercise (266 µmol/min). Given

the constant retention values, the two factors that appeared to control the total uptake of xylenes were the ambient

concentration of xylene and ventilation rates of the subjects.

Further studies on xylene uptake by inhalation have been conducted by Ogata et al., 1970; Astrand et al., 1978; Senczuk

and Orlowski, 1978; David et al., 1979.

Distribution

The Kow (octanol/water distribution coefficient) of xylene indicates that xylene is expected to partition primarily into

tissues containing a higher proportion of neutral lipids, such as adipose, liver, and brain tissue.

In their study on the uptake and distribution of ethylbenzene and xylenes, Riihimäki and Savolainen (1980) found that

10–20% of the xylene was distributed to the adipose tissue. Adipose has the highest concentration of neutral fat and the

highest affinity for xylene of all tissues. Therefore, once sequestered in adipose tissue, xylene is expected to have the

lowest rate of metabolism, the slowest movement to blood, and the longest persistence in the body. The concentration of

xylene in gluteal subcutaneous fat was about 10-fold higher than in venous blood following the last day of exposure (5

days exposure + weekend without exposure + 1 day of exposure).

Additional information on the distribution of xylenes in the body is available from Astrand et al. (1978), Engstrom and

Bjurstrom (1978) and Kumarathasan et al. (1998).

Metabolism and elimination

Proposed metabolic pathways for o-xylene are shown in Figure 4.1 as a model for all xylene isomers. The principal

metabolic fate involves oxidation of one of the methyl groups to a methylbenzoic acid derivative via methylbenzyl

alcohol and methylbenzaldehyde intermediates. The methylbenzoic acid derivative is mostly conjugated to glycine,

producing methylhippuric acid derivatives that can be excreted in urine. Conjugation to glucuronic acid is a minor

pathway. Oxidation of the benzene ring to produce xylenols (i.e., dimethylphenols) is expected to be a negligible

metabolic pathway, based on analysis of urinary metabolites. The liver is expected to be the principal site of metabolism

for xylenes.

The INDEX project Final report

180

* o-xylene used as a model for all isomers of xylene

** significant production of glucuronic derivative under conditions of high levels of administration

Figure 4.1. Metabolic pathways for xylenes (Source: Adapted from Ogata et al., 1970; Riihimaki and

Savolainen, 1980; Riihimaki, 1979; Bray et al., 1949; Sedivec and Flek, 1976a,b; Ogata et al., 1980;

Carlsson, 1981; Senszuk and Orlowski, 1978; David et al., 1979)

Riihimäki and Savolainen (1980), in their study of inhalation exposure to xylene by human subjects under sedentary

and physically active conditions, found that 95% of the eliminated xylene was in the form of methylhippuric acid, with

the remainder lost as unmetabolized xylene in expired air. No deposition sites, such as lipid-rich tissues, were studied.

The excretion rate of xylene from the blood followed biphasic, first-order kinetics, with the initial loss of xylene having

a half-life in the venous blood of 0.5–1 hour, followed by a second phase with a half-life of 20–30 hours. The authors

proposed that the two phases representing the rapid loss of xylene from the blood, mostly through conversion to

methylhippuric acid followed by excretion, indicate that well-perfused organs reach equilibrium within minutes and

muscles reach equilibrium within a few hours, whereas adipose tissues may require several days of continuous exposure

to reach equilibrium.

Riihimäki (1979) evaluated the metabolism and excretion of xylene and toluene derivatives in humans. A volunteer was

administered a single dose of 7.4 mmole m-methylbenzoic acid or 7.8 mmole m-methylhippuric acid. Urine was

analyzed for 30 hours following administration for the presence of metabolites. All of the administered xylene

derivatives appeared in urine as methylhippuric acid, indicating that, under the conditions of this study, once xylene has

been oxidized to methylbenzoic acid, the only route of metabolism was as the glycine conjugate.

Following methylbenzoic acid administration, only methylhippuric acid was detected in urine. The rate of loss (i.e,

excretion rate) was greater with the methylhippuric acid treatment than with the methylbenzoic acid treatment. This

study is limited by the fact that it was conducted on a single individual. The determination that loss of xylene is limited

The INDEX project Final report

181

by the availability of glycine suggests that the rate of utilization of glycine may vary with such factors as age and

nutritional status of the individual.

5. Health effects

Differences among individual xylene isomers

Although differences in the toxicity of individual xylene isomers have been detected, no consistent pattern following

inhalation exposure has been identified.

In rats exposed by inhalation for 30 minutes, EC

s (the concentrations producing half-maximal decreases in response

rate) for effects on an operant behavior test showed a relative toxicity order of o-xylene > p-xylene > m-xylene, whereas

s for a motor performance test showed a toxicity order of p-xylene > o-xylene = m-xylene. The range of EC

values among the isomers was not considered large (Moser et al., 1985). In contrast, p-xylene - but not o-xylene or m-

xylene - altered audiometric variables in rats exposed to 7800 mg/m

, 6 hours per day, 5 days per week, for 13 weeks

(Gagnaire et al., 2001). A different order of toxicity has been described for effects on motor coordination (rotarod

performance) in rats following 6 hours exposure to concentrations of 13 g/m

xylenes: o-xylene > m-xylene > p-xylene

(Korsak et al., 1990). All three isomers caused decreased fetal body weights in rats exposed for 24 hours per day on

GDs 7–14, but o-xylene caused the effect at a lower concentration (1.5 g/m

) than did either p-xylene or m-xylene (3

g/m

) (Ungváry et al., 1980). No available studies compare the potency of isomers for affecting neurological endpoints

following subchronic or chronic inhalation exposure. Other animal studies that have examined the toxicity of individual

xylene isomers are: Condie et al. (1988) ; Molnár et al. (1986); Fang et al. (1996).

Humans are most likely exposed to a mixture of xylenes rather than to individual isomers. In commercial mixtures the

m-isomer usually predominates (44–70% of the mixture) (Fishbein, 1988; ATSDR, 1995), the exact composition of the

isomers depending on the source. In technical product ethylbenzene is commonly present at significant levels. Thus,

most of the environmental and occupational exposures and toxicological studies are conducted on this mixture of

xylenes containing ethylbenzene. Other minor contaminants of xylenes include toluene and C

aromatic fractions.

Effects of short-term exposure

Despite its structural similarity to benzene, xylene does not influence hematopoiesis. Acute controlled exposure studies

have identified self-reported symptoms of irritation (e.g., watering eyes and sore throat) or neurological impairment

(e.g., mild nausea, headache, altered reaction time, altered balance) as potential effects of xylene following inhalation

exposure in humans.

Levels of 434-868 mg/m

(100-200 ppm) are associated with nausea and headache; 868-2170 mg/m

(200-500 ppm)

with dizziness, irritability, weakness, vomiting, and slowed reaction time; 3480-43000 mg/m

(800-10000 ppm) with

lack of muscle coordination, giddiness, confusion, ringing in the ears, and changes in sense of balance; and 43000

mg/m

(>10000 ppm) with loss of consciousness (HESIS, 1986). Other documented neurological effects include

impaired short term memory, impaired reaction time, performance decrements in numerical ability, and impaired

equilibrium (dizziness) and balance (Carpenter et al., 1975; Dudek et al., 1990; Gamberale et al., 1978; Riihimaki and

Savolainen, 1980; Savolainen and Linnavuo, 1979; Savolainen and Riihimaki 1981; Savolainen et al., 1979b; 1984;

1985b).

Results from subchronic animal studies identify neurological impairment and possible developmental effects as

potential health hazards from repeated inhalation exposure. Scattered reports of body weight changes and adaptive liver

changes in animals are available, but the results do not consistently identify these effects as potential health hazards.

Nelson et al. (1943) exposed 10 healthy human volunteers for periods of 3 to 5 minutes to estimated concentrations of

434 or 869 mg/m

(100 or 200 ppm) technical grade xylene. The subjects reported eye, nose, and throat irritation at 869

mg/m

but not at 434 mg/m

. A significant area of uncertainty arising from the Nelson et al. (1943) study is the use of

estimated rather than measured exposure concentrations. Carpenter et al. (1975) evaluated eye irritation in 7 human

volunteers exposed for 15 minutes to 460, 1000, 2000, or 3000 mg/m

. One volunteer noted mild throat discomfort at

460 mg/m

, but not at 2000 mg/m

. No subjects reported eye irritation at 460 mg/m

(106 ppm).

Hastings et al. (1984) exposed 50 healthy individuals to 434, 869, or 1738 mg/m

(100, 200, or 400 ppm) mixed xylenes

for 30 minutes to evaluate eye, nose, and throat irritation. The percent of subjects reporting eye irritation was 56 for

The INDEX project Final report

182

controls (clean air), 60 at 434 mg/m

, 70 at 869 mg/m

, and 90 at 1738 mg/m

. The authors concluded there was no

effect on eye irritation at 434 mg/m

because the incidence of irritation was as low as the control group. The data from

Nelson et al (1943), Carpenter et al. (1975), and Hastings et al. (1984) taken together are consistent with a human

NOAEL for eye irritation of about 434 mg/m

(100 ppm) for at least a 30-minute exposure.

Exposure of sedentary or exercising subjects to a 10-minute peak concentration of 1736 mg/m³ (400 ppm) resulted in

significantly increased uncontrolled body sway in these subjects. Exposure to 868 mg/m³ (200 ppm) xylene for up to 5

hours did not result in CNS disturbances measured by increased body sway (Laine et al., 1993). Riihimaki and

Savolainen (1980) reported that a single 5-minute exposure to 1736 mg/m³ xylene (isomeric form unknown) resulted in

lightheadedness and inebriation similar to alcohol intoxication. Deleterious effects on EEG, reaction time, body

balance, and manual dexterity were found in 8 healthy volunteers following exposure to 434 mg/m³ (100 ppm) m-

xylene for 6 hours/day for 6 days (Savolainen et al., 1980a).

Exposure of 15 volunteers to 434 mg/m³ technical xylene mixed with 20% ethylbenzene for 70 minutes, including 30

minutes of exercise, resulted in significant impairments in short-term memory and other CNS performance tests

(Gamberale et al., 1978). Because ethylbenzene may have contributed to the CNS effects, definitive conclusions about

the effects of xylene cannot be drawn from this study.

Nine healthy male volunteers were exposed to 868 mg/m³ m-xylene 4 hours a day, with or without exercise for 10

minutes at the beginning of each session (Savolainen et al., 1985a). There were no changes in reaction times, but

average and maximal body sway were decreased in a concentration dependent manner. Exercise had a sway reducing

effect. Male volunteers were exposed to 868 mg/m³ m-xylene vapor for 4 hours a day, either sedentary or with 10

minutes periods of exercise twice a day (Savolainen et al., 1984). The body balance of the subjects was impaired in the

anteroposterior direction. Nine healthy male students were exposed to 868 mg/m³ m-xylene for 4 hours per day at 6-day

intervals over 6 consecutive weeks (Savolainen et al., 1982a). Body sway tended to decrease with exposure. Only minor

electroencephalographic effects were noted on 4 hour exposures to 868 mg/m³ m-xylene exposure, and no other adverse

effects were noted (Seppalainen et al., 1991).

Five volunteers were exposed to 174 mg/m³ (40 ppm) xylene for 7 hours/day, 3 consecutive days/week in an inhalation

chamber. There was an 11-day break between each 3-day session (Mergler and Beauvais, 1992). Individual differences

in olfactory perception thresholds for toluene were noted, but there was no effect of exposure duration.

Summary of short-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

430

Study

215

1h-ADJ

860

Study

430

1h-ADJ

Eye, nose and throat irritation (subjective reports

of)

Volunteers, 30min Hastings et al., 1984

(supp.from other

studies)

OEHHA (1999);

REL: 22

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Acute Reference Exposure Level

Study Hastings et al., 1984 (with support from Carpenter et al., 1975;

Nelson et al., 1943)

Study population 50 healthy human volunteers

Exposure method 30 minute exposures to 430, 860 or 1720 mg/m

xylene (technical

grade)

Critical effects subjective reports of eye, nose, and throat irritation

LOAEL 860 mg/m

NOAEL 430 mg/m

(100 ppm)

Exposure duration 30 minutes

Equivalent 1 hour concentration 50 ppm (C1 * 60 min = 100 ppm * 30 min)

LOAEL uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

The INDEX project Final report

183

Cumulative uncertainty factor 10

Reference Exposure Level 5 ppm (22 mg/m³, 22,000 µg/m³)

With the possible exception of inconsistently observed developmental endpoints, irritation is the lowest reported human

health effect for xylene.

Effects of long-term exposure

Information on the toxicity of xylenes to humans is almost exclusively limited to case reports of acute exposures and

studies of occupational exposures in which persons often inhaled a mixture of hydrocarbon solvents 8 hours per day, 5-

6 days per week. These studies often have incomplete information on the airborne concentrations of xylene and other

hydrocarbons. One study examining chronic effects in humans from inhalation of predominantly mixed xylenes was

identified (Uchida et al., 1993) and one 4-week controlled exposure study examining the effects of p-xylene exclusively

was identified (Hake et al., 1981).

Xylene exposure has been associated with effects in a number of organ systems including the lungs, skin and eyes;

neurological system; heart and gastrointestinal system; kidney; and possibly the reproductive system.

Pulmonary effects have been documented in occupational exposures to undetermined concentrations of mixed xylenes

(and other solvents) and include labored breathing and impaired pulmonary function (Hipolito 1980; Roberts et al.,

1988). High levels of xylene exposure for short periods are associated with irritation of the skin, eyes, nose and throat

(ATSDR, 1995). Chronic exposure to xylenes has been associated with eye and nasal irritation (Uchida et al., 1993).

Although the mechanism by which xylenes exert their toxic effects on the nervous system and developing fetus are not

completely understood, some theories exist. The CNS toxicity observed during exposure to high concentrations has

been attributed to the liposolubility of xylenes in the neuronal membrane (Desi et al., 1967; Savolainen and Pfaffli,

1980; Tahti, 1992). It has been suggested that xylenes disturb the actions of proteins essential to normal neuronal

function. Changes in levels of various neurotransmitters and lipid composition have been observed in several brain

areas following acute- and intermediate-duration exposure to xylenes (Andersson et al., 1981; Honma et al., 1983). It is

unclear whether these represent direct effects of xylenes or are secondary changes resulting from nonspecific CNS

depression. Some authors have also suggested that metabolic intermediates such as arene oxides or methylbenzaldehyde

may be responsible for the toxic effects of xylenes (Savolainen and Pfaffli, 1980).

Chronic exposure to xylenes (with other hydrocarbons) has been associated with cardiovascular and gastrointestinal

effects. Heart palpitations, chest pain, and abnormal electrocardiogram were noted (Hipolito, 1980; Kilburn et al., 1985)

as were effects on the gastrointestinal system producing nausea, vomiting and gastric discomfort in exposed workers

(Goldie, 1960; Hipolito, 1980; Uchida et al., 1993; Klaucke et al., 1982; Nersesian et al., 1985).

Results of studies of renal effects of xylene are mixed and come from case reports and occupational studies where

multiple chemical exposures are common. The effects from subchronic exposure documented by Hake et al. (1981) and

from chronic exposure documented by Uchida et al. (1993) did not include renal effects. However, Morley et al. (1970)

found increased BUN and decreased creatinine clearance; Martinez et al. (1989) found distal renal tubular acidemia;

Franchini et al. (1983) found increased levels of urinary β-glucuronidase; and Askergren (1981, 1982) found increased

urinary excretion of albumin, erythrocytes, and leukocytes.

Available data (NTP, 1986; Wolfe et al., 1988a, b; Condie et al., 1988) do not consistently identify the kidney and the

liver as a sensitive target of xylenes in animals.

Reproductive effects were documented by Taskinen et al. (1994) who found increased incidence of spontaneous

abortions in 37 pathology and histology workers exposed to xylene and formaldehyde in the work place. The multiple

chemical exposures and the small number of subjects in this study limit the conclusions that can be drawn as to

reproductive effects of xylene in humans.

No hematological effects have been identified in studies where exposure was to xylene only. Previous studies

identifying hematological effects included known or suspected exposure to benzene (ATSDR, 1995; ECETOC, 1986).

One series of case reports identified lowered white cell counts in two women with chronic occupational exposure to

xylene (Hipolito, 1980; Moszczynsky and Lisiewicz, 1983; 1984), although they may also have had multiple chemical

exposures.

Groups of male volunteers (1 to 4 subjects/group) were exposed to p-xylene in a controlled-environment chamber for

7.5, 3, or 1 hr/day, 5 days/week for 4-weeks (Hake et al., 1981). The p-xylene concentration was changed on a weekly

The INDEX project Final report

184

basis starting at 434 mg/m

(100 ppm) the first week, followed by 87 mg/m

, 652 mg/m

, and 434 mg/m

(20, 150, and

100 ppm on average, with a range of 50 to 150 ppm) over subsequent weeks. In addition, groups of female volunteers (2

or 3/group) were exposed to 434 mg/m

p-xylene for 7.5, 3, or 1 hr/day for 5 days. The volunteers acted as their own

controls, with exposure to 0 ppm p-xylene occurring for two days (males) or one day (females) the week before and the

week after the xylene exposures. No serious subjective or objective health responses, including neurological tests,

cognitive tests and cardiopulmonary function tests were observed. Odor was noted, but the intensity decreased usually

within the first hour of exposure. The authors concluded that p-xylene may have a weak irritating effect on the soft

tissues starting at 434 mg/m

, but overall, the small sample size and high variability among the volunteers made all

results difficult to interpret.

The Uchida et al. (1993) study included a relatively large number of workers studied, exposure for an average of 7

years to xylenes predominately and a comprehensive set of medical examinations to document potential effects. A

survey of 994 Chinese workers involved in the production of rubber boots, plastic coated wire and printing processes

employing xylene solvents was carried out. The survey consisted of fitting individual workers with diffusive samplers

for an 8 hour shift. At the end of the 8 hour shift the samplers were recovered for analysis of solvent exposure, and urine

samples were collected for analysis of xylene metabolites. The following day workers answered a questionnaire

concerning subjective symptoms, and blood and urine were collected for analysis. Out of this group of xylene-exposed

workers, 175 individuals (107 men and 68 women) were selected for further study and analysis based on completion of

their health examinations and on results from diffusive samplers showing that xylene constituted 70% or more of that

individual’s exposure to solvents in the workplace. The control population consisted of 241 (116 men and 125 women)

unexposed workers from the same factories or other factories in the same region, of similar age distribution, of similar

time in this occupation (average of 7 years), and having a similar distribution of alcohol consumption and cigarette

usage. The xylene-exposed and unexposed groups were given health examinations which evaluated hematology (red,

white, and platelet cell counts, and hemoglobin concentration), serum biochemistry (albumin concentration, total

bilirubin concentration, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, alkaline

phosphatase, leucine aminopeptidase, lactate dehydrogenase, amylase, blood urea nitrogen, creatinine), and subjective

symptoms (survey of symptoms occurring during work and in the previous three months).

Results of analysis of the diffusive samplers showed that workers were exposed to a geometric mean of 61.7 ± 11.3

mg/m

(14.2 ± 2.6 ppm) xylene (arithmetic mean of 92.5 ± 93.8 mg/m

). This was broken down into geometric means

of 5.2 mg/m

o-xylene, 31.7 mg/m

m-xylene, 16.5 mg/m

p-xylene, 14.8 mg/m

ethyl benzene, and 4.5 mg/m

toluene.

n-Hexane was rarely present and no benzene was detected. Analysis of data from the health examinations found no

statistically significant difference (p<0.10) between hematology and serum biochemistry values for xylene-exposed and

unexposed populations. The frequency of an elevated ratio of aspartate aminotransferase to alanine transferase and of

elevated ratio of alkaline phosphatase to leucine aminopeptidase was significantly (p<0.01) higher in exposed men than

in the control population of men.

Results of the survey of subjective symptoms found differences in symptoms occurring during work and during a

similar analysis over the preceding three month period, apparently related to effects on the central nervous system and

to local effects on the eyes, nose and throat. The frequency of five symptoms experienced during work was significantly

(p<0.01) elevated in either xylene-exposed men or women including: dimmed vision, unusual taste, dizziness, heavy

feeling in the head, and headache. The frequency of four symptoms experienced during work were significantly

(p<0.01) elevated in both men and women including irritation in the eyes, nasal irritation, sore throat, and floating

sensation. Ten subjective symptoms occurring in the previous three months were significantly (p<0.01) elevated in

exposed men and women including nausea, nightmare, anxiety, forgetfulness, inability to concentrate, fainting after

suddenly standing up, poor appetite, reduced grasping power, reduced muscle power in the extremities, and rough skin.

Dose dependency appeared to exist for 3 subjective symptoms noted during work: irritation in the eyes, sore throat,

floating sensation, and for one symptom occurring in the last three months, poor appetite.

Although the human evidence for persistent effects on the nervous system or other persistent effects from repeated

inhalation exposure to xylenes is inadequate, results from animal studies more clearly identify potential persistent

neurological impairment and possible developmental effects as potential health hazards from repeated inhalation

exposure.

A number of subchronic studies in animals provide evidence for neurological effects following repeated inhalation

exposure to xylenes (Korsak et al., 1992; Korsak et al., 1994; Gralewicz et al. 1995; Gralewicz and Wiaderna 2001;

Pryor et al. 1987; Nylén and Hagman 1994).

The lowest exposure level that produced changes in a number of neurological endpoints was identified in several studies

of rats exposed to 434 mg/m

(100 ppm) m-xylene, 6 hours per day, 5 days per week, for 3 months. These studies

observed statistically significant changes in neurobehavioral tests conducted at least 24 hours following cessation of

The INDEX project Final report

185

exposure: decreased rotarod performance indicative of impaired motor coordination (Korsak et al., 1992, 1994),

decreased spontaneous motor activity (Korsak et al., 1992), impaired radial maze performance indicative of a learning

deficit (Gralewicz et al., 1995), and decreased latency to paw lick in the hot plate test, indicating increased sensitivity to

pain (Korsak et al., 1994).

At a lower exposure level (217 mg/m

by the same exposure protocol), rats showed statistically significantly decreased

latency in the paw lick response but no statistically significant effects on rotarod performance (Korsak et al., 1994).

Rats exposed to 434 mg/m

m-xylene by the same daily protocol for a shorter duration (4 weeks) displayed no

statistically significant changes in tests of radial maze performance, open-field activity, or active avoidance, but paw

lick latency was increased in the hot plate test and step-down time was shortened in 1/6 trials in the passive avoidance

test (Gralewicz and Wiaderna, 2001).

There are no available studies on the possible developmental toxicity of inhaled xylenes in humans, but a number of

studies have examined standard developmental toxicity endpoints and neurobehavioral endpoints in offspring of

animals exposed to mixed xylenes or individual xylene isomers.

The most significant effects on developmental endpoints were decreased fetal body weight or fetal survival in rats at

xylene isomer doses of 1520 or 3041 mg/m

(350 or 700 ppm) 24 hours per day (Ungváry et al., 1980) or at mixed

xylenes concentration of 3388 mg/m

(780 ppm) 24 hours per day (Ungváry and Tátrai, 1985) and increased abortions

in rabbits exposed to 1000 mg/m

(230 ppm) 24 hours per day (Ungváry and Tátrai, 1985). These effects, although of

concern, occurred at concentrations above those at which neurobehavioral effects were reported in adult animals.

Hass and Jakobsen (1993) exposed groups of 36 pregnant Wistar rats to clean air or 868 mg/m

(200 ppm) of xylene for

6 h/day on days 4-20 of gestation. There was no sign of maternal toxicity and no decrease in fetal weights and no

increase in soft-tissue or skeletal malformations. A large increase in the incidence of delayed ossification of the os

maxillare of the skull, however, was observed (53% of experimental fetuses as opposed to 2% of the controls). Potential

neurological/muscular changes measured as performance on a rotorod were also noted upon testing of 2-day-old rat

pups.

Developmental and reproductive effects

There are no available studies on the possible developmental toxicity of inhaled xylenes in humans, but a number of

studies have examined standard developmental toxicity endpoints and neurobehavioral endpoints in offspring of

animals exposed to mixed xylenes or individual xylene isomers.

The most significant effects on developmental endpoints were decreased fetal body weight or fetal survival in rats at

xylene isomer doses of 1500 or 3000 mg/m

(350 or 700 ppm) 24 hours per day (Ungváry et al., 1980) or at mixed

xylenes concentration of 3400 mg/m

(780 ppm) 24 hours per day (Ungváry and Tátrai, 1985) and increased abortions

in rabbits exposed to 1000 mg/m

24 hours per day (Ungváry and Tátrai, 1985). These effects, although of concern,

occurred at concentrations above those at which neurobehavioral effects were reported in adult animals.

Gestational exposure of animals to xylenes has resulted in neurodevelopmental effects (Hass et al., 1995; 1997; Hass

and Jakobsen, 1993) and other possible developmental effects (Ungváry et al., 1980; Ungváry and Tátrai, 1985), but

only at levels above those at which neurobehavioral effects in adult male rats were reported.

Exposure of pregnant rats for 6 hours/day on days 4-20 of gestation to 870 mg/m³ (200 ppm) technical (mixed) xylene

resulted in significantly increased incidence of delayed ossification of the skull in the offspring (Hass and Jakobsen,

1993). The rat pups exposed prenatally to 870 mg/m³ xylene displayed significantly decreased motor performance

during adolescence. However, a study using p-xylene showed no significant embryotoxic or developmental effects on

the CNS as measured by acoustic startle response in rats following exposure to 7000 mg/m³ throughout gestation

(Rosen et al., 1986).

The INDEX project Final report

186

Summary of long-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

61.7

Study

22.1

ADJ

Eye irritation, sore throat, floating sensation,

poor appetite

Occupational, 7y Uchida et al. 1993 OEHHA (1999);

REL: 0.7

ATSDR (1995);

MRL: 0.6

217

EXP

HEC

434

EXP

HEC

CNS; impaired motor coordination

Rats, 3m Korsak et al., 1994 EPA-IRIS (2003)

; RfC: 0.1

868

EXP

Developmental

Rats Hass and Jakobsen,

1993

RIVM (1999)

250

EXP

177

HEC

Developmental Rats Ungváry and Tátrai,

1985

Health Canada

(1991); TC: 0.18

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Chronic Reference Exposure Level (REL):

Study Uchida et al. (1993)

Study population 175 xylene-exposed factory workers and control population of 241

factory workers

Exposure method Inhalation

Critical Effects Dose related increase in the prevalence of eye irritation, sore throat,

floating sensation, and poor appetite.

LOAEL 14.2 ppm (geometric mean of exposure concentrations)

NOAEL Not applicable

Exposure continuity 8 hr/d (10 m3/day occupational inhalation rate), 5 d/wk

Exposure duration Occupational exposure for an average of 7 years

Average occupational exposure 5.1 ppm for LOAEL group (14.2 x 10/20 x 5/7)

Human equivalent concentration 5.1 ppm for LOAEL group

LOAEL uncertainty factor 3

Subchronic uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 30

Inhalation reference exposure level 0.2 ppm (200 ppb; 0.7 mg/m3; 700 µg/m3) for mixed xylenes or for

total of individual isomers

A number of issues are important in considering the uncertainty associated with this REL. For ATSDR (1995), the

animal and human toxicity data suggest that mixed xylenes and the different xylene isomers produce similar effects,

although different isomers are not equal in potency for producing a given effect. Therefore exposure of workers to a

mix of xylenes in the Uchida et al. (1993) study would be expected to generate a similar spectrum of responses as

exposure to single isomers, however the intensity of particular effects could be different.

The use of a neurological endpoint for derivation of a REL is supported by the large number of inhalation and oral

studies, which associate neurological effects with xylene exposure. ATSDR (1995) indicates that neurological effects

are a sensitive endpoint. The observation that floating sensation is apparently related to dose, further supports the

concept that this subjective symptom related to neurological effects was due to xylene exposure.

A UF of 3, rather than 10, was applied for the LOAEL to NOAEL extrapolation due to the generally mild adverse

effects observed and the principally low incidence (<50%) of the effects. A factor of 1 was used for subchronic

uncertainty. Although the average occupational exposure was only 7 years, there were 176 xylene-exposed workers of

average age 29.7 ± 9.0 years (arithmetic mean ± SD) for whom, according to the report, there had been essentially no

change in workplace in their working life. Thus, many workers would likely have been exposed for more than 8.4 years,

the cut-off point for chronic human exposure. Another issue is the use of diffusive samplers in the Uchida et al. (1993)

study. These samplers provide a time weighted average concentration of hydrocarbon and cannot indicate the maximum

concentrations a worker is exposed to. It is unknown whether peak concentrations alter the response to xylenes in

humans.

The INDEX project Final report

187

For comparison with the proposed REL of 0.7 mg/m

(200 ppb) based on human studies, (1) the free-standing NOAEL

of 339 mg/m

(78 ppm) o-xylene obtained by Jenkins et al. (1970) in rats and guinea pigs continuously exposed for 90

days was used to estimate a REL based on animal data. Use of an RGDR of 1, a subchronic UF of 3, an interspecies UF

of 3, and an intraspecies UF of 10 results in a REL of 3.5 mg/m

(800 ppb) for o-xylene for systemic effects. (2) Tatrai

et al. (1981) found a free standing LOAEL of 4761 mg/m

(1096 ppm) o-xylene for body weight gain in male rats

exposed every day for 8 hours. Time adjustment to continuous exposure and use of an RGDR of 1, a LOAEL UF of 3

for a mild effect, an interspecies UF of 3, and an intraspecies UF of 10 result in a REL of 17.4 mg/m

(4000 ppb). (3)

Ungváry and Tátrai (1985) exposed mice by inhalation continuously to 521 or 1000 mg/m

(120 or 230 ppm) xylene for

24 h/day on days 7-15 of gestation. The LOAEL was 1000 mg/m

and the NOAEL was 521 mg/m

. No time adjustment

is needed. Use of an RGDR of 1, a subchronic UF of 1, an interspecies UF of 3, and an intraspecies UF of 10 results in

a REL of 17.4 mg/m

(4000 ppb) for xylene for developmental effects.

ATSDR (2003)

Agency for Toxic Substances and Disease Registry

Derivation of the Minimal Risk Level (MRL):

Using the most well defined Uchida et al. (1993) study, ATSDR estimated a chronic inhalation MRL of 0.6 mg/m

from

a LOAEL of 61 mg/m

in humans for less serious effects; divided by a total uncertainty factor of 100 (10 for use of a

LOAEL and 10 for human variability).

U.S.EPA - IRIS (2003)

Integrated Risk Information System

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC)

Critical effect Experimental doses* UF MF RfC

Impaired motor coordination

(decreased

rotarod performance)

Subchronic inhalation study in

male rats (Korsak et al., 1994)

NOAEL: 50 ppm

NOAEL

(HEC)

: 39 mg/m

LOAEL: 100 ppm

LOAEL

(HEC)

: 78 mg/m

300 1 0.1 mg/m

*Conversion Factors and Assumptions - MW = 106.17. Assuming 25°C and 760 mmHg, NOAEL(mg/m

) = 50 ppm x 106.17/24.45 = 217 mg/m

NOAEL

[ADJ]

= 217 mg/m

x 6 hrs/day x 5 days/7 days = 39 mg/m

. The NOAEL*

[HEC]

was calculated for extrarespiratory effects of a Category 3 gas

(U.S. EPA, 1994). Blood/gas partition coefficients: H(b/g)

rat

= 46.0; H(b/g)

human

=26.4 (Tardif et al., 1995). (H

b/g

)

rat

/(H

b/g

)

human

=1.7; value of 1 is used

when the ratio is >1 (U.S. EPA, 1994). NOAEL*

[HEC]

= NOAEL

[ADJ]

x (H

b/g

)

rat

/(H

b/g

)

human

= 39 mg/m

The subchronic rat study by Korsak et al. (1994) was selected as the principal study for derivation of the RfC. A

NOAEL of 217 mg/m

(50 ppm) and a LOAEL of 434 mg/m

(100 ppm) were identified for decreased rotarod

performance (impaired motor coordination). This neurologic test was administered 24 hours after termination of the

exposure period, when xylenes would be expected to have been eliminated from the body. Other subchronic rat studies

provide support for the finding that 434 mg/m

exposure produces statistically significant changes in a number of

neurological endpoints: decreased rotarod performance (Korsak et al., 1992), decreased spontaneous motor activity

(Korsak et al., 1992), and impaired radial maze performance indicative of a learning deficit (Gralewicz et al., 1995).

Additional support for 434 mg/m

as an exposure level that may produce mild neurologic deficits comes from the report

(Gralewicz and Wiaderna, 2001) that rats exposed to 434 mg/m

m-xylene for 4 weeks showed shortened step-down

time in 1/6 trials in the passive avoidance test 50 days postexposure. These studies collectively identify 434 mg/m

the lowest reliable subchronic animal LOAEL and 217 mg/m

as a NOAEL for deficits in neurologic endpoints.

The NOAEL of 217 mg/m

was duration adjusted to 39 mg/m

, and a NOAEL[HEC] of 39 mg/m

was calculated on the

basis of species differences in blood/gas partition coefficients. The NOAEL[HEC] of 39 mg/m

was divided by a total UF

of 300 (10

1/2

x 10 x 10

1/2

x 10

1/2

; 3 for animal-to-human extrapolation using default dosimetric adjustments, 10 for

intrahuman variability, 3 for extrapolation from subchronic to chronic duration, and 3 for deficiencies in the data base,

including lack of studies in two species [available studies are predominantly in male rats] and a two-generation

reproductive toxicity study) to give the RfC of 0.1 mg/m

. Alternative approaches using available PBPK models to

extrapolate rat exposure concentrations to HECs arrived at similar points of departure for the RfC as the NOAEL[HEC] of

39 mg/m

The INDEX project Final report

188

Health Canada (1991)

Determination of a critical effect:

In inhalation studies conducted to date, xylenes have not been teratogenic but have induced fetotoxic effects, sometimes

at doses below those that were toxic to the mothers. With the exception of unconfirmed results reported by Mirkova et

al. (1983), the lowest concentration of xylenes reported to induce fetotoxic effects in the absence of maternal toxicity in

available investigations was 500 mg/m

. At this concentration, in an incompletely documented study, moderate

embryotoxic effects such as retardation of skeletal development and of body weight increase were observed in the

offspring of rabbits exposed continuously on days 7 to 20 of gestation to xylenes of unspecified composition (Ungváry

and Tátrai, 1985). In rats, maternal toxicity and fetotoxicity were observed following exposure to 250 mg/m

on days 7

to 15 of pregnancy, the lowest concentration administered, indicating that rats may be a more sensitive species

(Ungváry and Tátrai, 1985).

Derivation of a tolerable concentration (TC):

Inhalation is considered to be the most important route of exposure to xylenes for the general public. A provisional

tolerable concentration (TC) for xylenes has been derived on the basis of the results of studies in animal species

exposed by inhalation. Based on reported effect levels, developmental toxicity was considered the critical endpoint. The

lowest concentration at which fetotoxic effects in the absence of maternal toxicity have been observed following

exposure by inhalation to xylenes was 500 mg/m

in rabbits in a limited study (Ungváry and Tátrai, 1985); maternal and

fetal toxicity were observed in rats at the lowest administered concentration (250 mg/m

A provisional TC of 0.18 mg/m

has been derived, therefore, on the basis of the lowest LOEL (250 mg/m

) for

meaningful effects reported in a bioassay of adequate quality (though documentation was incomplete) in the most

sensitive species (Ungváry and Tátrai, 1985). This was adjusted for the ratio of inhalation volume to body weight of rats

[(0.11 m

/day)/0.35 kg] to humans aged 5 to 11 years [(12 m

/day)/27 kg]. An uncertainty factor of 1000 was applied

[10 for intraspecies variation, 10 for interspecies variation, 10 for LOEL rather than NOEL (although observed effects

at the LOEL were only moderately fetotoxic; documentation was also limited)]. No additional factor was incorporated

for the limited period of exposure since fetotoxic effects occur at doses below those which induce adverse effects in

subchronic and chronic studies.

The lowest concentration of the individual xylene isomers reported to induce adverse effects in studies in animal species

following inhalation is 150 mg/m

, in which embryo- and fetotoxic effects were observed in the absence of maternal

toxicity following continuous exposure of pregnant rats to p-xylene on days 7 to 14 of gestation (Ungváry et al., 1980).

This is only slightly less than the LOEL used above in the derivation of a TC for xylenes (177 mg/m

Carcinogenic potential and genotoxicity

Associations between occupational exposure to xylenes and increased risk of leukemia (Arp et al., 1983; Wilcosky et

al., 1984), non-Hodgkin's lymphoma (Wilcosky et al., 1984), and cancer of the rectum (Gérin et al., 1998), colon (Gérin

et al., 1998), or nervous system (Spirtas et al., 1991) have been reported. However, a number of limitations preclude the

usefulness of these data, including small sample sizes, no quantified exposure concentrations, and/or concurrent

exposures to other solvents.

Adequate human data on the carcinogenicity of xylenes are not available and the available animal data are inconclusive

as to the ability of xylenes to cause a carcinogenic response. Evaluations of the genotoxic effects of xylenes have

consistently given negative results.

The genotoxicity of commercial xylenes and all three individual isomers has been studied and the results are, for the

most part, negative (IARC, 1989). All studies cited in the GENE-TOX database are negative with the exception of one

study for which no conclusion was drawn. Xylenes are not mutagenic in bacterial test systems with Salmonella

typhimurium (Bos et al., 1981; Florin et al., 1980; NTP, 1986) and Escherichia coli (McCarroll et al., 1981) or in

cultured mouse lymphoma cells (Litton Bionetics, 1978). Xylenes do not induce chromosomal aberrations or sister

chromatid exchanges in Chinese hamster ovary cells (Anderson et al., 1990) or cultured human lymphocytes (Gerner-

Smidt and Friedrich, 1978), chromosomal aberrations in rat bone marrow (Litton Bionetics, 1978), micronuclei in

mouse bone marrow (Mohtashamipur et al., 1985), or sperm head abnormalities in rats (Washington et al., 1983).

Technical grade xylenes, but not o- and m-xylene, are weakly mutagenic in Drosophila recessive lethal tests (Donner et

al., 1980). No increase in the frequency of sister chromatid exchanges was observed in peripheral lymphocytes in

individuals exposed to xylenes in an occupational setting (Haglund et al., 1980; Pap and Varga, 1987) or an

experimental setting (Richer et al., 1993).

The INDEX project Final report

189

Interactions with other chemicals

The interaction of xylene with alcohol, drugs (aspirin, phenobarbitol), and various solvents (1,l,ltrichlorethane, benzene,

toluene, ethylbenzene, methyl ethyl ketone) has been evaluated in experimental studies with humans and animals.

Xylene has a high potential to interact with numerous substances because the isomers induce microsomal enzymes in

the liver (Blanchard and Morris 1994; Liira et al. 1991), while microsomal enzymes in the lungs are inhibited by xylene

exposure (Blanchard and Morris 1994; Elovaara et al. 1987; Pate1 et al. 1978; Silverman and Schatz 1991; Toftgard and

Nilsen 1982). Which enzymes will be affected is isomer dependent. For example, m-xylene is a more potent inducer of

P-450 2B enzymes than p-xylene (Backes et al. 1993). The isomer differences, as well as organ differences in effects on

xenobiotic metabolizing enzymes, make it difficult to predict the interaction of xylene with other substances.

Acute inhalation exposure to a mixture of toluene and xylene resulted in more than additive respiratory and central

nervous system toxicity (Korsak et al. 1988, 1992). Elevated blood levels of xylene and toluene and decreased excretion

of the major metabolites of xylene and toluene in the urine (Tardif et al. 1992) suggest mutual metabolic inhibition.

However, simultaneous exposures in humans indicate that a threshold exists for this interaction (Tardif et al. 1991). No

increase in blood levels of these substances was observed during combined exposures to 50 ppm toluene and 40 ppm

xylene over 3 consecutive days, whereas increases in blood levels and levels in exhaled air were observed during a

combined 4-hour exposure to 95 ppm toluene and 80 ppm xylene. Thus, combined exposures at below threshold level

are unlikely to produce greater than additive toxicity (Tardif et al. 1991). A physiologically based toxicokinetic

modeling study using rat data suggests that the interaction between toluene and xylene is competitive, with toluene a

more potent inhibitor of xylene metabolism than xylene is of toluene metabolism (Tardif et al. 1993a, 1993b).

Exposure to xylene combined with benzene or ethylbenzene may also produce mutual inhibition of the metabolism of

both solvents (Engstrom et al. 1984; Nakajima and Sato 1979b). Ethylbenzene is found in commercial xylene. In

contrast, ethyl acetate exposure in combination with exposure to

m-xylene caused a reduction in blood xylene levels

(Freundt et al. 1989).

Possibly because of competition for the enzymes involved in conjugation with glycine during the concurrent

metabolism of

m-xylene and aspirin by human volunteers, saturation of the conjugation pathway occurred that led to

decreases in the metabolism of both aspirin and

m-xylene (Campbell et al. 1988). Administration of aspirin to pregnant

rats during inhalation exposure to xylene caused greater than additive potentiation of maternal and fetal toxic effects

(Ungváry 1985). This was postulated to be due to the interference with metabolism of aspirin by xylene and vice versa.

Inhalation of

m-xylene following pretreatment with phenobarbital was associated with both increased pulmonary

retention of

m-xylene and increased urinary excretion of m-methylbenzoic acid (David et al. 1979). Surprisingly,

inhalation of

m-xylene and l,l,l-trichloroethane has been associated with slight improvements in certain

psychophysiological parameters, including reaction time and equilibrium in humans as compared with pre-exposure

measurements (Savolainen et al. 1982a, 1982b), and impairment in others such as visual evoked potentials and

equilibrium (Savolainen et al. 1982a; Seppalainen et al. 1983).

Ethanol appears to inhibit the metabolism of xylene, resulting in elevated blood levels of xylene and decreased

excretion of methylhippuric acid (Elovaara et al. 1980; Riihimaki et al. 1982a, 1982b; Romer et al. 1986; Savolainen

1980; Savolainen et al. 1978, 1979b, 1980b). A kinetic study in rats (Kaneko et al. 1993) suggests that ethanol

inhibition of xylene metabolism occurs only at high concentrations (500 ppm). Paradoxically, ethanol pretreatment

causes additive effects with xylene in inducing microsomal enzymes in the liver (Wisniewska-Knypl et al. 1989). This

would enhance the metabolic capacity of the liver and modify biological effects of other chemicals that are either

detoxified or converted to toxic metabolites by the microsomal enzymes. In summary, it cannot be stated with certainty

whether alcohol and xylene would interact to produce synergistic or antagonistic effects in humans and animals because

there are reasons why both would occur.

Odour perception

Source: ATSDR (1995)

Odor thresholds in air

Mixture: 4.3 mg/m

(1.0 ppm)

m-isomer: 16 mg/m

(3.7 ppm)

o-isomer: 0.34-0.74 mg/m

(0.08–0.17 ppm)

p-isomer: 2.0 mg/m

(0.47 ppm)

The INDEX project Final report

190

Source: Carpenter et al. (1975)

Mixed xylenes: 4.3 mg/m

(1 ppm) (detection)

Source: AIHA (1989)

Mixture:: 87 mg/m

(20 ppm) (detection);: 174 mg/m

(40 ppm) (recognition)

m-Xylene: 1.5 – 4.8 mg/m

(0.35 - 1.1 ppm) (detection)

o-Xylene: 23 mg/m

(5.4 ppm) (detection)

p-Xylene: 9.1 mg/m

(2.1 ppm) (detection)

Summary of Xylenes Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: not relevant

Toxicokinetics: ~60% uptake of the inhaled amount (no difference among isomers); accumulates in adipose tissue of

chronically exposed

Health effect levels of short- and long-term exposure

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

430

Study

215

1h-ADJ

860

Study

430

1h-ADJ

Eye, nose and throat irritation (subjective reports

of)

Volunteers, 30min Hastings et al., 1984

(supp.from other

studies)

OEHHA (1999);

REL: 22

Long-term exposure

61.7

Study

22.1

ADJ

Eye irritation, sore throat, floating sensation,

poor appetite

Occupational, 7y Uchida et al. 1993 OEHHA (1999);

REL: 0.7

ATSDR (1995);

MRL: 0.6

217

EXP

HEC

434

EXP

HEC

CNS; impaired motor coordination

Rats, 3m Korsak et al., 1994 EPA-IRIS (2003)

; RfC: 0.1

868

EXP

Developmental

Rats Hass and Jakobsen,

1993

RIVM (1999)

250

EXP

177

HEC

Developmental Rats Ungváry and Tátrai,

1985

Health Canada

(1991); TC: 0.18

Carcinogenicity: IARC: 3; inadequate evidence in both humans and experimental animals

Genotoxicity: Studies consistently negative

Odour threshold: Mixture of isomers: 4.3 mg/m³ (ATSDR, Carpenter); 0.25 - 0.3 mg/m³ (Devos)

Susceptible population: No evidence

Remarks: RD

: 6.1 g/m

; Although differences in the toxicity of individual xylene isomers have been detected, no

consistent pattern following inhalation exposure has been identified; nevertheless, human exposure most likely occurs

to the mixture of isomers

The INDEX project Final report

191

6. Risk Characterization

Health hazard evaluation of short- and long-term exposure

Animal and human toxicity data suggest that mixed xylenes and the different xylene isomers produce similar effects,

although different isomers are not equal in potency for producing a given effect. Therefore exposure to a mix of xylenes

in occupational studies would be expected to generate a similar spectrum of responses as exposure to single isomers,

however the intensity of particular effects could be different.

The use of a neurological endpoint for the derivation of an exposure limit EL is supported by the large number of

inhalation and oral studies, which associate neurological effects with xylene exposure. The observation that floating

sensation is apparently related to dose, further supports the concept that this subjective symptom related to neurological

effects was due to xylene exposure.

Using a well defined occupational study (Uchida et al. (1993), a chronic inhalation EL of 0.2 mg/m

was derived from a

LOAEL of 62 mg/m

in humans for generally mild adverse effects observed and principally low incidence (<50%). This

LOAEL was adjusted to 22 mg/m

accounting for intermittent to continuous exposure and divided by a total assessment

factor of 100 (10 for use of a LOAEL and 10 for human variability). The endpoint considered was: dose related increase

in the prevalence of eye irritation, sore throat, floating sensation, and poor appetite.

An acute EL was derived refering on a study on volunteers and subjective reports of eye, nose, and throat irritation as

shown in Table 6.1.

Table 6.1: Derivation of short- and long-term exposure limits (EL)

Effect level - mg/m

Assessment

factor

mg/m

Toxicological endpoint

Short-term

Exposure

Limit

Human NOAEL

Volunteers

200 10

Eye, nose and throat irritation (subjective

reports of)

Long-term

Exposure

Limit

Human LOAEL

Occupational

22 100

0.2

Dose related increase in the prevalence of

eye irritation, sore throat, floating

sensation, and poor appetite

not considering a NOAEL (10);

intraspecies variability (10)

Relevance of EU-population exposure to xylenes

Table 6.2: Percentage of population exposed beyond the derived EL and margins of safety

Available exposure data

Xylenes EL

Margins of

Safety (MOS)

Description (Study, Year) N 0.2 mg/m

(90

)

Athens 30-h TWA (Expolis, 96-98) 42 < 7 (4)

Basel 30-h TWA (Expolis, 96-98) 47 < 25 (10)

Helsinki 30-h TWA (Expolis, 96-98) 188 < 25 (10)

Milan 30-h TWA (Expolis, 96-98) 38 < 6 (2)

Oxford 30-h TWA (Expolis, 98-00) 40 < 32 (7)

Prague 30-h TWA (Expolis, 96-98) 46 < 10 (5)

German Survey 48-h PEM (GerEs II,

1990-1992)

113 < 10 (3)

The INDEX project Final report

192

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

102 < 20 (5)

Avg. MOS: 18 (6)

< out of the evaluation range (i.e. <5% of the environments investigated);

Result

A chronic exposure limit of 200 µg/m

has been derived based on generally mild adverse effects associated with CNS

and increase in the prevalence of eye irritation and sore throat. The results of eight monitoring surveys (Table 6.2)

indicate that background levels of xylenes in European residences are of no concern to human health since median (90th

percentile) levels are, on average, 20 (6) times lower than the EL established. Acute exposure data indicate that it is

very unlikely that xylenes emissions associated with human indoor activities would generate levels in the order of the

proposed short-term EL of 20 mg/m

, considered protective for irritative effects in the general population. Although

human exposure most likely occurs to the mixture of xylene isomers, animal and human toxicity data suggest that

mixed xylenes and the different xylene isomers produce similar effects.

The INDEX project Final report

193

7. Aromatic compounds: Comparison of relevant data

Short-term exposure : derivation of limits of exposure (EL)

Effect level - mg/m

Assessment

factor

mg/m

Toxicological endpoint

Toluene

Human NOAEL

Volunteers

150 10

Impaired reaction time, headache,

dizziness, feeling of intoxication, slight eye

and nose irritation

Styrene

Human NOAEL

Volunteers

217 100

2 Eye and throat irritation

Xylenes

Human NOAEL

Volunteers

200 10

Eye, nose and throat irritation (subjective

reports of)

not considering a NOAEL (10);

intraspecies variability (10) ;

susceptible population (10)

Long-term exposure : derivation of limits of exposure (EL)

Effect level - mg/m

Assessment

factor

mg/m

Toxicological endpoint

Toluene

Human LOAEL

Occupational

30 100

0.3 CNS: color vision impairment

Styrene

Human LOAEL

Occupational

25 100

0.25

CNS; neuropsychological tests: reaction

time, s/l term logic memory, s/l term verbal

memory, digit-symbol association, block

design and figure identification.

Xylenes

Human LOAEL

Occupational

22 100

0.2

Dose related increase in the prevalence of

eye irritation, sore throat, floating

sensation, and poor appetite

not considering a NOAEL (10);

intraspecies variability (10) ;

susceptible population (10)

Percentage of population exposed beyond given threshold limits

EXPOSURE DATA

Toluene EL Styrene EL Xylenes EL

Aromatic HC

Mixture

Description (Study, Year) N 0.3 mg/m

0.25 mg/m

0.2 mg/m

refer to text*

Athens 30-h TWA (Expolis, 96-98) 42 5 % < < 20%

Basel 30-h TWA (Expolis, 96-98) 47 < < < <

Helsinki 30-h TWA (Expolis, 96-98) 188 < < < <

Milan 30-h TWA (Expolis, 96-98) 38 < < < 35%

Oxford 30-h TWA (Expolis, 98-00) 40 < < < 5%

Prague 30-h TWA (Expolis, 96-98) 46 < < < 15%

England 28-d TWA (BRE, 2002) 796 < n.a. n.a.

German Survey 48-h PEM (GerEs II,

1992-93)

113 5 % < <

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

102 -

110

< < <

< out of the evaluation range (i.e. <5% of the environments investigated);

* see reciprocal calculation procedure

The reciprocal calculation approach for mixtures

A further element is included, based on a ACGIH approach commonly used in occupational settings to describe

Threshold Limit Values (TLVs) resulting from the exposure to solvent mixtures. Supposing the additivity of substances

having similar toxicologic effects, an exposure limit for aromatic hydrocarbons could be defined as follows :

[Benzene]/EL

Benz

+ [Toluene]/EL

Tol

+ [Styrene]/EL

Styr

+ [Xylenes]/EL

Xyls

≤ 1

with EL

= single Exposure Limits considering neurologic endpoints

The INDEX project Final report

194

100

012345

Neurological Endpoint:

Animal LOAEL

Benz

= 2.5 mg/m

TC = 0.025 mg/m

Human LOAEL

Tol

= 31 mg/m

GL = 0.3 mg/m

Human LOAEL

Styr

= 25 mg/m

GL = 0.25 mg/m

Animal NOAEL

Xyls

= 39 mg/m

TC = 0.4 mg/m

TC = tolerable concentration

GL = indoor guideline value proposed

Athens

Basel

Helsinki

Milan

Oxford

Prague

concentration relative to proposed guidelines

Percentile

In order to consider overall neurologic endpoints, the following critical effects and limits were chosen for benzene and

the xylenes (toluene and styrene were still classified as neurologic toxicants in the present RA).

Benzene: increased rapid response time, Animal LOAEL: 2.5 mg/m

, Assessment factor: 100, EL: 0.025

Xylenes: impaired motor coordination, Animal NOAEL: 39 mg/m

, Assessment factor: 100, EL: 0.4

Figure 6.1: Cumulative frequency distribution of EXPOLIS data, applying the

reciprocal calculation approach on a mixture of four aromatic hydrocarbons

The INDEX project Final report

195

Styrene

Synonyms: phenethylene, phenylethene, phenylethylene, styrol, styrole, styrolene,

vinylbenzene, vinylbenzol, cinnamene, cinnamol

CAS Registry Numbers: 100-42-5

Molecular Formula: C

1. Compound identification

Styrene is a colourless, viscous liquid with a pungent odour. It is one of the most important monomers, worldwide

because of its common use in the production of polystyrene, acrylonitrile-butadiene-styrene resins, styrene-butadiene

rubbers and latexes, and in the reinforced plastics industry. Styrene has been produced since the 1920s by catalytic

dehydrogenation of ethylbenzene.

2. Physical and Chemical properties

Molecular weight (g/mol) 104.15

Melting point (°C) -30.6

Boiling point (°C) 145.2

Density (at 20 °C, 1 atm) 0.91

Relative density (air =1) 3.6

Solubility: slightly soluble in water, soluble in ethanol

and very soluble in benzene and petroleum ether

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 4.323 µg/m

1 µg/m

= 0.231 ppb

Sources: (WHO 1983, WHO 2000, Verschueren 2001, EPA/Cal 2003)

3. Indoor Air Exposure assessment

Emission sources

Styrene exposures, at levels of milligrams per day, have been measured in the manufacture of styrene-based plastics and

rubbers and fibre glass-reinforced polyester products and at much higher levels in the glass fibre-reinforced plastics

industry. General population is typically exposed to levels of micrograms per day resulted from inhalation of ambient

air and cigarette smoke, emissions from newly installed carpets containing a styrene-butadiene rubber latex adhesive,

and from intake of food that has been in contact with styrene-containing polymers (WHO, 1983, IARC 1994, IARC

2000, EPA/Cal 2003).

Natural styrene sources are such as cinnamic acid containing plants, e.g. balsamic trees, and a by-production of fungal

and microbial metabolism. Ambient air concentration of styrene is relatively low compared with toluene and xylene,

because of its reactivity with ozone to yield irritants such as benzaldehyde and peroxides such as peroxybenzyol nitrate,

which is a potent eye irritant. (WHO 2000).

The INDEX project Final report

196

Indoor air and exposure concentrations

Usually indoor styrene concentrations were low, but higher than respective outdoor concentrations ranging 1-6 µg/m

and 1-2 µg/m

, respectively (Table 4.0.1). Arithmetic mean in Milan was an exception, being as high as 30.3 µg/m

due

to few extremely high data points. Indoor concentrations were at the same level or higher in Helsinki and Prague, but

less than one half in Athens compared to personal exposures. Probably occupational emissions caused this difference in

Athens, because concentrations in workplaces were clearly higher, 7.1 µg/m

, than residential indoor concentrations

(Annex 3, Jantunen et al 1999).

In Northern America styrene concentrations in outdoor air have been generally <1 µg/m

. In an Canadian study, carried

out in 1988–1990, the mean concentrations at the ambient 18 sites ranged from 0.09 to 2.35 µg/m

. In indoor air the

mean concentrations were slightly higher (<1–6 µg/m

). Smoking has been found a significant contributor to indoor

concentrations of styrene. The styrene content of cigarette smoke has been reported to be 18–48 µg/cigarette (WHO

1983, WHO 2003). Nazaroff and Singer (2002) assessed average daily residential exposures to styrene in ETS for US

non-smokers who live with smokers. Using exposure–relevant emission factors and assuming the mean consumption of

20 cigarettes smoked per day, they assessed a concentration of 0.6 µg/m

being caused by ETS.

Lee and Wang (2004) studied styrene emissions from incense burning in the chamber (18 m

) tests. They found

maximum styrene concentrations peaking up to 13 µg/m

during incense burning.

Cumulative distributions of the indoor concentrations of styrene in some European cities are presented in Figure 3.1 and

distributions of 48-hour personal exposures in Figure 3.2.

100

0 5 10 15 20

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of 30-hour indoor air concentrations of styrene in Athens (Ath, n=42),

Basel (Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=40) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS

2002).

The INDEX project Final report

197

100

01020

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.2. Cumulative frequency distributions of 48-hour personal exposure concentrations of styrene in Athens (Ath),

Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean concentrations of the

German Survey GerES II (Hoffman et al 2000).

4. Toxicokinetics

Absorption

In animal and human controlled studies, the uptake of styrene has been found to be rapid. The principal routes of

exposure are pulmonary and, to a lesser extent, dermal. In several studies with workers and volunteers, the pulmonary

retention of styrene was 60–70% of the inhaled dose. Styrene in ambient air is absorbed through the skin at 2–5 % of

the dose absorbed in the respiratory tract. (IARC, 1994).

Distribution

Styrene is widely distributed throughout the body. The distribution and sequestration of styrene to fat tissue and its

subsequent slow elimination indicate a potential for accumulation in situations of repeated daily exposure. However, in

a study of workers exposed to 160 mg/m

(37 ppm; 8-hour TWA) styrene, no evidence of accumulation was found in

monitoring samples during a working week (Pekari et al., 1993).

Metabolism and elimination

Styrene is oxidized to styrene-7,8-oxide by the cytochrome P-450-mediated monooxygenase system. The responsible

isozymes are the ethanol-inducible CYP2E1, CYP2B6 and CYP1A2, and additional isozymes with lower activity for

styrene may be involved (Nakajima et al., 1994). Styrene can also undergo oxidation by other mechanisms and it can be

cooxidized to styrene-7,8-oxide during the lipoxynase-mediated formation of arachidonic acid peroxides (Belvedere, G.

et al., 1983). In vitro, styrene oxidation to styrene-7,8-oxide has been shown to be mediated by oxyhaemoglobin and

myoglobin (Tursi et al., 1983; Ortiz de Montellano and Catalano, 1985).

Enzymatic hydrolysis of styrene-7,8-oxide yields phenylethylene glycol which is further oxidized to mandelic acid and

phenylglyoxylic acid, the principal urinary metabolites of styrene in humans (IARC, 1994). Conjugation of styrene-7,8-

oxide with glutathione leads to urinary excretion of thioethers or mercapturic acid, quantitatively a minor metabolite in

humans. In a field study carried out on workers exposed to 43–311 mg/m

of styrene, styrene-7,8-oxide concentration in

blood was found to be linearly correlated with ambient styrene (Korn et al. 1994).

The INDEX project Final report

198

The elimination of styrene is linear at lower concentrations of atmospheric styrene. At concentrations of less than 852

mg/m

there is a fast elimination phase with a half-life of 0.5–0.7 hours, and a second phase of slow elimination with a

half-life of 13 hours (IARC, 1994). Several physiologically-based pharmacokinetic models have been developed

(33,34) which simulate the behaviour of styrene and styrene-7,8-oxide in the body. Partially saturated metabolism can

be predicted by the models at high exposures exceeding 1280 mg/m

(300 ppm).

Urinary excretion of mandelic acid and phenylglyoxylic acid is biphasic. After an 8-hour exposure of workers to styrene

at concentrations of 26–130 mg/m

(6.1–30.5 ppm), both metabolites have a half-life of about 2.5 hour for the first

phase and 30 hours for the second (IARC, 1994). These compounds are used in assessing occupational exposure to

styrene. According to several studies, an exposure to 80 mg/m

(20 ppm) styrene corresponds, the following morning, to

a combined mandelic acid and phenylglyoxylic acid level of about 2.9 mmol/litre (Pekari et al., 1993). Coexposure to

ethanol inhibits the metabolism of styrene, resulting in a lag in the excretion of mandelic acid (Wilson et al. 1983).

5. Health effects

Effects of short-term exposure

Styrene may irritate the eyes and mucous membranes and may be toxic to the central nervous system (IARC, 1979).

Immediate eye and throat irritation, increased nasal mucus secretion, listlessness, impairment of balance, and

drowsiness followed by unsteadiness, muscle weakness, and depression were reported in a study of 2 human volunteers

exposed to 3408 mg/m

(800 ppm) styrene for 4 hours (Carpenter et al., 1944). Other symptoms include a feeling of

being “lightheaded” or “drunk” (Lorimer et al., 1976).

In an exposure chamber study, volunteer subjects complained of an objectionably strong odor when exposed to 852-

1704 mg/m

(200-400 ppm) styrene; exposure to 256 mg/m³ (60 ppm) resulted in detectable odor but no irritation (Wolf

et al., 1956). The duration of exposure and number of subjects were not specified. Investigators at a fiberglass plant

could not withstand more than 1-2 minute exposure to concentrations of 2130-3410 mg/m

(500-800 ppm) styrene

(Götell et al., 1972). However, workers exposed to this concentration of styrene for hours complained of only mild to

moderate complaints of irritation, suggesting that tolerance may have developed.

Stewart et al. (1968) found eye and throat irritation in 3 out of 6 volunteers exposed to 422 mg/m

(99 ppm) styrene for

20 minutes. No symptoms were reported in 3 subjects after exposure to 217 mg/m

for 1 hour. Exposure of these

subjects to 1579 mg/m

styrene for 25 minutes resulted in nausea, significant discomfort, and an abnormal Romberg

test, indicative of cerebellar dysfunction. Significant decrements were noted in 3/5 subjects in other tests of

coordination and manual dexterity at 50 minutes. Exposure to 920 mg/m

or less for up to 1-hour did not cause

measurable impairment of coordination and balance.

The neurotoxic effects mediated by styrene consist of slowing in sensory, but not motor, nerve conduction velocity and

CNS depression (Cherry and Gautrin, 1990). Reaction time was significantly impaired in 12 males exposed to 1470

mg/m

(350 ppm) styrene for 30 minutes, whereas no significant impairment was observed at 1065 mg/m

(250 ppm) or

lower (Gamberale and Hultengren, 1974). In this study, no effects on perceptual speed or manual dexterity were

detected. In another study of 12 workers exposed during the workday to 111 mg/m³ (26 ppm), Edling and Ekberg

(1985) measured reaction time before and after work and found no significant differences. Other non-CNS symptoms

were reported in a neuropsychiatric questionnaire completed by the subjects.

The CNS depressant effects of acute exposures to high styrene levels are probably mediated by the direct effect of the

lipophilic, unmetabolized styrene on nerve cell membranes.

Abnormal electroencephalograms were correlated with urinary levels of the styrene metabolite, mandelic acid, of 700

mg/l or higher in workers exposed to styrene (Harkonen et al., 1978).

Consumption of ethanol has been shown to decrease formation of the metabolites mandelic and phenylglyoxylic acid in

human volunteers exposed to 426 mg/m³ (100 ppm) styrene for 8 hours (Cerny et al., 1990). Lowered levels of these

metabolites have been associated with a reduced risk of CNS disturbances in volunteer workers (Cherry and Gautrin,

1990). Co-exposure to inhaled acetone was reported to alter cytochrome-P450 enzymes as measured by altered urinary

steroid metabolites and glucaric acid in workers who consumed moderate amounts of alcohol (Dolara et al., 1983).

However, the clinical significance of the presence of these compounds in the urine is unknown.

Styrene is bioactivated to styrene 7,8-oxide, a reactive metabolite which binds to tissue proteins, acts as a hapten, and

elicits contact allergy in some individuals (Sjoborg et al., 1984). Analyses of styrene oxide adducts bound to human

serum albumin have been used as biomarkers for exposure to styrene (Rappaport et al., 1993). In a study comparing 9

The INDEX project Final report

199

styrene-exposed workers with 24 healthy controls, hematocrit, blood lead levels, and delta-aminolevulinate dehydrase

(ALA-D) levels were measured (Fujita et al., 1987). The workers were exposed to at least 213 mg/m³ (50 ppm) styrene

for 7 days. Styrene oxide was shown to inhibit the formation of ALA-D, an important enzyme in heme biosynthesis, in

these workers. Styrene oxide is also known to bind covalently to DNA in vitro (Hemminki and Hesso, 1984).

Two subjects with occupational asthma due to prior exposure to styrene were exposed to 64 mg/m

(15 ppm) styrene in

a chamber (Moscato et al. 1987). Immediate bronchoconstriction was observed in both subjects while a late rash was

also observed in one of the subjects.

Predisposing Conditions for Styrene Toxicity

Medical: Asthmatics may be more sensitive to adverse pulmonary effects from styrene exposure (Moscato et al., 1987).

Chemical: Ethanol consumption and acetone inhalation may inhibit the metabolism and clearance of styrene (Cerny et

al., 1990; Dolara et al., 1983; Elovaara et al., 1990).

Summary of short-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

217

EXP

422

EXP

Eye and throat irritation

Volunteers,

1h (NOAEL),

20m (LOAEL)

Stewart et al., 1968 OEHHA (1999);

REL: 21

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Acute Reference Exposure Level (protective against mild adverse effects; for a 1-hour exposure)

Study Stewart et al., 1968

Study population three human volunteers

Exposure method inhalation

Critical effects eye and throat irritation

LOAEL 99 ppm (for 20 minutes)

NOAEL 51 ppm

Exposure duration 1 hour

Extrapolated 1 hour concentration 51 ppm

LOAEL uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 10

Reference Exposure Level 5.1 ppm (21 mg/m³; 21,000 µg/m³)

ATSDR

Agency for Toxic Substances and Disease Registry

Derivation of the acute Minimal Risk Level (MRL).

The possibility for brief human exposure to high concentrations of styrene exists in occupational settings, and might

also exist near major spills. Exposure of the general public to episodic high concentrations of styrene at hazardous waste

sites, in the home, or in the general environment is unlikely.

The respiratory tract and central nervous system are the likely target organ systems for inhaled styrene (Alarie 1973;

Carpenter et al. 1944; DeCeaurriz et al. 1983; Kankaanpaa et al. 1980; Murray et al. 1978; Spencer et al. 1942; Stewart

et al. 1968). The data are not considered sufficient to establish an inhalation acute-duration MRL. Furthermore, ATSDR

acknowledges additional uncertainties inherent in the application of the procedures to derive less than lifetime MRLs.

As an example, acute inhalation MRLs may not be protective for health effects that are delayed in development or are

acquired following repeated acute insults, such as hypersensitivity reactions, asthma, or chronic bronchitis. As these

kinds of health effects data become available and methods to assess levels of significant human exposure improve, these

MRLs will be revised. However, the potential carcinogenicity of styrene prevents the design of controlled laboratory

exposures in humans.

The INDEX project Final report

200

Effects of long-term exposure (noncancer)

Chronic exposures to styrene result in central nervous system (CNS) and peripheral nervous system effects, although

the latter are not as pronounced (ATSDR, 1992; Rebert and Hall, 1994; Murata et al., 1991). Irritation or discomfort of

the upper respiratory tract resulting from styrene exposure has not been reported in long-term occupational studies

(Foureman, 1994). However, sensory irritation and neurological impairment does occur in acute human studies at

concentrations above 426 mg/m

(Stewart et al., 1968). The evidence for styrene induced hepatic changes is either

negative or equivocal (ATSDR, 1992). Evidence for nephrotoxicity due to long-term occupational exposure is also

negative or equivocal (ATSDR, 1992; Verplanke and Herber, 1998; Kolstad et al.,1995). Some human studies suggest

that chronic exposure to styrene results in reproductive effects, but the limited data are difficult to interpret because of

the small sample numbers (Brown, 1991; Lindbohm, 1993). Immunologic alterations (e.g., altered phenotypic profiles

among lymphocyte subsets, decreased natural killer cell activity, and decreased chemotaxis) have also been observed,

but the limited data prevent quantitative interpretation (Bergamaschi et al., 1995; Governa et al., 1994).

The CNS depressant effects of acute exposures to high styrene levels are probably mediated by the direct effect of the

lipophilic, unmetabolized styrene on nerve cell membranes. Long-term effects of styrene exposure may result from the

action of one or more metabolites of styrene (Savolainen, 1977; Mutti et al., 1988).

One postulated mechanism for the chronic non-cancer toxicity of styrene is the binding of the highly reactive styrene

oxide to components of nervous tissue. Another postulated mechanism is an alteration in the levels of circulating

catecholamines (e.g., dopamine) due to the binding of PGA to these biogenic amines (Mutti, 1993; Mutti et al., 1984a;

Checkoway, 1994) and the subsequent changes in physiological functions that are under biogenic amine control.

Although long-term exposures to styrene are associated with decrements in physiological functions, the exact

mechanism(s) for these effects have not been clearly established (see reviews by ATSDR, 1992; Mutti, 1993; Rebert

and Hall, 1994).

Kolstad et al. (1995) estimated excess deaths due to four major non-malignant disease groups for 53847 male workers

in the Danish RPI. Low and high styrene exposures were based on companies with less than 50% (low) and those with

50% or more (high) employees involved with reinforced plastics. An internal comparison was made with workers

unexposed to styrene to account for more similar activities and lifestyles. Statistically significant (p < 0.05) excess

deaths due to pancreatitis and degenerative disorders of the myocardium and non-significant excess deaths due to

degenerative diseases of the nervous system were observed. Non-significant excess deaths due to glomerulonephritis

were also observed.

Checkoway et al. (1994) described a cross-sectional study of 59 male boat plant workers exposed to <4,3 to 613 mg/m

mean = 158 mg/m

styrene. Monoamine oxidase B (MAO-B) activity in platelets was measured as an indicator of

catecholamine metabolism. When the styrene exposed workers were divided into quartile exposures, a dose dependent

decrease in MAO-B activity was observed after adjustments were made for age, smoking, alcohol and medication use.

Female workers employed in the reinforced plastics industry (RPI) were studied for levels of substances associated with

neuroendocrine function (Mutti et al., 1984a). Serum prolactin, thyroid stimulating hormone, human growth hormone,

follicle stimulating hormone, and luteinizing hormone were measured in 30 women who were between the 5th and 15th

day of the menstrual cycle. Exposure was based on the nextmorning MA+PGA, and levels of the neuroendocrine

substances were measured in venous blood samples taken the next morning before the start of work. On the basis of a

relationship (not detailed in the report) between urinary metabolites and styrene air concentration, the authors estimated

that the average styrene TWA/8 hr was about 554 mg/m

(130 ppm). Controls consisted of women factory workers

living in the same area as the styrene-exposed women, but not knowingly exposed to styrene. After controlling for age

and exposure time, the increased prolactin and thyroid stimulating hormone levels were correlated with the

concentration of next-morning urinary MA+PGA, although only the increased prolactin levels were statistically

significant.

Numerous occupational studies have noted CNS disturbances in styrene-exposed workers. Decreased manual dexterity,

increased reaction times, and/or abnormal vestibuloocular reflex (ability to track moving objects) were observed by

Götell et al. (1972), Gamberale et al. (1975), Lindstrom et al. (1976), Mackay and Kelman (1986), Flodin et al. (1989),

Moller et al. (1990), and Cherry and Gautrin (1990) for air styrene levels of about 51 mg/m

(12 ppm) to more than 426

mg/m

(100 ppm). However, in each of these studies, there were difficulties in quantifying the effect. The difficulties

included small sample size, unknown exposure duration, lack of concurrent control group, lack of dose-response data,

and either unknown ethanol consumption or lack of adjustment for ethanol consumption. In the Cherry and Gautrin

(1990) investigation, however, the authors determined that accounting for ethanol consumption did not reduce the

correlation between increased reaction time and exposure.

The INDEX project Final report

201

Decrements in other CNS functions were observed among workers in the well-controlled studies of Fallas et al. (1992),

Chia et al. (1994), and Mutti et al. (1984b). Fallas et al. (1992) studied 60 male workers (average age = 29.5 years,

average air styrene = 104 mg/m

). The styrene-exposed population was compared to non-exposed worker controls and

matched for age, intellectual level, and ethnic origin. The results from a standardized test battery showed decrements in

the aiming response and 22/60 styrene exposed workers exhibited increased reaction times compared to 7/60 controls.

Acquired color vision loss (dyschromatopsia) was also observed in the styrene-exposed workers compared to controls.

Chia et al. (1994) also observed decrements in CNS function as defined by altered visual retention, audio-digit

recognition, and digit recognition. However, a dose-response relationship did not exist. These workers also exhibited a

statistically nonsignificant dose-dependent dyschromatopsia.

In the most comprehensive occupational study to date on CNS effects of styrene exposure, Mutti et al.(1984b) assessed

memory and sensory/motor function in a group of 50 male styrene-exposed workers (average exposure = 8.6 years) and

a control group of 50 manual workers. In addition to matching for age, sex, and educational level, a vocabulary test was

included to match for general intelligence. Eligibility criteria included absence of metabolic, neurologic, or psychiatric

disorders, limited ethanol intake, and limited cigarette usage. All subjects were instructed to avoid intake of alcohol and

drugs for two days prior to testing. Styrene exposure was assessed from urinary MA+PGA levels the morning after the

last workday in the week, followed immediately by participation in a battery of 8 neuropsychological tests designed to

measure CNS function. The tests included reaction time, short and long term logic memory, short and long term verbal

memory, digit-symbol association (using a reference code), block design (reproducing a displayed design using colored

blocks), and embedded figures (timed identification of figures in Rey’s table). The mean + 2 SDs of the values found in

the control group was set as the normal range limit for each neuropsychological test. The results were expressed as

continuous and quantal data. Expressed as continuous data, styrene-exposed workers exhibited significantly poorer

performances than controls in all tests, except in the digit-symbol test. Also, urinary metabolite concentration and

duration of exposure were found to be significantly correlated with the scores of several tests. As a subgroup, workers

with metabolite levels of up to 150 mmoles MA+PGA/mole creatinine (mean = 75 mmoles/mole creatinine + 33 [SD],

which is equivalent to a mean styrene concentration of 64 mg/m

) appeared to have no significant effects. The authors

state that this level of urinary metabolites corresponds to a mean daily 8-hour exposure to air styrene of 107 mg/m

(25

ppm). Based on greater urinary excretion of styrene metabolites, significantly poorer performances in four or more

neuropsychological tests were recorded in the other three subgroups (150-299, 300-450, and > 450 mmoles MA +

PGA/mole creatinine).

Mutti et al. (1984b) expressed the quantal data as the fraction of tested subjects who responded abnormally to ≥ 1, ≥ 2,

and ≥ 3 tests (see Table 5.1). Positive dose-response relationships existed between intensity of styrene exposure

(mmoles MA + PGA/mole creatinine) and abnormal scores, whether it was expressed as abnormal responses in at least

one, at least two, or at least three neuropsychological tests. The chi-square test and validity calculations were performed

by constructing 2 x 2 tables selecting different levels of urinary excretion of MA and PGA as a cut-off point. The

highest values for chi-square and predictive validity were found when the cut-off of 150 mmol/mol creatinine was

chosen, suggesting that the quantal isolation of the low dose subgroup from the next subgroup is appropriate.

When the quantal data for the low dose subgroup were analyzed by OEHHA using the Fisher’s Exact Test, a significant

level of abnormal responses were observed for ≥1 (p = 0.005) and ≥3 (p = 0.04) tests. The abnormal responses for ≥2

tests were statistically marginal (p = 0.06). For each of the remaining exposure groups, the p-values were <0.05. Unlike

the assumptions made concerning the continuous data, quantal data results suggest that the low dose subgroup

represents a LOAEL, and that a NOAEL is not available from the data. Mutti et al. (1984b) also expressed the data in a

quantal three-way representation including prevalence (number of respondents for at least one, two or three abnormal

tests), duration (years at work), and intensity (metabolite level). This representation revealed a positive correlation of

neuropsychological deficits with duration as well as intensity.

Table 5.1: Subjects Classified Positive on Neuropsychological Tests as a Function of Styrene Exposure

MA+PGA, mmoles per mole

creatinine

Total

Subjects

Number of Abnormal Tests

≥ 1 ≥ 2 ≥ 3

Controls 50 4/50 2/50 0/50

< 150 (mean = 75 ± 33)

14 6/14 3/14 2/14

150-299 (mean = 216 ± 45) 9 6/9 5/9 3/9

300 - 450 (mean = 367 ± 49) 14 10/14 7/14 5/14

> 450 (mean = 571 ± 108) 13 11/13 8/13 6/13

Data from Table IV in Mutti et al. (1984b).

“Next-morning” styrene urinary metabolites.

The INDEX project Final report

202

The quantal grouping of the number of subjects that performed abnormally in >3 tests based on their styrene urinary metabolite

concentrations, both shown in bold, were used in a benchmark concentration (BMC) analysis for the derivation of the REL (see

Section VI below).

Based on Guillemin et al. (1982), a linear relationship exists for converting the urinary metabolite concentrations to ppm air

styrene levels (4.97 mmoles MA+PGA/mole creatinine is equivalent to 1 ppm styrene). Thus, the mean styrene concentrations

per group are 0, 15, 44, 74, and 115 ppm.

In addition to dyschromatopsia observed by Chia et al. (1994), Gobba and Cavalleri (1993) and Campagna et al. (1995)

also reported this visual dysfunction among styrene workers in the reinforced plastics industry. Workers (n=36) exposed

to an average of 16 ppm styrene exhibited significantly greater dyschromatopsia than controls, matched for age, ethanol

consumption and tobacco smoking (Gobba and Cavalleri, 1993). Among the study population, only 1/36 styrene-

exposed workers (compared to 16/36 controls) performed the test with 100 percent accuracy. When a different group of

styrene-exposed workers was tested, those exposed to > 50 ppm styrene exhibited greater dyschromatopsia than those

exposed to ≤ 50 ppm, and within this group, a subset exhibited a similar decrement after returning from a one month

vacation. In the Campagna et al. (1995) study, the test for dyschromatopsia was given to 81 reinforced plastics industry

workers (79 male and 2 female) exposed to 4.6, 10.1, and 88.8 ppm styrene (first quartile, median, and third quartile,

respectively). No control group was used in this study. Statistical analysis revealed a correlation of color vision loss

with exposure to styrene (defined as next-morning urinary mandelic acid), age, and ethanol consumption.

Exposure to styrene may affect the peripheral nervous system (PNS). In a case report (Behari et al., 1986), a man

working for 5 years with a photostat process that used styrene was diagnosed with peripheral neuropathy. However, in

occupational studies, the relationship between exposure to styrene and PNS effects has been inconsistent (Triebig et al.,

1985; Cherry and Gautrin, 1990). A major difficulty in understanding the potential for this relationship is the lack of

knowledge about the appropriate surrogate for dose that leads to PNS disturbance (Murata et al., 1991). In one study,

however, chronic exposure indices were developed which included work method, years at work, time spent laminating

(source of high exposure), styrene air concentration, and end-of-shift urinary mandelic acid (Matikainen et al. (1993).

Numbness in the extremities increased with the exposure index, although statistically the effect was marginally

insignificant (p < 0.1). The styrene TWA/8 hr was 136 mg/m

for the 100 study subjects.

Female reproductive toxicity has been inconsistently reported among humans (Brown, 1991; Lindbohm, 1993). These

studies are difficult to interpret because of the high background rates of endpoints such as spontaneous abortion and

menstrual disorders in combination with confounding exposures. In those studies that showed no reproductive effects

due to styrene exposure, the power of the studies was low due to the small numbers of women. Hence the evidence for

any adverse effects of exposure to styrene on female reproductive function is inconclusive.

Kishi et al. (1992a, 1992b) reported that Wistar rat offspring exposed to airborne styrene in utero at 256 mg/m

(60

ppm) for 6 hour/day during days 7 to 21 of gestation, had significantly reduced pup weights at day 1, delayed

development of righting reflex and auditory startle reflex, and nonsignificantly decreased levels of serotonin and its

metabolites 5-hydroxyindoleacetic acid (5-HT) in the cerebrum. Exposure to 1270 mg/m

(293 ppm) significantly

reduced pup body weight and the levels of serotonin and 5-HT, and induced alterations in a wider range of behavioural

measures. Thus, the lowest LOEL for neurotoxic effects in animals following inhalation of styrene in an adequately-

conducted study is 260 mg/m

(60 ppm) in Wistar rats exposed to the compound in utero (Kishi et al., 1992a,b).

Male workers employed in the reinforced plastics industry were examined for effects on sperm chromatin structure and

semen quality (Kolstad et al., 1999a) and time to pregnancy (Kolstad et al., 1999b). No indications of an exposure-

response relationship were seen when individual changes in semen quality were related to the postshift urinary mandelic

acid concentrations among 23 exposed workers. A weak increase in sperm DNA-susceptibility to in situ denaturation as

a function of mandelic acid concentration was indicated, but was within the interassay variability. No detrimental effect

of styrene exposure was observed with regard to male fecundity among 188 exposed workers when compared to 353

unexposed workers.

Immune system alterations were reported in a study conducted by Bergamaschi et al. (1995). Reinforced plastics

industry workers (n=32 female/39 male, average age = 32 years, average exposure duration = 7 years) were compared

with non-styrene exposed factory workers and matched for age, sex, tobacco use and ethanol consumption. Air styrene

levels, among the different factories, varied between 43 – 213 mg/m

(10-50 ppm), and individual worker exposure was

measured by urinary metabolites the morning after the last shift (15 hours post-exposure). Among all workers in the

study (median exposure = 68 mg/m

- according to the data of Guillemin et al. (1982)), the proportion of 12/18

lymphocyte subsets and the prevalence of abnormal values of immunologic phenotypes for 11/18 subsets were

statistically different from the controls (p < 0.001 to < 0.05). When the workers were placed into three exposure groups

(0, < 107 mg/m

, and > 107 mg/m

styrene), dose-response relationships were observed for prevalences of abnormal

responses for four lymphocyte subsets and, in the case of two subsets, abnormal responses were observed in the group

exposed to < 107 mg/m

styrene. Natural killer cell activity (a lymphocyte function), measured in a different group of

The INDEX project Final report

203

workers in the same study, was decreased compared to unexposed worker controls. The median exposure, given in

terms of urinary metabolites, was calculated as 89 mg/m

based on the data of Guillemin et al. (1982). The data show

that exposure of these workers to air styrene levels below 213 mg/m

, and probably at levels near 107 mg/m

, resulted

in alterations of the immune system.

Governa et al. (1994) observed reduced chemotactic responses of polymorphonuclear lymphocytes (PMNs) obtained

from 21 styrene-exposed workers. However, the lack of exposure data prevents a quantitative assessment. In the same

study, 0.1 - 0.6 mM styrene inhibited the chemotaxis of isolated healthy PMNs.

Reproductive and teratogenic effects

The frequency of spontaneous abortions among women with definite or assumed exposure to styrene has been

investigated in a number of studies; the majority do not indicate an increased risk in association with occupational

exposure to styrene (IARC, 1994). A study in Canada (McDonald et al., 1988) found an increased risk for spontaneous

abortions (18 observed; SMR, 158; 90% CI, 102–235) among women employed in polystyrene manufacture. The

expected figures were derived from the experience of 47 316 pregnant women who had worked for 30 hours or more

per week at the start of pregnancy. No styrene concentrations were given, and most of these women had mixed

exposures. Two studies in Finland found no increase in the frequency of spontaneous abortion or congenital

malformation among the wives of men exposed occupationally to styrene (Taskinen et al., 1989; Lindbohm et al.,

1991).

The conclusion thus remains that no clear association has been established between occupational exposure of either

mothers or fathers to styrene and the frequency of spontaneous abortions or congenital malformations.

Summary of long-term exposure effect levels (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Study

MLE

7.2

BC05

2.6

ADJ

CNS; neuropsychological tests: reaction time, s/l

term logic memory, s/l term verbal memory,

digit-symbol association, block design and figure

identification.

Occupational, 8.6y Mutti et al. 1984b

OEHHA (1999);

REL: 0.9

95%lCL

ADJ

>94

95%lCL

same as above Occupational, 8.6y;

Mutti et al. 1984b

EPA-IRIS (1992)

; RfC: 1

107

Study

ADJ

same as above Occupational 8.6y; Mutti et al. 1984b ATSDR (1992);

MRL: 0.25;

WHO (2001);

GV: 0.26;

RIVM (2000)

260

EXP

ADJ

HEC

Neurological developmental; body weight,

biochemical parameters in the brain and

behaviour

Rat offspring Kishi et al.,1992a,b Health Canada

(1993);

TC: 0.092

Final statement (UNIMI) : Human LOAEL: 25 mg/m³

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration ;

95%lCL

lower limit of the 95% confidence interval

The INDEX project Final report

204

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of the Chronic Reference Exposure Level (BMC Approach) :

Study Mutti et al. (1984b)

Study populations Human (occupational)

Exposure method Inhalation

Critical effects Central nervous system

LOAEL 15 ppm

NOAEL Not established

BMC

1.7 ppm

Exposure continuity 8 hr/d (10 m

per 20 m

day), 5 d/wk

Exposure duration 8.6 years (average years at work)

Average occupational exposure 0.61 ppm (1.7 x 10/20 x 5/7)

Human equivalent concentration 0.61 ppm

LOAEL uncertainty factor Not needed in the BMC approach

Subchronic uncertainty factor 1 (average exposure 12.3% of lifetime)

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 3

Cumulative uncertainty factor 3

Inhalation reference exposure level 0.2 ppm (200 ppb; 0.9 mg/m

3; 900 µg/m

)

The most relevant chronic noncancer effect due to styrene exposure is neurotoxicity. The Mutti et al. (1984b)

occupational study presented convincing dose-response information and was well designed and executed in terms of

experimental protocol and statistical evaluation, which included tests for false positive and false negative responses.

While not all confounders could be ruled out (e.g., compensatory mechanisms, biorhythms, workers who leave because

of styrene related illness), careful attention was paid to include eligibility criteria for the control group that correct for

confounders unique for this population (e.g., limited ethanol intake, a control work-force not exposed to neurotoxic

substances, and a test to allow a match for general intelligence). The use of urinary metabolites to measure exposure

dose is based on the observation that the next-morning urinary mandelic acid (MA) and phenylglyoxylic acid (PGA) is

directly related to the air level of styrene. The Guillemin et al. (1982) study provides the basis for the conversion of

urinary MA+PGA levels to styrene exposure levels used by Mutti et al. (1984b). The quantal dose-response data by

Mutti et al. (1984b) is applicable for use in a benchmark concentration (BMC) approach. The quantal grouping of the

number of subjects that performed abnormally in >3 tests based on their urinary metabolite concentrations was chosen

for a BMC analysis. Basing the BMC on abnormal responses to >3 tests reduces the complexity of multiple test

comparisons and the potential for inappropriate comparison of different neuropsychological tests between control and

exposure groups for statistical purposes. Also, the potential for false positive responses is reduced due to the zero

background level of abnormal responses in the control group when the criteria are >3 abnormal tests. Using a log-

normal probit analysis (Tox-Risk, version 3.5; ICF-Kaiser Inc., Ruston, LA) the maximum likelihood estimate (MLE)

for a 5% response was 17 mg/m

(4.0 ppm). The resulting 95% lower confidence limit at the MLE provided a BMC

7.2 mg/m

(1.7 ppm). A BMC

is considered to be similar to a NOAEL in estimating a concentration associated with a

low level of risk. Following adjustment for exposure continuity (10 m

per 20 m

day for 5 d/wk) and application of an

UF of 3 to account for human intraspecies variability, a REL of 0.9 mg/m

(0.2 ppm) was attained. For exposure data

that utilizes healthy human subjects, the resulting BMC represents a less than 10% incidence in the general population.

When combined with an UF of 3, as carried out above, the resulting REL will be protective of the vast majority of

individuals. This analysis of the quantal data is supported by recognizing that, in a population of 50 subjects, individual

test-specific effects that occur at low doses may not have been observed. If the criterion for abnormality is expressed in

terms of CNS dysfunction, defined by all tests, the sensitivity of the testing procedure is increased and the low dose

effects are more easily observed. The quantal data of Mutti et al. (1984b), i.e., the proportion of subjects responding

abnormally to the tests, therefore provide a more sensitive approach to detecting low dose eff ects. Collasping a battery

of test data to increase sensitivity may introduce the dilemma of multiple test comparisons, as noted above. However,

OEHHA believes that a statistical method to correct for this, known as a Bonferroni correction, is unnecessary. The

REL development is based on calculating a statistic of one effect of a complex of responses (or a syndrome) that results

from CNS dysfunction, and not based on calculating a statistic for each test within the group of tests. The apparent

global nature of the neurological syndrome resulting from long-term styrene exposure, in addition to basing the BMC

on abnormal responses to >3 tests, should more than adequately address any concerns that may result from combining

neurological test data. Applying NOAEL/LOAEL methodology to the Mutti et al. (1984b) quantal data yields an

exposure value similar to that attained with the BMC approach. The LOAEL of 64 mg/m

(15 ppm) is adjusted to an

equivalent continuous exposure of 23 mg/m

(64 mg/m

x 10/20 m3 x 5/7 d/wk). Use of a LOAEL UF of 3 and an

intraspecies UF of 10 resulted in an estimated REL of 0.8 mg/m

(0.2 ppm).

The INDEX project Final report

205

The U.S. EPA (1996) calculated a reference concentration (RfC) of 1 mg/m

(0.3 ppm), which is slightly higher than the

OEHHA-derived chronic REL of 0.9 mg/m

(0.2 ppm). The RfC for styrene is also based on the findings of Mutti et al.

(1984b), but utilized the continuous data for development of the RfC and used standard NOAEL methodology for the

RfC derivation. U.S. EPA (1996) established a NOAEL for the lowest exposure group (<150 MA+PGA mmole/mole

creatinine; equivalent to < 106 mg/m

styrene).

However, OEHHA staff believe that the use of the continuous data to establish a NOAEL overlooks the advantages of

using the BMC approach using the quantal data. These advantages are that the BMC

05 reflects the shape of the dose-

response curve and takes into account the number of subjects involved in the study. In addition, OEHHA staff evaluated

the quantal data with the Fisher’s Exact Test and determined the probabilities of abnormal responses among the exposed

subjects based on the unexposed subjects whose responses were assumed to be normal. At the lowest exposure, the

probability that the proportion of subjects responding abnormally to >1 and >3 tests was within the expected range was

p = 0.005 and p = 0.04, respectively, indicating that neuropsychological deficits due to styrene occur in the low dose

subgroup. Thus, the quantal data indicate that a NOAEL was not established in this study.

U.S.EPA - IRIS (1992)

Integrated Risk Information System

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC).

Critical effect Exposures* UF MF RfC

CNS effects

Occupational Study

Mutti et al. (1984b)

NOAEL: 94 mg/cu.m (25 ppm = 150 mmole

urinary styrene metabolites/mole creatinine

adjusted to lower 95%

confidence limit = 22 ppm)

NOAEL(HEC): 34 mg/cu.m

LOAEL: >94 mg/cu.m (>22 ppm

derived as in NOAEL listing)

30 1 1 mg/m

*Conversion Factors: MW = 104.15. Assuming 25 C and 760 mmHg, NOAEL (mg/cu.m) = NOAEL (ppm) x MW/24.45 = 94 mg/cu.m. The NOAEL

exposure level is based on a back extrapolation from worker urinary concentration of styrene metabolites reported in the principal study and adjusted

to the lower 95% confidence limit listed in Guillemin et al. (1982), which was 88%, 25 ppm x 0.88 = 22 ppm. The NOAEL(HEC) is calculated using

an 8-hour TWA occupational exposure. MVho = 10 cu.m/day, MVh = 20 cu.m/day. NOAEL(HEC) = 94 mg/cu.m x MVho/MVh x 5 days/7 days =

34 mg/cu.m. The feasibility of applying the exposure model of Perbellini et al. (1988) for extrapolation of the values in the principal study is currently

being investigated. Application of this model may result in changes in the NOAEL(HEC) value and, therefore, the RfC.

The concentration-response relationship between urinary metabolite concentration (mandelic acid and phenylglyoxylic

acid levels normalized to creatinine in "morning-after" urine) and test results indicated a significant effect level in the

subgroup whose urine contained 150-299 mmole urinary metabolites/mole creatinine. Workers with metabolite

concentrations of up to 150 mmoles/mole appeared to have no significant effects, and this level is therefore designated

as the NOAEL in this study. The authors state that this level of urinary metabolites corresponds to a mean daily 8-hour

exposure to air styrene of 107 mg/m

(25 ppm). Derivation of this air level is from the creatinine-normalized, combined

concentration of the styrene metabolites, MA and PGA, in urine collected from the workers on Saturday mornings.

Guillemin et al. (1982) demonstrated a logarithmic relationship (r = 0.871) between the summation of urinary

metabolites (MA + PGA, next morning) and air concentrations of styrene (ppm x hours). Guillemin calculated the mean

combined urinary metabolite concentration (next morning) for an 8-hour exposure to 426 mg/m

. This relationship was

used by both Mutti et al. (1984b) and Guillemin and Berode (1988) in a proportional manner to obtain styrene air levels

at lower urinary metabolite concentrations. The 95% confidence interval was also calculated for an 8-hour exposure at

426 mg/m

, the lower limit of the confidence calculation being 88% of the mean styrene exposure. This factor was

applied directly to the NOAEL of 107 mg/m

[107 mg/m

x 0.88 = 94 mg/m

]. Due to the construction of the

subgroups, designation of a LOAEL was the lower limit of the subgroup in which adverse effects were observed [i.e.,

greater than the NOAEL of 94 mg/m

Confidence in the Inhalation RfC: The study of Mutti et al. (1984b) documents concentration-response relationships of

CNS effects in a relatively small worker population. However, the results of this study are consistent with a number of

other studies showing central effects in chronically exposed worker populations, most notably that of Moller et al.

(1990). The urinary metabolites, MA and PGA, are direct biological indicators of exposure to styrene. Numerous

studies have demonstrated the relationship between urinary metabolites and air levels of styrene to be reliable and

quantitative. Physiologically based pharmacological modeling of this exposure methodology demonstrates that it

reflects and incorporates at least a portion of intrahuman variability related to pharmacokinetics. The study is therefore

assigned a medium confidence level. The data base can be considered medium to high as chronic laboratory animal

studies addressing noncancer endpoints are not yet available, but a number of human exposure studies support the

choice of critical effect. Preliminary information in mice indicate that styrene is a respiratory tract irritant in mice at

concentrations lower than 47.5 mg/cu.m. The RfC is assigned an overall confidence rating of medium.

The INDEX project Final report

206

ATSDR (1992)

Agency for Toxic Substances and Disease Registry

Derivation of the acute Minimal Risk Level (MRL):

Studies in workers exposed to styrene in workplace air suggest that neurological effects are probably the most sensitive

indicator of styrene toxicity. Available data do not.provide a clear picture of the NOAEL for neurological effects

following acute- or intermediate-duration inhalation exposure, but two chronic studies in workers identify LOAELs of

107 mg/m

(25 ppm) (Mutti et al. 1984a) and 132 mg/m

(31 ppm) (Harkonen et al. 1978). Based on the lower LOAEL

(25 ppm), a chronic inhalation MRL of 0.25 mg/m

(0.06 ppm) has been derived using factors of 8/24 and 5/7 to

account for exposure 8 hours/day, 5 days/week, and an uncertainty factor of 100 (10 to account for human variability,

and 10 to account for use of a LOAEL).

WHO (2001)

Air Quality Guidelines for Europe, 2000

Although genotoxic effects in humans have been observed at relatively low concentrations, they were not considered as

critical endpoints for development of a guideline, in view of the equivocal evidence for the carcinogenicity of styrene.

In occupationally exposed populations, subtle effects such as reductions in visuomotor accuracy and verbal learning

skills, and subclinical effects on colour vision have been observed at concentrations as low as 107–213 mg/m

(25–50

ppm). Taking the lower number of this range for precautionary reasons and adjusting this value in order to allow for

conversion from an occupational to a continuous pattern of exposure (a factor of 4.2), and incorporating a factor of 10

for interindividual variation and 10 for use of a lowest-observed-adverse-effect level (LOAEL) rather than a no-

observed-adverse-effect level (NOAEL) this results in a guideline of 0.25 mg/m

(weekly average). This value should

also be protective for the developmental neurological effects observed in animal species.

The air quality guideline could also be based on the odour threshold. In that case, the peak concentration of styrene in

air should be kept below the odour detection threshold level of 70 µg/m

as a 30-minute average.

Health Canada

Canadian Environmental Protection Act (CEPA).

Derivation of a tolerable concentration (TC):

The available data on nonneoplastic effects (principally neurological) in humans are considered to be inadequate to

serve as the basis for the development of a tolerable concentration (TC), due to such factors as the lack of precise

exposure data, simultaneous exposures to other chemicals, the short duration of available clinical studies, and the small

numbers of subjects in many studies. It should be noted, however, that a TC derived on the basis of neurotoxic effects in

cross-sectional epidemiological studies and clinical studies in human volunteers would not vary greatly from that

derived below on the basis of studies in animal species.

The lowest reported levels inducing meaningful effects in experimental animals exposed to styrene by inhalation fall

within a similar range; indeed, there is no clearly superior critical study and TCs derived on the basis of the lowest

effect levels from several studies would be similar. For this assessment, the study by Kishi et al. (1992a,b) has been

selected for the derivation of a TC since it leads to the most conservative value and since observed effects included both

body weight changes and manifestations of neurotoxicity (including biochemical and behavioral effects).

On the basis of the 260 mg/m

(60 ppm) LOEL of the Kishi et al. (1992a,b) studies, a TC of 0.092 mg/cu.m was derived

for inhalation exposure. A factor of 6/24 was used to convert from intermittent to continuous exposure. The ratio of

inhalation volume to body weight of rats [(0.11 m

/day)/0.35 kg] to humans aged 5 to 11 [(12 m

/day)/27 kg] was

applied. An uncertainty factor of 500 (10 for interspecies variation, and 10 for intraspecies variation and 5 for use of a

LOEL as the effects observed at this concentration were not clearly adverse).

Limited available data indicate that humans form less of the putative toxic metabolite, styrene-7,8-oxide, and hydrolyze

it more quickly than experimental animals; however, available data were insufficient to take these differences into

account in the development of the uncertainty factor, and the relevance of this metabolite to developmental and

neurotoxic effects is not clear.

This tolerable concentration is within the range of that which would be derived on the basis of neurological effects in

occupationally exposed populations.

The INDEX project Final report

207

Macromolecular adducts and genetic toxicity

The genetic toxicology of human styrene exposure and styrene macromolecular binding have been reviewed by several

authors (IARC, 1994; Phillips and Farmer, 1994; Barale 1991; Norppa and Sorsa, 1993) and have mostly been studied

in the reinforced plastics industry where styrene is the main chemical exposure. Only a few studies are available on

other branches of industry, such as styrene production, where exposure to styrene is low. Exposure levels in the studies,

evaluated on the basis of urinary styrene metabolites or workplace air samples, have ranged widely.

The levels of the N-terminal valine adduct of haemoglobin, N-(1-hydroxy-2-phenylethyl)valine, have been found to be

four times higher in styrene-exposed workers than in controls. In a Swedish-Finnish study (Christakopoulos et al.,

1993), increased levels of adducts to N-terminal valine in haemoglobin were seen in reinforced plastics workers (mean

28 pmol × g

-1

globin) in relation to control persons (mean <13 pmol × g

-1

globin). This level was low, while a

background adduct level was also seen among the controls. A significant correlation of individual adduct levels and free

styrene-7,8-oxide and styrene glycol in blood, as well as mandelic acid in urine, was seen in a regression analysis.

Extrapolation from the metabolite monitoring data gave an estimate of about 320 mg/m

(75 ppm) for the workplace air

styrene concentration. In another study, using similar techniques, but at lower exposure levels (30 mg/m

7 ppm), no

styrene-7,8-oxide adducts in N-terminal valine of haemoglobin were seen (Severi et al., 1994).

The first DNA adducts studies on humans exposed to styrene were published by Liu et al. (1988) who detected styrene-

7,8-oxide modified guanine adducts in a single exposed person but none in a single unexposed individual.

Extensive studies on biomonitoring of styrene exposure using DNA adducts have been performed (Canadian

Environmental Protection Act, 1993; Vodicka and Hemminki, 1993). Lamination workers from work sites representing

high exposures (mean air styrene level 370 mg/m

87 ppm) and moderate exposures (mean air styrene level 210 mg/m

49 ppm) were studied. Using synthetic standards and the

P-postlabelling technique, O

-(2-hydroxy-1-phenylethyl)-2'-

deoxy-guanosine-3'-monophosphate (O

-dGMP) adducts were detected among the exposed workers, at 7–16 times

higher level than in the controls.

In a further study, where DNA adducts were investigated before and after a 2-week vacation, no decline in adduct levels

was seen, indicating a slow removal of the specific O

-guanine adducts of styrene from DNA (Vodicka et al., 1994).

Horvath et al. (1994) studied DNA adducts by postlabelling in mononuclear blood cells of lamination workers and

found two types of adducts, one identified as guanine N

adduct, to be increased with a significant linear relationship to

individual measurements of airborne styrene, but not to SCE frequencies among the same workers. DNA adduct levels

were highly elevated among the styrene exposed workers, with a correlation to DNA strand breakage measured by the

comet assay, but not to mutant frequencies (hypoxanthine-guanine phosphoribosyl transferase (HPRT) locus) (Vodicka

et al., 1995). These results suggest that no simple quantitative relationships exist between the adducts and other

parameters of DNA damage in human lymphocytes. In this study, control subjects from the administration of the same

plastics factory also showed elevated adduct levels in comparison with unexposed control persons.

In studies on cytogenetic parameters – chromosomal aberrations, SCEs, and micronuclei – in peripheral lymphocytes of

workers employed in the reinforced plastics industry (Sorsa et al., 1992; Mäki-Paakkanen et al., 1991; Brenner at al.,

1991; Meretoja et al., 1977; Meretoja et al., 1978; Fleig and Thiess, 1978; Thiess and Fleig, 1978; Högstedt et al., 1979;

Thiess et al., 1980; Andersson et al., 1980; Watanabe et al., 1981; Watanabe et al., 1983; Högstedt et al., 1983; Camurri

et al., 1983; Camurri et al., 1984; Hansteen et al., 1983; Nordenson and Beckman, 1984; Pohlova and Sram, 1985;

Mäki-Paakkanen, 1987; Forni et al., 1988; Jablonicka et al., 1988; de Jong et al., 1988; Hagmar et al., 1989; Kelsey et

al., 1990; Yager et al., 1990; Yager et al., 1993; Perera et al., 1992; Tomanin et al., 1992; Hallier et al., 1994; Van

Hummelen et al., 1994; Artuso et al., 1995; Anwar and Shamy, 1995; Norppa et al., 1981), positive findings have

primarily concerned chromosomal aberrations, which have been observed to be increased in the majority of the studies.

SCEs and micronuclei were also found to be increased in some studies.

On the basis of rough exposure estimates and observed frequencies of chromosomal aberrations, the NOAEL for long-

term styrene exposure has been proposed to be 85–128 mg/m

(20–30 ppm) (Knudsen and Sorsa, 1993) or 213 mg/m

(50 ppm) (Barale, 1991).

A recent meta-analysis on cytogenetic data from 25 reports on occupational styrene exposure (Bonassi et al., 1996)

showed a significant increase in weighed frequency ratio for chromosomal aberrations from studies with median styrene

exposure above the chosen dichotomization point, 125 mg/m

(30 ppm); results for SCEs and micronuclei were

inconclusive. Fleig and Thiess (1978) observed an increase in chromosomal aberrations in a worker group with high

exposure to styrene (mean 486 mg/m

, 114 ppm) but not in two other groups with lower exposure levels (2 and 8

mg/m

). Anderson et al. (1980) observed an increase in the frequency of chromosomal aberrations with increasing total

styrene exposure (expressed as the average concentration in mg/m

multiplied by the number of years of employment)

The INDEX project Final report

208

in a low-dose group (mean total styrene exposure 137 mg/m

) but not in a high-dose group (mean 1204 mg/m

). Forni et

al. (1988) found an increase of chromosomal aberrations in workers from a factory where current exposure to styrene

was high (41–198 mg/m

) but not in workers from another factory with low current exposure (1.7–17 mg/m

); the latter

group of workers – who had experienced a high exposure to styrene in the past and had, therefore, a high cumulative

exposure – showed an increase in chromosomal aberrations. Mäki-Paakkanen et al. (1991) obtained a significant

correlation between the number of cells with chromosomal aberrations and years of exposure. Tomanin et al. (1992)

could demonstrate an elevation of chromosomal aberrations in a high-exposure group (115–443 mg/m

) but not in a

low-exposure group (21–102 mg/m

). Artuso et al. (1995) obtained a significant styrene-exposure dependent trend for

both chromosomal aberrations and SCEs among a high-dose group (gel coaters, laminators, rollers and assemblers;

exposure range 85–1389 mg/m

), a low-dose group (other tasks; exposure range 2–119 mg/m

), and controls.

Camurri et al. (1984) observed a significant increase in SCEs in workers in reinforced plastics plants with mean

airborne styrene levels of more than 200 mg/m

but not in plants with mean levels of 30–200 mg/m

. Yager et al. (1993)

obtained a clear relationship among a group of boat manufacturers between SCE frequency and styrene exposure

measured either as the concentration of styrene in workroom air (0.9– 234 mg/m

; mean 64 mg/m

) or in exhaled air.

Hallier et al. (1994) observed an increase in SCEs among laminators exposed to high concentrations of styrene

(approximately 170 mg/m

) but not in formers exposed to much lower levels (43 mg/m

). When hygienic and technical

improvements at the workplace reduced the styrene exposure of the laminators to approximately 85 mg/m

, a significant

reduction in the SCE levels of the laminators was noted, although their values were still higher than in unexposed

controls.

The few studies available on DNA strand breakage and N-acetylaminofluorene-induced unscheduled DNA synthesis in

peripheral leukocytes have given positive results. DNA strand breaks were shown to disappear quickly from the

peripheral leukocytes following the exposure, so that blood samples collected before the work shift could be used as

control samples (Walles et al., 1993). Reinforced plastics workers showed elevated frequencies of GPA variant

erythrocytes – a measure of mutations in the glycophorin A locus – with a significant effect especially among workers

belonging to a high-exposure group (breathing zone styrene concentration ≥ 85 mg/m

, 20ppm) (Bigbee et al., 1996).

Another study also suggested an increase in GPA variant frequencies in a high-exposure group (average exposure

concentration 136 mg/m

, 32 ppm) as compared with a low exposure group (average exposure to 5.1 mg/m

, 1.2 ppm

styrene), but final conclusions were complicated by inadequate matching of the groups (Compton-Quintana et al.,1993).

Vodicka et al. (1995) observed a weak elevation in the frequency of HPRT mutations in T-lymphocytes, while another

study (Tates et al., 1994) was considered inconclusive.

In summary, genotoxic effects have been observed in the blood cells of reinforced plastics workers for various

endpoints at styrene exposure levels of around 85–128 mg/m

(20-30 ppm) and above. DNA breakage has been

observed at exposure levels below 43 mg/m

. The role of styrene in generating the genotoxic effects seen in reinforced

plastics workers is supported by the elevated levels of styrene-7,8-oxide DNA adducts observed in their peripheral

leukocytes. Such adducts appear to be distinguishable following occupational exposures to only a low concentration of

styrene.

Genotoxicity assays on styrene in experimental animals have given conflicting results. The findings suggest a weak

genotoxic activity for styrene in vivo, as revealed by the most sensitive techniques (SCEs and DNA strand breakage),

especially in the mouse – the rodent species expected to be more susceptible than the rat or Chinese hamster to styrene.

Carcinogenic potential

IARC has classified styrene as a Group 2B, possibly carcinogenic to humans. EPA does not have a carcinogen

classification for styrene; the chemical currently is undergoing an EPA Integrated Risk Information System (IRIS)

review to establish such a classification. ATSDR, Health Canada, and RIVM have evaluated the carcinogenicity data

for styrene. Health Canada classified styrene as "possibly carcinogenic to humans" (Group III) and also a "possible

human germ cell mutagen"(Group III). RIVM concluded that styrene is not a genotoxic compound, and that the

carcinogenic potential of styrene is related to the metabolite styrene oxide, but concentration of this metabolite in

humans is very low due to rapid biotransformation to styrene glycol. Consequently, RIVM derived risk values based on

noncancer endpoints. ATSDR has published a Toxicological Profile for Styrene. Although ATSDR discusses the

carcinogenicity data in its Toxicological Profiles, it does not currently assess cancer potency or perform cancer risk

assessments.

Cases of leukaemia and lymphoma were identified among workers exposed in the manufacture of styrene and

polystyrene, and in the production or manufacture of styrene–butadiene (IARC, 1994).

The INDEX project Final report

209

Epidemiological studies of styrene have been conducted in three types of industry: production of glass-reinforced

plastics products, production of styrene monomers and styrene polymerization, and production of styrene–butadiene

rubber. The malignancies observed in excess most frequently are of the lymphatic and haematopoietic system.

Reinforced plastics industry

In a European multinational study of more than 40 000 workers in the glass-reinforced plastics industry (660 plants), no

overall excess of deaths from lymphatic and haematopoietic cancers was observed in comparison with national controls

(Kogevinas et al., 1994). An increased risk for lymphatic and haematopoietic tumours was observed in Poisson

regression models for years since first exposure (P = 0.012) and for average exposure (P = 0.19) but not for cumulative

exposure. Within the models, there was an increasing trend in risk for lymphatic and haematopoietic cancer with

average intensity of exposure, culminating in a relative risk (RR) of 3.6 (95% confidence interval (CI), 1.0–13) for the

highest category, >852 mg/m

; for more than 20 years since first employment, the RR was 4.0 (95% CI, 1.3–12).

Nonsignificant increases in risk with time since first exposure or cumulative exposure were noted for cancers of the

pancreas, kidney, and oesophagus.

A study of cancer incidence in the reinforced plastics industry in Denmark involved 12 800 male workers who had been

included within the above study and a further 24 000 workers with a lower probability of exposure to styrene. The mean

annual air concentrations of styrene calculated for 128 of the companies studied ranged from 767 mg/m

in 1964–1970

to 183 mg/m

in 1976–1988. Within this cohort, there were 112 malignancies of the lymphatic and haematopoietic

system, with 93.7 expected (standardized incidence ratio (SIR), 1.2; 95% CI, 0.98–1.4). In workers with more than 10

years since first employment, the SIR for leukaemia was 157 (107–222). The excess was concentrated mainly in those

workers not previously included in the European study, in short-term workers with at least 10 years since first

employment and in those employed before 1970 (Kolstad et al., 1994). The same authors also reported the occurrence

of deaths from solid cancers among the 36.610 reinforced plastics workers (Kolstad et al., 1995). An increased

incidence rate ratio (IRR) for pancreatic cancer was found (IRR 2.2, 95% CI, 1.1– 4.5).

In a large study on 5826 employees who had worked for at least 6 months between 1948 and 1977 in 30 reinforced

plastics plants in the United States, no overall increase in risk for lymphatic and haematopoietic malignancies was

observed (Wong, 1990; Wong et al., 1994). The follow-up continued until the year 1989. The overall mortality rate was

108 (95% CI, 103–113), and the mortality rate from all cancers was 116 (95% CI, 105–127). Total lympho-

haematopoietic cancers showed no excess. For workers involved in open-mould processing with high exposure to

styrene, the standardized mortality ratio (SMR) for lymphatic and haematopoietic cancers was 141 (based on four

cases). For the highest cumulative exposure (>426 mg/m

× years) and more than 20 years of latency, the SMR was 134

(5–373). Mortality for cancers at a number of sites was increased significantly. These included oesophagus (mortality

rate 198; 95% CI, 105–322), bronchus, trachea, and lung (141; 120–164), cervix uteri (284; 136–521), and other female

genital organs (202; 107–345). No positive dose–response relationship was found for the lympho-haematopoietic

cancers or any other cancer in excess; IARC (1994), however, noted that the possibility that the two exposure variables

included in the regression model were correlated may have reduced the likelihood of accurate assessment

A smaller study of 5021 employees in two reinforced plastics boat-building facilities in the United States showed no

deaths from leukaemia or lymphomas (1.7 and 2.1 expected, respectively) (Okun et al., 1985). 2060 individuals were

considered to have had high exposure to styrene (mean levels of styrene in the air in two facilities, 181 and 305 mg/m

);

48% of them had worked only for one month to one year and only 5% for more than 5 years. In this group, no

lymphatic or haematopoietic cancers were seen (about 1 expected).

A deficit of deaths from lymphatic and haematopoietic malignancies (6 observed, 14.9 expected) was reported in a

cohort of 7949 men and women employed during 1947–1984 in eight companies in the United Kingdom manufacturing

glass-reinforced plastics involving high exposure to styrene (Coggon et al., 1987). A nonsignificant excess was seen in

deaths from cancers of the lung, pleura and mediastinum (SMR, 126; 95% CI, 94–166); this finding particularly related

to workers who had had 1–9 years of exposure to styrene, but the risk did not increase with time from the first exposure.

Follow-up of this cohort was later extended to 1990 as part of an international collaborative study (Kogevinas et al.,

1994). By that time the previous deficit of lymphatic and haematopoietic cancer had largely disappeared (13 deaths;

SMR, 88; 95% CI, 47–151) and the excess of lung cancer was less marked (77 deaths; SMR, 106; 95% CI, 84–132).

Styrene manufacture and polymerization

A study of chemical workers in the production of styrene and styrene derivatives in the United States found seven

deaths from lymphatic and haematopoietic malignancies (except leukaemias) (SMR, 132; 95% CI, 58–272) and six

from leukaemias (SMR 176; 95% CI, 64–383) (Ott et al., 1980). In an update, a total of 28 deaths from lymphatic and

haematopoietic cancer were recorded (SMR 144; 95% CI, 95–208) (Bond et al., 1992).

The INDEX project Final report

210

A smaller United Kingdom study of 622 men exposed for at least a year in the production, polymerization and

processing of styrene found an excess of deaths from lymphoma (3 observed, 0.56 expected) (Hogdson and Jones,

1985).

Styrene–butadiene rubber production

The large cohorts of styrene–butadiene rubber workers showed increased risks for lymphatic and haematopoietic

malignancies, but nested case–control analyses found little evidence for relationship to styrene (Matanoski et al., 1990;

Santos-Burgoa et al., 1992; Matanoski et al., 1993). The styrene exposures in these industries are at least one order of

magnitude lower than in the reinforced plastics industry.

In summary, several epidemiological studies have suggested that workers exposed to styrene in the reinforced plastics

industry have increased risk of lymphatic and haematopoietic tumours. The studies are not, however, fully conclusive

because the observed associations are often based on small numbers, and in larger studies, dose relations are somewhat

obscure.

Interactions with other chemicals

Styrene metabolism is known to be inhibited by the presence of other chemicals such as toluene, trichloromethylene,

and ethyl benzene. The biotransformation of styrene in rats to PGA, MA, and hippuric acid was suppressed by

coadministration of toluene (Ikeda et al. 1972). This may be due to competitive inhibition of oxidative mechanisms.

Similar results were reported by Ikeda and Hirayama (1978) in rats when styrene metabolism was inhibited by the

administration of trichloroethylene. Urinary metabolites of styrene may be markedly reduced when humans or animals

are concurrently exposed to organic solvents that inhibit styrene metabolism.

Odour perception

Source: WHO (2001)

The odour threshold for styrene is only 70 µg/m

(0.016 ppm). Its characteristic pungent odour is recognized at

concentrations 3–4 times higher than this threshold value. Some individuals can perceive the odour at levels lower than

70 µg/m

, but, in general, odour problems are not likely to occur if peak concentrations in the ambient air are kept

below this threshold value. When styrene is emitted into the air, its half-time is estimated to be 2 hours. In ambient air it

is chemically transformed into benzaldehyde and formaldehyde, both of which are odorous air pollutants (WHO, 1987).

Source: AIHA (1989)

Detection: 72 – 8100 µg/m

(0.017-1.9 ppm)

Recognition: 640 µg/m

(0.15 ppm)

Source: New Jersey department of health and senior services - Hazardous Substance Fact Sheet

Odor threshhold = 85 – 2000 µg/m

(0.02 - 0.47 ppm)

Source: (Hoshika et al., 1993)

Barely perceptible concentration level of styrene: 0.068/0.140 mg/m

(in The Nederlands and Japan, respectively)

International comparison of odor threshold values of several odorants in Japan and in The Netherlands, given as the

barely perceptible concentration level revealed striking similarities for hydrogen sulfide (in Japan 0.0005 ppm/in The

Netherlands 0.0003 ppm), phenol (0.012/0.010), styrene (0.033/0.016), toluene (0.92/0.99), and tetrachloroethylene

(1.8/1.2) but not for m-xylene (0.012/0.12). Such a similarity was not found with any other literature sources.

The INDEX project Final report

211

Summary of Styrene Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Dermal absorption from ambient air <5% of the dose absorbed in the respiratory tract

Toxicokinetics: 60-70% pulmonary retention; widely distributed throughout the body; styrene-oxide in blood linearly

correlated with ambient styrene; no evidence of accumulation at occupational levels (160 mg/m

) due to fast enzymatic

oxidation/hydrolysis to MA and PGA and biphasic elimination; MA+PGA used for biological monitoring.

Health effect levels of short- and long-term exposure (noncancer)

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

217

EXP

422

EXP

Eye and throat irritation

Volunteers,

1h (NOAEL),

20m (LOAEL)

Stewart et al., 1968 OEHHA (1999);

REL: 21

Long-term exposure

Study

MLE

7.2

BC05

2.6

ADJ

CNS; neuropsychological tests: reaction time, s/l

term logic memory, s/l term verbal memory,

digit-symbol association, block design and figure

identification.

Occupational, 8.6y Mutti et al. 1984b

OEHHA (1999);

REL: 0.9

95%lCL

ADJ

>94

95%lCL

see above Occupational, 8.6y;

Mutti et al. 1984b

EPA-IRIS (1992)

; RfC: 1

107

Study

ADJ

see above Occupational 8.6y; Mutti et al. 1984b ATSDR (1992);

MRL: 0.25;

WHO (2001);

GV: 0.26;

RIVM (2000)

260

EXP

ADJ

HEC

Neurological developmental; body weight,

biochemical parameters in the brain and

behaviour

Rat offspring Kishi et al.,1992a,b Health Canada

(1993);

TC: 0.092

Carcinogenicity: IARC: 2B ; ACGIH: A4 ; Little evidence of carcinogenicity in humans (WHO)

Genotoxicity: Genotoxic in vitro and in vivo; increased chromosomal aberrations evidenced in numerous occupational

studies; proposed NOAELs based on this endpoint were >85 mg/m

. Although genotoxic effects in humans have been

observed at relatively low concentrations, they were not considered as critical endpoints for development of a guideline,

in view of the equivocal evidence for the carcinogenicity of styrene (WHO).

Odour threshold: Range: 0.072 - 8.1 mg/m

, recognition: 0.64 mg/m

(AIHA); Threshold: 0.07 mg/m

(WHO); 0.23

(Devos); Barely perceptible concentration levels: 0.140/0.068 mg/m

in Japan/The Nederlands, respectively (Hoshika et

al., 1993).

Susceptible population: Asthmatics may be more sensitive to adverse pulmonary effects; immediate

bronchoconstriction observed in occupational (styrene) asthmatics at 64 mg/m

Remarks: RD

: 3.3 g/m

; styrene toxicity is enhanced (metabolism slowed down, with increase of styrene oxide in

blood) by exposure to other solvents (incl.ethanol)

________________________________________________________________________________________________

The INDEX project Final report

212

6. Risk Characterization

Health hazard and cancer risk evaluation of short- and long-term exposure

(Adapted from WHO, 2001 - Air Quality Guidelines for Europe)

Potentially critical effects for the derivation of an exposure limit (EL) for styrene are considered to be

carcinogenicity/genotoxicity and neurological effects, including effects on development. The value of the available

evidence for an association between exposure to styrene and small increases in lymphatic and haemotopoietic cancers

observed in workers in some studies is limited by concurrent exposure to other substances, lack of specificity and

absence of dose–response. In limited studies in animals, there is little evidence that styrene is carcinogenic. IARC has

classified styrene in group 2B.

Styrene was genotoxic in vivo and in vitro following metabolic activation. In cytogenetic studies on peripheral

lymphocytes of reinforced plastics workers, there were increased rates of chromosomal aberrations at mean levels of

styrene of more than 120 mg/m

. Elevated levels of single-strand breaks and styrene-7,8-oxide adducts in DNA and

haemoglobin have also been observed. Although these genotoxic effects have been observed at relatively low

concentrations, they were not considered as critical endpoints for the development of a WHO Air Quality Guideline, in

view of the equivocal evidence of carcinogenicity for styrene.

The available data, although limited, indicate that neurotoxicity in the form of neurological developmental impairments,

is among the most sensitive of endpoints. In the offspring of rats exposed to styrene at a concentration of 260 mg/m

there were effects on biochemical parameters in the brain and behaviour.

Studies in workers exposed to styrene in workplace air suggest that neurological effects are probably the most sensitive

indicator of styrene toxicity. Available data do not.provide a clear picture of the NOAEL for neurological effects

following acute- or intermediate-duration inhalation exposure. A chronic study in workers (Mutti et al. 1984a) identified

subtle effects such as reductions in visuomotor accuracy and verbal learning skills, and subclinical effects on colour

vision have been observed at concentrations as low as 107–213 mg/m

. Taking the lower number of this range as a

LOAEL (for precautionary reasons) and adjusting this value in order to allow for conversion from an occupational to a

continuous pattern of exposure (dividing by 4.2) a concentration of 25 mg/m

is obtained. Incorporating a factor of 10

for interindividual variation and 10 for use of a LOAEL rather than a NOAEL this results in an EL of 0.25 mg/m

(to be

applied as a weekly average; see also Table 6.1). This value should also be protective for the developmental

neurological effects observed in animal species.

Since available data do not.provide a clear picture of the NOAEL for neurological effects following acute exposure, a

short-term EL was here derived based on irritation of the eyes and mucous membranes. Stewart et al. (1968) found eye

and throat irritation in 3 out of 6 volunteers exposed to 422 mg/m

styrene for 20 minutes. No symptoms were reported

in 3 subjects after exposure to 217 mg/m

for 1 hour. A short-term EL is here derived applying an assessment factor of

100 for interindividual variation (10) and the possibility (limited evidence) that asthmatics may be more sensitive to

adverse pulmonary effects (10) from styrene exposure (Moscato et al., 1987).

The odour threshold for styrene is only 70 µg/m

. Its characteristic pungent odour is recognized at concentrations 3–4

times higher than this threshold value. Some individuals can perceive the odour at levels lower than 70 µg/m

, but, in

general, odour problems are not likely to occur if peak concentrations in the ambient air are kept below this threshold

value.

Table 6.1: Derivation of short- and long-term exposure limits

Effect level - mg/m

Assessment

factor

mg/m

Toxicological endpoint

Short-term

Exposure

Limit

Human NOAEL

Volunteers

217 100

2 Eye and throat irritation

Long-term

Exposure

Limit

Human LOAEL

Occupational

25 100

0.25

CNS; neuropsychological tests: reaction

time, s/l term logic memory, s/l term verbal

memory, digit-symbol association, block

design and figure identification.

not considering a NOAEL (10);

interindividual variation (10) ;

susceptible population (10)

The INDEX project Final report

213

Relevance of EU-population exposure to styrene

Table 6.2: Percentage of population exposed beyond the derived EL and margins of safety

Available exposure data

Margins of

Safety (MOS)

Description (Study, Year) N 0.25* mg/m

(90

)

Athens 30-h TWA (Expolis, 96-98) 42 < 181 (71)

Basel 30-h TWA (Expolis, 96-98) 47 < 470 (182)

Helsinki 30-h TWA (Expolis, 96-98) 188 < 184 (103)

Milan 30-h TWA (Expolis, 96-98) 38 < 77 (20)

Oxford 30-h TWA (Expolis, 98-00) 40 < 225 (64)

Prague 30-h TWA (Expolis, 96-98) 46 < 164 (64)

German Survey 48-h PEM (GerEs II,

1990-92)

113 < 125 (36)

French National Survey 7d-TWA (IAQ

observatory; 2003-04)

109 < 250 (96)

Avg. MOS: 235 (79)

< out of the evaluation range (i.e. <5% of the environments investigated);

* corresponds with WHO guideline value (weekly average)

Result

A long-term exposure limit of 250 µg/m

has been derived based on the assumption that neurological effects are

probably the most sensitive indicator of styrene toxicity. When examining the results of eight monitoring surveys (Table

6.2) it can be concluded that background styrene concentrations in European residences are of no concern to human

health since median levels are, on average, two orders of magnitude below the established EL. Although no acute

exposure data were available, it is unlikely that styrene emissions associated with human indoor activities would

generate levels up to the proposed short-term EL of 2000 µg/m

, considered protective for irritative effects in

asthmatics. Although genotoxic effects in humans have been observed at relatively low concentrations, they were not

considered as critical endpoints for the derivation of the exposure limit, in view of the equivocal evidence for the

carcinogenicity of styrene in humans (WHO).

The INDEX project Final report

214

Ammonia

Synonyms: -

CAS Registry Numbers: 7664-41-7

Molecular Formula: NH

1. Compound identification

Ammonia is a colourless gas with a sharp, irritating odour. Ammonia has both natural and anthropogenic sources. It is a

key compound in the global nitrogen cycle. It is formed in the body during decomposition of organic materials. The

most common natural sources of ammonia include the natural breakdown of manure and dead plants and animals.

Majority of all manufactured ammonia is used as fertilizers. It is also released to the atmosphere by leaks during

commercial synthesis, production, and transportation, or sewage, burning of coal, wood, and other natural products, and

volcanic activity. Ammonia is used to manufacture synthetic fibres, plastics, explosives, and many cleaning products. It

is also used as a refrigerant. (WHO 1997, ATSDR, 2002)

Since ammonia occurs naturally in the environment, the general population is exposed to ammonia in air, soil, and

water. People are frequently exposed to ammonia while using household products that contain ammonia, such as

cleaning solutions, window cleaners, floor waxes, and smelling salts. Household ammonia solutions usually contain 5-

10% ammonia in water. Also those who live near farms or cattle feedlots, poultry confinement buildings, or in the

vicinity of other areas with high animal populations may be exposed to elevated ammonia levels (ATSDR 2002,

EPA/Cal 2003).

Typical environmental ammonia concentrations have not been reported to cause adverse health effects in general

population. However, low levels of ammonia may harm some asthmatics and other sensitive individuals (ATSDR

2002).

2. Physical and Chemical properties

Molecular weight (g/mol) 17.03

Melting point (°C) -77.7

Boiling point (°C) 33.35

Density (g/l at 0 oC, 1 atm) 0.77

Relative density (air =1) 0.59

Solubility: Soluble in water and oxygenated solvents

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 0.707 µg/m3

1 µg/m3 = 1.414 ppb

Source: Verschueren 2001, HSDB 2003.

3. Indoor Air Exposure assessment

Indoor air and exposure concentrations

There is a lack of knowledge and publications about indoor concentrations and exposures of ammonia. Ammonia is

studied mostly as an ambient pollutant. Typical ambient background concentrations of ammonia are between 0.71

The INDEX project Final report

215

µg/m

and 3.55 µg/m

(1 and 5 ppb) (ATSDR 2002). Fisher et al (2003) determined indoor and outdoor concentrations

of ammonia in an unoccupied residence in Clovis California from October 2000 to January 2001. Monthly mean indoor

concentrations ranged from 10.9 µg/m

to 15.1 µg/m

, while mean outdoor concentrations were always lower ranging

from 8.80 µg/m

to 9.66 µg/m

Considerably high ammonia concentrations have been measured in new buildings built in 1994–1997 in Finland.

Average concentrations exceeded 50 µg/m

. Construction materials such as glues, fillers used in floors, walls and

ceilings, have been identified as ammonia sources in indoor environments. Ammonia is emitted from these materials by

chemical breakdown processes, especially in the presence of moisture.

Puhakka et al (2000) studied reparation methods to reduce ammonia concentrations in ‘ammonia problem’ buildings.

They found the highest indoor air concentrations during preparation and very high levels inside the floor structures up

to 15370 µg/m

. The reparation of floor did not reduce ammonia concentrations in two months after the reparation, but

in other studies indoor concentrations started to decline in 6–8 months after the reparation.

Residential indoor concentrations of ammonia in homes with and without known indoor air quality (IAQ) problems

have been monitored in Finland (VTT 2003). Average indoor concentrations of ammonia in homes with and without

IAQ problems were 30 µg/m

and 24 µg/m

, respectively. Especially, if the building materials became wet due to water

damage or poor ventilation of the building structures, elevated ammonia concentrations were usually detected in indoor

air. Casein-containing materials were identified a source of ammonia in these cases (Saarela, 2003). Cumulative

distributions of ammonia in homes with and without IAQ problems in Finland are presented in Figure 3.1. Short time

mean concentrations of ammonia are summarised in Table 3.1.

100

0 10203040506070

Indoor concentration (µg/m

)

Cumulative frequency (%)

normal IAQ

IAQ problems

Figure 3.1. Cumulative distributions of residential indoor concentrations of ammonia in Finland in homes without and

with indoor air quality (IAQ) problems (VTT 2003).

The INDEX project Final report

216

Table 3.1. Short time ammonia concentrations related to specific microenvironments or emission sources.

Averaging time Reference

Mrange

Problem buildings before floor repairs 27 - 79 Puhakka et al 2000

during repairs 46 - 237

2 months after repairs 29 - 99

inside the floor structure 3810 - 15370

Dwellings empty dwellings 100-min 21 5 - 49 Saarinen et al 2002

inhabited dwellings 100-min 9 - 90

New apartments 2-hour 21 - 42 Tuomainen and Pirinen 2002

Toilets with septic tank 10-minute 1244 Hyung-Suk and Byung-Kee 1993

traditional vault system 7395

Homes kerosene heaters 8 -12-hour 11 Leaderer et al 1993

gas range 8 -12-hour 8.5

no gas appliances 8 -12-hour 7.1

Archives of the university library Helsinki downtown 4-12 Raiala et al 1993

Mikkeli < 5

AM = arithmetic mean

Concentration (µg/m

)

Environment or

emission source

4. Toxicokinetics

Absorption

At low concentrations, inhaled ammonia dissolves in the mucous fluid lining of the upper respiratory tract and little

reaches the lower airways. At ammonia levels associated with ambient air (i.e., 1 - 200 µg/m

), very little, if any, is

absorbed through the lungs.

Experiments with volunteers show that ammonia, regardless of its tested concentration in air (range = 40–350 mg/m

is almost completely retained in the nasal mucosa (83–92%) during short-term exposure, i.e., up to 120 seconds

(Landahl and Herrmann 1950). However, longer-term exposure (10–27 minutes) to a concentration of 350 mg/m

resulted in lower retention (4–30%), with 244-279 mg/m

eliminated in expired air by the end of the exposure period

(Silverman et al. 1949), suggesting an adaptive capability or saturation of the absorptive process. Nasal and pharyngeal

irritation, but not tracheal irritation, suggests that ammonia is retained in the upper respiratory tract. Unchanged levels

of blood-urea-nitrogen (BUN), non-protein nitrogen, urinary-urea, and urinary-ammonia are evidence of low absorption

into the blood.

Distribution

Absorption data from human inhalation exposure suggest that only small amounts of ammonia are absorbed into the

systemic circulation (Silverman et al. 1949; WHO 1986). Toxic effects reported from inhalation exposure suggest local

damage, or changes resulting from necrotic tissue degradation, rather than the presence of elevated levels of NH

, per

se, in tissues other than the respiratory/pharyngeal tissues.

Information on the distribution of endogenously-produced ammonia suggests that any NH

absorbed through

inhalation would be distributed to all body compartments via the blood, where it would be used in protein synthesis or

as a buffer, and that excess levels would be reduced to normal by urinary excretion, or converted by the liver to

glutamine and urea. If present in quantities that overtax these organs, NH

is distributed to other tissues and is known

to be detoxified in the brain (Takagaki et al. 1961; Warren and Schenker 1964).

Metabolism and elimination

Quantitative data on human metabolism of exogenously introduced ammonia were not located in the available

literature.

The INDEX project Final report

217

Ammonia and ammonium ion are metabolized to urea and glutamine mainly in the liver (Fürst et al., 1969; Pitts,1971).

However, it can be rapidly converted to glutamine in the brain and other tissues as well (Takagaki et al. 1961; Warren

and Schenker 1964).

Studies using low levels of ammonia show that inhaled ammonia is temporarily dissolved in the mucus of the upper

respiratory tract, and then a high percentage of it is released back into the expired air.

The quantitative difference between inspired and expired ammonia suggests that small amounts are absorbed across the

nasopharyngeal membranes into the systemic circulation. Absorbed ammonia is excreted by the kidneys as urea and

urinary ammonium compounds (Gay et al. 1969; Pitts 1971; Richards et al. 1975; Summerskill and Wolpert 1970), as

urea in feces (Richards et al. 1975), and as components of sweat (Guyton 1981; Wands 1981), but quantitative data are

lacking. Toxic levels do not develop as a result of chronic inhalation exposure because the body has multiple effective

mechanisms for detoxifying and excreting it.

5. Health effects

Effects of short-term exposure

The available human and animal data provide strong evidence that acute-duration exposure to ammonia can result in

site-of-contact lesions primarily of the eyes and the respiratory tract. Even fairly low airborne concentrations (35

mg/m

) of ammonia produce rapid onset of eye, nose, and throat irritation, coughing, and narrowing of the bronchi.

More severe clinical signs include immediate narrowing of the throat and swelling, causing upper airway obstruction

and accumulation of fluid in the lungs. This may result in low blood oxygen levels and an altered mental status.

Mucosal burns to the tracheobronchial tree can also occur. Children may be more vulnerable to corrosive agents than

adults because of the smaller diameter of their airways.

The eye is especially sensitive to alkali burns. Ammonia combines with moisture in the eyes and mucous membranes to

form ammonium hydroxide. Ammonium hydroxide causes saponification and liquefaction of the exposed, moist

epithelial surfaces of the eye and can easily penetrate the cornea and damage the iris and the lens (CCOHS, 1988; Way

et al., 1992). Damage to the iris may eventually lead to cataracts (CCOHS, 1988).

Silverman and coworkers (1949) exposed 7 volunteers to 348 mg/m³ (500 ppm) ammonia for 30 minutes using an oral-

nasal mask. Symptoms due to ammonia inhalation varied widely among the 7 subjects. All seven subjects experienced

upper respiratory irritation, which was graded as severe in 2 subjects. Only 2 subjects were able to continue nasal

breathing throughout the 30 minute exposure. Reactions included irritation of the nose and throat, hypoesthesia of the

exposed skin, and lacrimation. In two subjects, the nasopharyngeal irritation persisted for 24 hours after the exposure.

One of the 7 subjects was only exposed to ammonia for 15 minutes rather than the full 30 minutes. The reason for this

deviation in the exposure regimen was not given. In a previous experiment, brief exposure to 696 mg/m³ (1000 ppm)

reportedly resulted in immediate coughing in human subjects.

Ferguson and coworkers (1977) used six human subjects to demonstrate that a tolerance to ammonia exposure of 70

mg/m³ (100 ppm) can be developed with a two-to-three week inurement period during which volunteers were exposed

to lesser concentrations. The results tended to support the belief that persons with no recent history of ammonia

exposure are more sensitive to the irritating effects than those who are acclimated to the noxious gas.

Verberk (1977) exposed sixteen volunteers (8 science faculty with knowledge of the effects of ammonia and 8 non-

science university students not familiar with ammonia health effects) were exposed, 4 at a time, to 35, 56, 77, 98 mg/m³

(50, 80, 110, and 140 ppm) ammonia for 2 hours. Each group was exposed to each exposure level with 1 week in

between exposures. Immediately before and after exposure, respiratory function tests (vital capacity [VC], forced

expiratory volume in the first second [FEV

1], and forced inspiratory volume in the first second [FIV1]) were done.

During exposure, each participant recorded subjective effect levels for smell, taste, irritation of eyes, irritation of nose,

irritation of throat, irritation of breast, urge to cough, headache, and general discomfort. The scale used was: 0=no

sensation, 1=just perceptible, 2=distinctly perceptible, 3=nuisance, 4=offensive, and 5=unbearable. A (+) or (–) could

be used to interpolate between the levels. A few weeks after the experiments, the histamine threshold was determined

for 13 of the 16 volunteers as a measure of pre-existing non-specific reactivity of the airways to exogenous stimuli.

None of the participants was hypersusceptible to nonspecific irritants. No participant had a decrease of more than 10%

of pre-exposure values for VC, FEV

1, or FIV1. There was a difference between the science faculty group (experts) and

the students for the subjective scoring. Students consistently scored higher for smell and there was little increase in

score with concentration. Score for irritation of the eyes increased with concentration and there was no difference

The INDEX project Final report

218

between groups. Irritation of the throat had a sharp increase in score with concentration and scores were higher for

students. All students left the exposure chamber between 0.5 and 1.25 hours in the 98 mg/m³ exposure because of

severe irritation. Scores for urge to cough and general discomfort were low in the expert group, but increased with

concentration in the student group. All students left the chamber before 2 hours of exposure to 98 mg/m³.

At the end of the initial 30 minutes of the 2-hour exposure period, nuisance level smell, eyes, nose, or throat irritation,

or cough urge were reported by 7 of 16 (44%), 9 of 16 (56%), 12 of 16 (75%), or 15 of 16 (94%) individuals at

concentrations of 35, 56, 77, 98 mg/m³, respectively.

MacEwen et al. (1970) exposed groups of 5 and 6 human subjects to respective ammonia concentrations of 21 and 35

mg/m³ (30 and 50 ppm). The volunteers subjectively rated irritation for the 10-minute exposures. No moderate or higher

irritation was discerned by the group at the lower exposure level; however, 4 of the 6 subjects rated the 10 minute

exposure at 35 mg/m

as causing moderate irritation.

The Industrial Bio-Test Laboratories (1973) evaluated ten human subjects for the irritation threshold of ammonia from

exposures to ammonia gas at four different concentrations: 22, 35, 50, and 93 mg/m³. Irritation was taken to be any

annoyance to the eyes, nose, mouth, throat, or chest which persisted throughout the 5-minute exposure period. At 50

mg/m³ three subjects experienced eye irritation, two had nasal irritation, and three had throat irritation. At 93 mg/m³,

five of the ten subjects experienced lacrimation and eye irritation, seven complained of nasal irritation, eight had throat

irritation, and one experienced chest irritation. The authors only used 5-minute exposure durations; and it is possible

that irritation symptoms could have developed with longer exposure durations at the lower exposures. The authors

discounted the significance of nasal dryness reported at the two lowest levels.

Douglas and Coe (1987) determined a lachrymatory threshold of 38 mg/m³ for ammonia following approximately 15

second exposures of volunteers via tight-fitting goggles. The threshold for bronchoconstriction, determined as a 20%

increase in airway resistance, was slightly higher at 59 mg/m³ following 10 breaths of ammonia via mouthpiece.

Tolerance appears to develop with repeated exposure (Sekizawa and Tsubone 1994; Verberk 1977). Thus, subjects

exposed to 70, but not 35 mg/m

ammonia 6 hours/day, 5 days/week for 6 weeks experienced nose and throat irritation

only during the first week (Ferguson et al. 1977).

Estimates of odor thresholds for ammonia vary from 0.03-72 mg/m³ (Ferguson et al., 1977; Henderson and Haggard,

1943; Ruth, 1986). Near the odor threshold, persons exposed to ammonia can experience annoyance and believe the

odor to be a nuisance. Exposure to ammonia may result in an exacerbation of preexisting asthma. Shim and Williams

(1986) surveyed 60 patients with a history of asthma worsened by certain odors. Nearly 80% of these patients claimed

to have an exacerbation of asthma following exposure to household cleaners containing ammonia.

Summary of short-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

14.0

1h-ADJ-MLE

9.5

BC05

eye and respiratory irritation volunteers,

1h-adjusted

Benchmark approach

Industrial Biotest

Laboratories, 1973;

MacEwen et al., 1970;

Silverman et al., 1949;

Verberk, 1977

OEHHA 1999;

REL: 3.2

EXP

irritation to nose and throat; urge to cough volunteers, 2h Verberk, 1977 ATSDR 2002;

MRL: 1.2

No study with asthmatics!

Final statement (UNIMI): Human LOAEL: 35 mg/m³

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

The INDEX project Final report

219

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of Acute Reference Exposure Level protective against mild adverse effects (for a 1-hour exposure)

Study Industrial Biotest Laboratories, 1973; MacEwen et al., 1970;

Silverman et al., 1949; Verberk, 1977

Study population humans

Exposure method inhalation

Critical effects eye and respiratory irritation

LOAEL varied (see the text)

NOAEL varied (see the text)

Exposure duration varied (see the text)

Extrapolated 1 hour concentration 9.5 mg/m

(BC05)

LOAEL uncertainty factor not needed in BC approach

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 3

Cumulative uncertainty factor 3

Reference Exposure Level 4.5 ppm (3.2 mg/m³; 3200 µg/m³)

The exposure concentrations from the 4 studies were adjusted to 1-hour durations using the formula C

x T = K (Table

5.2). The value for the exponent n was empirically derived from the preceding data sets. The value of n (in the formula

x T = K) was sequentially varied for the log-normal probit relationship analysis. Using a chi-square analysis, a value

of n = 4.6 was found to be the best fit.

The REL was calculated by a benchmark concentration (BC) approach using a log-normal probit analysis (Crump and

Howe, 1983; Crump. 1984). The 95% lower confidence limit of the concentration expected to produce a response rate

of 5% is defined as the BC

05. The maximum likelihood estimate for a 5% response was 14.0 mg/m

(20.1 ppm) and the

95% LCL on this value (BC

05) for ammonia from this analysis was 9.5 mg/m

(13.6 ppm).

An uncertainty factor (UF) of 3 was used to account for intraspecies variation in the human population.

Table 5.2: Ammonia, Human Irritation, 60 Minute Exposures (adjusted), mg/m

Study Study

concentration

Exposure time 60 min

adjusted

concentration

Response

Industrial Biotest Laboratories, 1973 22 5 13 0/10

MacEwen et al., 1970 21 10 14 0/5

Industrial Biotest Laboratories, 1973 35 5 20 0/10

MacEwen et al., 1970 35 10 24 4/6

Industrial Biotest Laboratories, 1973 50 5 29 3/10

Verberk, 1977 35 120 30 7/16

Verberk, 1977 56 120 48 9/16

Industrial Biotest Laboratories, 1973 93 5 54 8/10

Verberk, 1977 77 60 66 12/16

Verberk, 1977 98 60 84 15/16

Silverman et al., 1949 348 30 299 7/7

Table adapted from: Verberk, 1977; Industrial Biotest Laboratories, 1973; MacEwen et al., 1970; and Silverman et al., 1949. The two

lowest concentrations were combined for the log-probit analysis since this improved the fit of the data.

Response rate MLE (mg/m

) 95% LCL

(mg/m

)

1% 9.3 5.4

5% 14.0 9.5 (BC

)

The INDEX project Final report

220

Graphic Representation of Benchmark Concentration Determination

ATSDR (1999)

Agency for Toxic Substances and Disease Registry

Derivation of a Minimal Risk Level (MRL):

An acute inhalation MRL of 1.2 mg/m³ (1.7 ppm) was derived from the Verberk (1977) study. Dose and end point used

for MRL derivation are 35 mg/m³ (50 ppm) for mild irritation to the eyes, nose, and throat in humans exposed to

ammonia gas for 2 hours. An Uncertainty Factors of 30 was used in MRL derivation: 3 for use of a minimal LOAEL

and 10 for human variability

The MRL is supported by other observations of respiratory effects associated with acute- and intermediate-duration

exposure including transient irritation of the nose and throat of humans exposed to 70 mg/m

(Ferguson et al. 1977);

nasal discharge in rats at 262 mg/m

(Coon et al. 1970); nasal lesions in rats at 104 mg/m

(Broderson et al. 1976); and

nasal inflammation and lesions in rats at 348 mg/m

(Richard et al. 1978). A study of piggerie workers exposed to a

mean level of 5.5 mg/m

ammonia measured lung function change over a workshift; a small but borderline significant

decrease in lung function was noted (Heederik et al. 1990). This was not used as a basis for MRL derivation because the

workers were also exposed to other potential respiratory toxicants (dust and endotoxins).

Effects of long-term exposure

Several studies have examined the relationship between chronic exposure to ammonia and respiratory effects. Studies of

farmers working in enclosed livestock facilities provide evidence that ammonia may contribute to transient respiratory

distress (Choudat et al. 1994; Cormier et al. 2000; Donham et al. 1995, 2000; Heederik et al. 1990, 1991; Melbostad

and Eduard 2001; Reynolds et al. 1996; Vogelzang et al. 1997, 2000); however, co-exposure to total dust, respirable

dust, carbon dioxide, total endotoxins, respirable endotoxins, fungi, bacteria, and/or molds complicates the

interpretation of these studies.

The INDEX project Final report

221

Comparisons were made between 52 workers and 31 control subjects in a soda ash plant for pulmonary function and

eye, skin and respiratory symptomatology (Holness et al., 1989). The pulmonary function tests included FVC (forced

vital capacity – the total amount of air the subject can expel during a forced expiration), FEV

1 (forced expiratory

volume in one second), FEF50 (forced expiratory flow rate at 50% of the FVC) and FEF75 (forced expiratory flow rate

at 75% of the FVC). Age, height, and pack-years smoked were treated as covariates for the comparisons. The workers

were exposed on average for 12.2 years to mean (time-weighted average) ammonia concentrations of 6.4 ± 1.0 mg/m

(9.2 ± 1.4 ppm), while controls were exposed to 0.21 ± 0.7 mg/m

(0.3 ± 0.1 ppm). No differences in any endpoints

(respiratory or cutaneous symptoms, sense of smell, baseline lung function, or change in lung function over a work shift

at the beginning and end of a workweek) were reported between the exposed and control groups.

Groups of human volunteers were exposed to 0, 17, 35, or 70 mg/m

(25, 50, or 100 ppm) ammonia 5 days/week for 2,

4, or 6 hours/day, respectively, for 6 weeks (Ferguson et al., 1977). Another group of 2 volunteers was exposed to 35

mg/m

ammonia for 6 hours/day for 6 weeks. Pulmonary function tests (respiration rate, FVC and FEVl) were measured

in addition to subjective complaints of irritation of the eyes and respiratory tract. The difficulty experienced in

performing simple cognitive tasks was also measured, as was pulse rate. There were reports of transient irritation of the

nose and throat at 35 or 70 mg/m

. Acclimation to eye, nose, and throat irritation was seen after two to three weeks (in

addition to the short-term subjective adaption). No significant differences between subjects or controls on common

biological indicators, in physical exams, or in performance of normal job duties were found. After acclimation,

continuous exposure to 70 mg/m

, with occasional excursions to 139 mg/m

, was easily tolerated and had no observed

effect on general health.

Broderson et al. (1976) exposed groups of F344 rats (6/sex/dose) continuously to 17, 35, 104 or 174 mg/m

ammonia

(HEC = 1.9, 3.7, 11.2 or 18.6 mg/m

, respectively) for 7 days prior to inoculation with Mycoplasma pulmonis and from

28-42 days following M. pulmonis exposure. Each treatment group had a corresponding control group exposed only to

background ammonia and inoculated with M. pulmonis in order to produce murine respiratory mycoplasmosis (MRM).

The following parameters were used to assess toxicity: clinical observations and histopathological examination of nasal

passages, middle ear, trachea, lungs, liver and kidneys. All levels of ammonia, whether produced naturally or derived

from a purified source, significantly increased the severity of rhinitis, otitis media, tracheitis and pneumonia

characteristic of M. pulmonis. Furthermore, there was a significant concentration response between observed

respiratory lesions and increasing environmental ammonia concentration for gross and microscopic lesions. All lesions

observed were characteristic of MRM. Gross bronchiectasis and/or pulmonary abscesses and the extent of gross

atelectasis and consolidation was consistently more prevalent in exposed animals at all concentrations than in their

corresponding controls. The severity of the microscopic lesions in the nasal passages, middle ears, tracheas and lungs

was significantly greater in all exposed groups compared with controls. Increasing ammonia concentration was not

associated with an increasing frequency of M. pulmonis isolations. Additionally, rats not exposed to M. pulmonis and

exposed to ammonia at 174 mg/m

developed nasal lesions (epithelial thickening and epithelial hyperplasia) unlike

those observed in inoculated rats.

The growth of bacteria in the lungs and nasal passages, and the concentration of serum immunoglobulin were

significantly increased in rats exposed to 70 mg/m

(100 ppm) ammonia over that seen in control rats (Schoeb et al.,

1982).

Guinea pigs (10/group) and mice (20/group) were continuously exposed to 14 mg/m

(20 ppm) ammonia for up to 6

weeks (Anderson et al., 1964). Separate groups of 6 guinea pigs and 21 chickens were exposed to 35 and 14 mg/m

(50

and 20 ppm) ammonia for up to 6 and 12 weeks, respectively. All species displayed pulmonary edema, congestion, and

hemorrhage after 6 weeks exposure, whereas no effects were seen after only 2 weeks. Guinea pigs exposed to 35 mg/m

ammonia for 6 weeks exhibited enlarged and congested spleens, congested livers and lungs, and pulmonary edema.

Chickens exposed to 140 mg/m

(200 ppm) for 17-21 days showed liver congestion and slight clouding of the cornea.

Anderson and associates also showed that a 72-hour exposure to 14 mg/m

ammonia significantly increased the

infection rate of chickens exposed to Newcastle disease virus, while the same effect was observed in chickens exposed

to 35 mg/m

for just 48 hours.

Coon et al. (1970) exposed groups of rats (as well as guinea pigs, rabbits, dogs, and monkeys) continuously to ammonia

concentrations ranging from 40 to 470 mg/m

. There were no signs of toxicity in 15 rats exposed continuously to 40

mg/m

for 114 days or in 48 rats exposed continuously to 127 mg/m

for 90 days. Among 49 rats exposed continuously

to 262 mg/m

for 90 days, 25% had mild nasal discharge. At 455 mg/m

50 of 51 rats died. Thus 127 mg/m3 (179 ppm)

is a subchronic NOAEL for upper respiratory effects in rats. Coon et al. (1970) also found no lung effects in 15 guinea

pigs exposed continuously to 40 mg/m

(28 ppm) ammonia for 114 days.

No chronic-duration oral or dermal data were located. Studies by these routes of exposure would provide useful

information on the identification of target organs especially after low-dose exposure.

The INDEX project Final report

222

Summary of long-term exposure effect levels

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

6.4

Study

2.3

ADJ

Pulmonary function, eye, skin, and respiratory

symptoms of irritation

occupational, 12.2y

Holness et al., 1989

OEHHA 1999;

REL: 0.2

EPA-IRIS 1991;

RfC: 0.1

EXP

1.9

HEC

Increased severity of

rhinitis and pneumonia

with respiratory

lesions

rats, subchronic Broderson et al., 1976 OEHHA 1999

EPA-IRIS 1991

8.7

Study

2.2

ADJ

Pulmonary function, eye, skin, and respiratory

symptoms of irritation, sense of smell

occupational, 12.2y Holness et al., 1989 ATSDR 2002;

MRL: 0.2

Final statement (UNIMI) : Human NOAEL: 2 mg/m³

Study

average concentration ;

EXP

experimental concentration ;

ADJ

concentration adjusted from an intermittent to a continuous exposure ;

1h-ADJ

concentration adjusted to 1-hour exposure duration ;

HEC

human equivalent concentration ;

MLE

maximum likelihood estimate for 5%

response

;

BC05

95% lower confidence limit of the concentration expected to produce a response rate of 5% (Benchmark concentration

approach) ;

STAT

lowest statistically significant effect concentration

OEHHA (1999)

Office of Environmental Health Hazard Assessment of the Californian Environmental Protection Agency

Derivation of Chronic Reference Exposure Level

Study Holness et al., 1989 (supported by Broderson et al., 1976)

Study population 52 workers; 31 controls

Exposure method Occupational inhalation

Critical effects Pulmonary function, eye, skin, and respiratory symptoms of irritation

LOAEL 25 ppm (Broderson et al., 1976) (rats)

NOAEL 9.2 ppm (Holness et al., 1989)

Exposure continuity 8 hours/day (10 m

3/day occupational inhalation rate), 5 days/week

Exposure duration 12.2 years

Average occupational exposure 3 ppm for NOAEL group (9.2 x 10/20 x 5/7)

Human equivalent concentration 3 ppm for NOAEL group

LOAEL uncertainty factor 1

Subchronic uncertainty factor 1

Interspecies uncertainty factor 1

Intraspecies uncertainty factor 10

Cumulative uncertainty factor 10

Inhalation reference exposure level 0.3 ppm ( 300 ppb; 0.2 mg/m

; 200 µg/m

)

The Holness et al. (1989) study was selected because it was a chronic human study and was published in a respected,

peer-reviewed journal. It is also the only chronic study available. The U.S.EPA (1995) based its RfC of 100 µg/m

the same study but included a Modifying Factor (MF) of 3 for database deficiencies. The criteria for use of modifying

factors are not well specified by U.S. EPA. Such modifying factors were not used by OEHHA.

For comparison with the proposed REL of 200 µg/m

based on human data, OEHHA estimated RELs from 2 animal

studies. (1) Anderson et al. (1964) exposed guinea pigs continuously to 35 mg/m

(50 ppm) ammonia for 6 weeks and

observed pulmonary edema. Use of an RGDR of 0.86 and a cumulative uncertainty factor of 3000 (10 for use of a

LOAEL, 10 for subchronic, 3 for interspecies, and 10 for intraspecies) resulted in a REL of 10 µg/m

. Staff note that the

nearly maximal total uncertainty factor of 3000 was used in this estimation. (2) Coon et al. (1970) exposed rats

continuously to 127 mg/m

ammonia for 90 days and saw no signs of toxicity. Use of an RGDR(ET) of 0.16 for nasal

effects (observed in rats exposed to higher levels of ammonia in Broderson et al. (1976) and a cumulative uncertainty

factor of 100 (3 for subchronic, 3 for interspecies, and 10 for intraspecies) resulted in a REL of 200 µg/m

Data Strengths and Limitations for Development of the REL: Significant strengths in the ammonia REL include (1) the

availability of long-term human inhalation exposure data (Holness et al., 1989), (2) the demonstration of consistent

effects in experimentally exposed human volunteers following short-term exposures (Ferguson et al., 1977), and (3)

reasonable consistency with animal data (Coon et al., 1970).

The INDEX project Final report

223

Major areas of uncertainty are: (1) the lack of a NOAEL and LOAEL in a single study, (2) a lack of animal data with

chronic exposure and histopathological analyses, and (3) difficulties in estimated human occupational exposures. The

overall database for this common chemical is limited.

ATSDR (2002)

Agency for Toxic Substances and Disease Registry

Derivation of the Minimal Risk Level (MRL):

In the Holness et al. (1989) study, the cohort was also divided into groups that were exposed to low (<4.35 mg/m

medium (4.35-8.7 mg/m

), and high (>8.7 mg/m

) ammonia levels and analyzed for change in lung function. Analysis

was performed using each worker’s personal exposure and his change in lung function over the workweek. Differences

due to number of years of ammonia exposure was also assessed. No statistically significant differences were seen

between the level of personal exposure and change in lung function or in lung function between low, medium, and high

exposed groups. No association was evident between years of exposure and lung function changes. The dose and end

point used for MRL derivation are: 8.7 mg/m

(NOAEL) for sense of smell, prevalence of respiratory symptoms

(cough, bronchitis, wheeze, dyspnea, and others), eye and throat irritation, and lung function parameters (FVC, FEV1,

FEV1/FVC, FEF50, and FEF75).

The NOAEL was adjusted for continuous exposure as follows: 8.7 mg/m

x 8.4/24 hours x 5/7 days = 2.2 mg/m

. An

Uncertainty Factors of 10 was used in MRL derivation for human variability: MRL = 0.2 mg/m

(0.3 ppm).

The MRL is supported by other observations of respiratory effects associated with chronic-duration exposure including

an association between exposure to pollutants, including ammonia, in livestock confinement buildings and an increase

in respiratory symptoms (such as bronchial reactivity/hyperresponsiveness, inflammation, cough, wheezing, or

shortness of breath) and/or a decrease in lung function (such as forced expiratory volume in the first second [FEV1.0],

maximum expiratory flow rates [MEF50 and MEF75], and maximal mid-expiratory flow rate [MMEF]) in farmers

exposed to ammonia levels of 1.6-14.4 mg/m

(Choudat et al. 1994; Cormier et al. 2000; Donham et al. 1995, 2000;

Heederik et al. 1990; Reynolds et al. 1996; Vogelzang et al. 1997, 2000). The farmers were also exposed to other

possible respiratory toxins, such as dust and endotoxins. A cross-sectional study of male workers at two fertilizer

factories in Saudi Arabia showed a significant association between exposure to ammonia gas and respiratory symptoms

and bronchial asthma (Ballal et al. 1998). No continuous exposure levels could be calculated for these workers because

the number of days worked per week was not provided.

USEPA-IRIS (1991)

Integrated Risk Information System

Determination of the Reference Concentration for Chronic Inhalation Exposure (RfC)

Critical effect Exposures* UF MF RfC

Lack of evidence of

decreased pulmonary

function or changes

in subjective

syptomatology

Occupational Study

Holness et al., 1989

NOAEL: 6.4 mg/cu.m (9.2 ppm)

NOAEL(ADJ): 2.3 mg/cu.m

NOAEL(HEC): 2.3 mg/cu.m

LOAEL: None

30 1 0.1 mg/m

Increased severity of

rhinitis and pneumonia

with respiratory

lesions

Rat Subchronic

Inhalation Study

Broderson et al., 1976

NOAEL: None

LOAEL: 17.4 mg/cu.m (25 ppm)

LOAEL(ADJ): 17.4 mg/cu.m

LOAEL(HEC): 1.9 mg/cu.m

*Conversion Factors: MW = 17.03 Holness et al., 1989: Assuming 25C and 760 mm Hg, NOAEL (mg/cu.m) = 9.2 ppm x 17.03/24.45 =

6.4 mg/cu.m. The NOAEL is based on an 8-hour TWA occupational exposure. MVho = 10 cu.m/day, MVh = 20 cu.m/day. NOAEL(ADJ)

= 6.4 mg/cu.m x (MVho/MVh) x 5 days/7 days = 2.3 mg/cu.m.

Broderson et al., 1976: Assuming 25C and 760 mm Hg, the LOAEL (mg/cu.m) = 25 ppm x 17.03/24.45 = 17.4 mg/cu.m. The

LOAEL(HEC) was calculated for a gas:respiratory effect in the ExtraThoracic region. MVa = 0.14 cu.m/day, MVh = 20 cu.m/day, Sa(ET)

= 11.6 sq. cm., Sh(ET) = 177 sq. cm. RGDR(ET) = (MVa/Sa) / (MVh/Sh) = 0.1068. NOAEL(HEC) = 17.4 x RGDR = 1.9 mg/cu.m.

The INDEX project Final report

224

The use of Holness et al. (1989) as the principal study can only be supported in the context of the data array. It is not

surprising that no effects were seen on screening spirometry since the exposure levels were low. Comparing the 9.2

TWA of Holness et al. (1989) with other data on the respiratory effects of ammonia, a trend is observed that at lower

concentrations the extrathoracic region of the respiratory system is affected due to the chemical's solubility and

reactivity; while at higher concentrations, the lower part of the respiratory system is involved in both experimental

animals (Dahlman, 1956; Gamble and Clough, 1976) and humans (Flury et al., 1983). Thus, no effects were observed in

the lower respiratory system as reflected by pulmonary function. Pulmonary function may not be a particularly sensitive

test because exposure to this type of agent at low concentrations is not expected to result in significant exposure of the

lower respiratory region. No objective investigation of the workers' nasal epithelium was performed and the complaint

of exacerbated upper respiratory symptoms suggests sensory irritation and supports the extrathoracic region as the

critical region for an effect. The possibility of selection bias against atopic predispositions in the population is suggested

by the significantly lower prevalence of hay fever in the exposed versus control cohort. Thus, there is a concentration-

response in the extrathoracic region in experimental animals beginning at a LOAEL at essentially the same HEC as the

NOAEL in Holness et al. (1989) and the NOAEL may be based on a less sensitive endpoint. Also the apparent

discrepancy of a lower LOAEL(HEC) from Broderson et al. (1976) and the identified NOAEL(HEC) of the Holness et

al. (1989) study may be the result of differences in air flow patterns since rats are obligate nose- breathers and humans

breathe oronasally. The use of the NOAEL from Holness et al. (1989) can be supported as marginal in this context due

to the symptomatology complaints and because human data engenders less uncertainty than extrapolation from the

experimental animal data.

Confidence in the Inhalation RfC: Confidence in the principal study is medium. Although a relatively small sample size

(males only) was studied and a free standing NOAEL was determined, mild extrathoracic effects were observed in rats

near the same HEC as reported in the Holness study. Additional human subchronic and acute studies support the

NOAEL. Confidence in the data base is medium to high. Although developmental, reproductive or chronic toxicity

following ammonia exposure has not been adequately tested, pharmacokinetic data suggests systemic distribution at the

HEC level is unlikely. Reflecting medium confidence in the principal studies and medium to high confidence in the data

base, confidence in the RfD is medium.

Reproductive, mutagenic and carcinogenic effects

Toxicological effects, such as mutagenic, reproductive, or carcinogenic effects, would not be expected from exposures

insufficient to cause local effects. There is no evidence that ammonia is carcinogenic, though it can produce

inflammatory lesions of the colon and cellular proliferation, which could increase susceptibility to malignant change.

There was no evidence that ammonia was responsible for the increased incidence of tumours with increased dietary

protein intake. Ammonia did not either cause tumours or increase the spontaneous incidence of tumours in life-time

studies on mice.

Interactions with other chemicals

Exposure to substances that would increase the pH of exposed tissues could be expected to enhance the alkalotic effects

of ammonia, and vice versa.

Data regarding exposure to mixtures of atmospheric contaminants indicate that, contrary to what might be expected,

increased carbon dioxide concentration (up to 5% in air) does not alter the hyperventilatory rate induced by

hyperammonemia in dogs (Herrera and Kazemi 1980). Ammonia in expired air may neutralize inhaled acid aerosols

(EPA 1979; Larson et al. 1980; Utell et al. 1989).

Odour perception

Odour is characterized as sharp, pungent and intensely irritating.

Source: Devos et al., (1990)

Odour threshold: 1 mg/m

Source: Amoore and Hautala (1983)

Odor threshold: 17 mg/m

(25 ppm)

The INDEX project Final report

225

Source: Leonardos et al. (1969)

Recognition threshold: 33 mg/m

(47 ppm)

The considerable variability in threshold data prompted work by Leonardos et al. (1969), who used a standardized

procedure to determine recognition thresholds rather than detection thresholds for 53 chemicals. The odour threshold

was defined as the first concentration at which all four panel members (trained odour analysts) were able to recognize

the characteristic odour of the chemical. The panel tested only one chemical per day. Concentrations examined were

multiples by 10 of 0.7, 1.47, and 3.3 mg/m

(1, 2.1, and 4.6 ppm). The recognition threshold for the odour of ammonia

was 32.6 mg/m

(46.8 ppm) (WHO, 1986).

Source: AIHA (1989)

Reported odour threshold values range from 0.03 to 37.5 mg/m

(0.041 to 53 ppm) with a geometric mean of 11.8

mg/m

(17 ppm)

Source: WHO (1986)

Estimates of odor thresholds for ammonia vary from 0.04-103 ppm (0.03-73 mg/m³) (Ferguson et al., 1977; Henderson

and Haggard, 1943; Ruth, 1986). Near the odor threshold, persons exposed to ammonia can experience annoyance and

believe the odor to be a nuisance.

Summary of Ammonia Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Endogenous ammonia (dietary proteins) used for protein synthesis and as a buffer,

rapid conversion of excesses to glutamine/urea (liver) and urinary excretion.

Toxicokinetics: At ambient air levels (<350 mg/m

), almost completely (83-92%) dissolved (NH

OH) in the upper

respiratory tract (nasal and pharyngeal, but not tracheal irritation), reduced absorption at prolonged exposure (>10min,

>350 mg/m

) due to saturation. Low absorption into the blood (systemic circulation), elimination mainly through

expired air at the end of exposure. Only small proportion absorbed across the nasopharyngeal membranes.

Health effect levels of short- and long-term exposure

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

14.0

1h-ADJ-MLE

9.5

BC05

eye and respiratory irritation volunteers,

1h-adjusted

Benchmark approach

Industrial Biotest

Laboratories, 1973;

MacEwen et al., 1970;

Silverman et al., 1949;

Verberk, 1977

OEHHA 1999;

REL: 3.2

EXP

irritation to nose and throat; urge to cough volunteers, 2h Verberk, 1977 ATSDR 2002;

MRL: 1.2

General remark: No study with asthmatics!

Long-term exposure

6.4

Study

2.3

ADJ

Pulmonary function, eye, skin, and respiratory

symptoms of irritation

occupational, 12.2y

Holness et al., 1989

OEHHA 1999;

REL: 0.2

EPA-IRIS 1991;

RfC: 0.1

EXP

1.9

HEC

Increased severity of rhinitis and pneumonia

with respiratory lesions

rats, subchronic Broderson et al., 1976 OEHHA 1999

EPA-IRIS 1991

8.7

Study

2.2

ADJ

Pulmonary function, eye, skin, and respiratory

symptoms of irritation, sense of smell

occupational, 12.2y Holness et al., 1989 ATSDR 2002;

MRL: 0.2

Carcinogenicity: No evidence in life-time studies on mice. Mutagenic or carcinogenic effects would not be expected

from exposures insufficient to cause local effects (inflammatory lesions and cell proliferation).

Genotoxicity: No evidence

Odour threshold: Range: 0.03-38 mg/m

, with geometric mean: 12 mg/m

(AIHA) and recognition: 33 mg/m

; 1

mg/m

(Devos); Coughing reflex and lachrymation starting from 35 mg/m

, immediate coughing at 700 mg/m

Susceptible population: Exposure to ammonia result in an exacerbation of preexisting asthma (48/60 asthmatics

claimed following exposure to household cleaners containing ammonia). Wide variability of symptoms depending on

recent history of ammonia exposure; tolerances up to 70 mg/m

after weekly inurement period.

Remarks: RD

: 211 mg/m

The INDEX project Final report

226

________________________________________________________________________________________________

6. Risk Characterization

Health hazard evaluation of short-term exposure

Toxicological endpoint: Irritation to nose and throat; urge to cough

Derivation of a limit of exposure (EL)

LOAEL

100

LOAEL

300

35 mg/m

3.5 mg/m

0.35 mg/m

0.1 mg/m

Human LOAEL ;

not considering a NOAEL (10);

intraspecies variability

(10) ;

susceptible population (3)

Odour threshold: Range: 0.03-38 mg/m

, with geometric mean: 12 mg/m

(AIHA) and recognition: 33 mg/m

; 1 mg/m

(Devos); Coughing reflex and lachrymation starting from 35 mg/m

, immediate coughing at 700 mg/m

Health hazard evaluation of long-term exposure

Toxicological endpoint: Pulmonary function, eye, skin, and respiratory symptoms of irritation

Relevance of EU-population exposure to Ammonia

Non population-based study

NOAEL

Margin of

safety (MOS)

Description (Study, Year) N 2 mg/m

0.2 mg/m

0.07 mg/m

(90

)

Helsinki (VTT 2003)

Homes without known IAQ problems

< < < 3.0 (1.8)

Helsinki (VTT 2003)

Homes with known IAQ problems

< < < 2.6 (1.3)

Occupational LOAEL, adjusted for continuous exposure;

intraspecies variability (10);

susceptible population;

< out of the evaluation range (i.e. <5% of the environments investigated)

Result

There is a lack of knowledge concerning indoor concentrations and exposures of ammonia. Exposure data are limited

on only one non population-based study describing concentrations of ammonia in Finnish homes with and without

known indoor air quality (IAQ) problems. In both cases measured concentrations were within the same order of

magnitude with both exposure limits here esteblished for short- and long-term effects (70 and 100 µg/m

, respectively)

relating on irritative effects and pulmonary funtions and taking into account the particular susceptibility of asthmatic

subjects. It is assumed that exposure concentrations in the order of the short-term EL could easily be attained during

domestic activities making use of ammonia containing household products.

The INDEX project Final report

227

Limonene

Synonyms: 1-methyl-4-(1-methylethenyl)cyclohexene

CAS Registry Numbers: 138-86-3

Molecular Formula: C

1. Compound identification

Limonene exists as two optical isomers, d-limonene (pleasing orange scent) and l-limonene (smells piney, turpentine

smell; mixture of both isomers: dipentene), and is used as a flavour and fragrance additive in food, household cleaning

products and perfumes or could be contained in solvents (e.g. turpentine).

Limonene is both a naturally occurring and a synthetic colourless mobile liquid, which is used in many food products,

for its characteristic lemon-like flavour and odour. It is also used as a solvent, wetting agent, in resins, and as a

monomer and copolymer. It may be emitted to indoor air also from cats and dogs repellent spray and shampoo, pet

sleeping quarters, household dwellings, indoor premises, and human clothing. Limonene emissions have been detected

in the industries such as extraction of pine gum, paper and pulp mills, plastics materials-synthetic resins, perfumes,

cosmetics and other toilet preparations, organic solvents and lubricating oils and greases (HSDB 2003).

Limonene may be emitted to household environments from furniture polishes and room fresheners. Occupational

exposure to limonene may occur by inhalation or dermal contact during its production, formulation, transport or use.

The main routes of human exposure to limonene in the general population is assessed to be inhalation of limonene and

ingestion of food in which it occurs naturally or to which it has been added as a flavour or fragrance (EPA 1994, HSDB

2003).

2. Physical and Chemical properties

Molecular weight (g/mol) 136.23

Melting point (°C) -74

Boiling point (°C) 176

Density (at 20 °C, 1 atm) 0.8402

Relative density (air =1) 4.7

Solubility: Miscible with alcohol and ether

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 5.654 µg/m

1 µg/m

= 0.177 ppb

Sources: Verschueren 2001, HSDB 2003

3. Indoor Air Exposure assessment

Contribution of inhalation exposure to total exposure

Food is the principal source of exposure to limonene (96%), the contribution from ambient air being approximately 4%

(WHO, 1998). d-limonene is generally recognized as safe in food by the Food and Drug Administration. Based on daily

US consumption of d-limonene per capita, the intake of d-limonene from food for the general population was estimated

to be 0.27 mg/kg body weight per day (Flavor and Extract Manufacturers Association, 1991). The intake of limonene

from indoor and outdoor air for the general population is estimated to be 10 and 0.1 µg/kg body weight per day,

The INDEX project Final report

228

respectively. This is based on the daily inhalation volume for adults of 22 m

, a mean body weight for males and

females of 64 kg, the assumption that 4 of 24 hours are spent outdoors (IPCS, 1994), and arithmetic mean limonene

levels in indoor and outdoor air of 0.04 and 0.002 mg/m

, respectively, in a study from Los Angeles (Wallace et al.,

1991).

Terpene/ozone reaction products

It has been proposed that reactions between unsaturated volatile compounds (e.g. limonene, α-pinene, styrene) and

ozone or hydroxyl (OH) radicals produce chemically reactive products more likely to be responsible for eye and airway

irritation than the chemically non-reactive VOCs usually measured indoors. Many of the known reaction products are

oxygenates (e.g. formaldehyde and other aldehydes, carboxylic acids, peroxides) which may cause irritation at low

concentrations or odor annoyance due to their usually low odour thresholds (Weschler and Shields, 1997; Wolkoff et

al., 1997; 2000). Ozone enters the indoor environment (<0.3 mg/m3) by infiltration of the outdoor air (<0.6 mg/m3) or

could be produced by photocopiers and certain air cleaners.

Indoor air and exposure concentrations

Residential indoor concentrations of limonene were clearly higher than respectively outdoor concentrations in all

regions of Europe based on EXPOLIS results (Table 4.0.1, Jantunen et al 1999). Average indoor concentrations were

lowest in Central Europe and highest in Athens (Figure 3.1, Annex 1). Personal exposures to limonene, ranging from 19

µg/m

to 56 µg/m

were lower than residential indoor concentrations (Jantunen et al 1999, Hoffman et al 2000).

Kostiainen et al (1995) reported clearly lower mean and median indoor concentrations of limonene in Helsinki, Finland,

14.2 µg/m

and 8.8 µg/m

respectively. Maroni et al (1995) reported typical limonene concentrations in residential

indoor environments in Europe and in the United States, being on average 30 µg/m

. Indoor concentrations measured in

offices in Milan, Italy were about half of the residential levels, being on average 17 µg/m

. Brown (2002) studied

limonene concentrations in new and established buildings. Limonene concentrations in a new building increased after

construction, being 12 µg/m

, 15 µg/m

, and 34 µg/m

2, 19, and 72 days after construction respectively. This

decreasing trend suggested the dominant source of limonene being human activities, possibly consumer products, rather

than building materials.

TEAM study carried out in Los Angeles, the USA showed a maximum 12-hour day time exposure of 990 µg/m

limonene (Wallace et al 1991). Respective residential indoor and outdoor concentrations showed concentrations of 230

µg/m

and 10 µg/m

. Maximum exposure concentration in California showed higher value than in European studies.

Maximum residential indoor concentration was in the range with European maximum levels.

WHO (1989) has assessed indoor air concentrations of VOCs in “typical homes” giving median and 90

percentile

concentrations of 15 µg/m

and 70 µg/m

for limonene. Brown et al (1994) calculated weighted geometric mean

(WAGM) indoor concentrations for several building types using mean values collected from the studies carried out

worldwide. Also they reported lower indoor values, WAGM of 21 µg/m

and 90

percentile of 85 µg/m

for limonene

than found in European population based studies.

Residential indoor air concentrations of limonene in European urban populations measured in the EXPOLIS study and

the National Survey in England are presented in Figure 3.1 and Figure 3.2. Respective exposure concentrations are

presented in Figure 3.3.

The INDEX project Final report

229

100

0 50 100 150 200 250

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of limonene in Athens (Ath, n= 42), Basel

(Bas, n= 47), Helsinki (Hel, n= 188), Milan (Mil, n= 41) Oxford (Oxf, n= 40) and Prague (Pra, n= 46) (EXPOLIS

2002).

100

0 1020304050

Indoor concentration (µg/m

)

Cumulative frequency (%)

Limonene 28-day

Figure 3.2. Cumulative frequency distribution of 28-day indoor air concentrations of limonene in UK (GM= 6.2 µg/m

max 308 µg/m

, n=796, Brown et al 2002).

The INDEX project Final report

230

100

0 50 100 150 200 250

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.3. Cumulative frequency distributions of 48-hour personal exposure concentrations of limonene in Athens

(Ath), Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean exposures of the

German Survey GerES II (GeS) (Hoffman et al 2000).

4. Toxicokinetics

Absorption

Limonene has a high partition coefficient between blood and air (λ

blood/air

= 42) and is easily taken up in the blood at the

alveolus (Falk et al., 1990). The net uptake of limonene in volunteers exposed to the chemical at concentrations of 450,

225, and 10 mg/m

for 2 hours during light physical exercise averaged 65% (Falk Filipsson et al., 1993). Orally

administered limonene is rapidly and almost completely taken up from the gastrointestinal tract in humans as well as in

animals (Igimi et al., 1974; Kodama et al., 1976).

Distribution

Limonene is rapidly distributed to different tissues in the body and is readily metabolized. Clearance from the blood

was 1.1 litre/kg body weight per hour in males exposed for 2 hours to d-limonene at 450 mg/m

(Falk Filipsson et al.,

1993). A high oil/blood partition coefficient and a long half-life during the slow elimination phase suggest high affinity

to adipose tissues (Falk et al., 1990; Falk Filipsson et al., 1993).

In rats, the tissue distribution of radioactivity was initially high in the liver, kidneys, and blood after the oral

administration of [

C] d-limonene (Igimi et al., 1974); however, negligible amounts of radioactivity were found after

48 hours. Differences between species regarding the renal disposition and protein binding of d-limonene have been

observed.

Metabolism and elimination

The biotransformation of d-limonene has been studied in many species, with several possible pathways of metabolism

(see Figure 4.1). Metabolic differences between species have been observed with respect to the metabolites present in

both plasma and urine. About 25-30% of an oral dose of d-limonene in humans was found in urine as d-limonene-8,9-

diol and its glucuronide; about 7-11% was eliminated as perillic acid (4-(1-methylethenyl)-1-cyclohexene-1-carboxylic

acid) and its metabolites (Smith et al., 1969; Kodama et al., 1976). d-Limonene-8,9-diol is probably formed via d-

limonene-8,9-epoxide (Kodama et al., 1976; Watabe et al., 1981). In another study, perillic acid was reported to be the

principal metabolite in plasma in both rats and humans (Crowell et al., 1992). Other reported pathways of limonene

metabolism involve ring hydroxylation and oxidation of the methyl group (Kodama et al., 1976).

The INDEX project Final report

231

Following the inhalation exposure of volunteers to d-limonene at 450 mg/m

for 2 hours, three phases of elimination

were observed in the blood, with half-lives of about 3, 33, and 750 minutes, respectively (Falk Filipsson et al., 1993).

About 1% of the amount taken up was eliminated unchanged in exhaled air, whereas about 0.003% was eliminated

unchanged in the urine. When male volunteers were administered (per os) 1.6 g [

C] d-limonene, 50-80% of the

radioactivity was eliminated in the urine within 2 days (Kodama et al., 1976). Limonene has been detected, but not

quantified, in breast milk of non-occupationally exposed mothers (Pellizzari et al., 1982).

Figure 4.1: Major pathways for d-limonene metabolism (Source: WHO 1998, from Kodama et al., 1976)

5. Health effects

Effects of short-term exposure

In general, limonene could be considered, with the exception of its irritative (skin, eyes) and sensitizing properties, to be

a chemical with fairly low acute toxicity (WHO, 1998). Potential hazard to the general population are skin irritancy

and sensitisation from use of consumer products, varying with the concentration of limonene in the product and, for

sensitisation, with its oxidation status (addition of oxidised limonene to the list of substances used in allergy testing has

been recommended)..

None of eight volunteers reported any discomfort, irritation, or symptoms related to central nervous system effects

during a 2-hour inhalation exposure in an exposure chamber (work load 50 W) to d-limonene at 10, 225, or 450 mg/m

;

however, a slight decline in vital capacity was observed following exposure to the highest concentration (Falk Filipsson

et al., 1993).

Eye irritation of l-limonene was measured using goggles instrumentation and a relatively short exposure time of 2

minutes. The threshold level for irritation in 12 volunteers was 1700-3400 mg/m

(Mølhave et al., 2000).

d-Limonene infused directly into the bile system of human volunteers to dissolve gallstones caused pain in the upper

abdomen, nausea, vomiting, and diarrhoea, as well as increases in serum aminotransferases and alkaline phosphatase

(Igimi et al., 1976, 1991). The oral administration of 20 g d-limonene to volunteers resulted in diarrhoea, painful

constrictions, and proteinuria, but no biochemical changes (total protein, bilirubin, cholesterol, aspartate

aminotransferase, alanine aminotransferase, alkaline phosphatase) in the liver (Igimi et al., 1976).

The INDEX project Final report

232

It has recently been proposed that reactions between limonene (or other unsaturated volatile compounds) and ozone or

reactive radicals produce chemically reactive products more likely to be responsible for eye and airway irritation than

the chemically non-reactive VOCs usually measured indoors (Weschler and Shields, 1997; Wolkoff et al., 1997; 2000).

Reaction mixtures of excess alpha-pinene, limonene and isoprene with ozone considerably below their NOEL

concentrations resulted in significant upper airway irritation (Wolkoff et al, 2000). The reduction of the respiratory rate

was from 30% to about 50%. Chemical analysis of reaction mixtures by conventional methods showed that readily

identified stable products and residual reactants at the concentrations found could not account for the observed

reductions of the respiratory rate, assuming additivity of the reaction products. The results suggest that, in addition to

known irritants (formaldehyde, acrolein, methacrolein, methyl vinylketone), one or more strong airway irritant(s) of

unknown structure(s) were formed (Wolkoff et al, 2000). Effects of variation of reaction time, relative humidity and

initial ozone concentration on irritant formation were described by Wilkins et al. (2003).

Kleno and Wolkoff (2004) exposed 8 male subjects for 20 min to limonene oxidation products (LOPs) and measured

relative changes in blink frequencies. The subjects were exposed locally in the non-dominant eye and single blind in

random order. Blinking was video-recorded and evaluated for full sessions of 36 min while the subjects viewed an

educational film. The mean blink frequency increased significantly at lower ppb levels of LOPs, 42% ( P<0.0001),

compared with that at baseline. Neither the residual reactants nor clean air changed the blink frequency significantly.

The findings coincided with qualitative reporting of weak eye irritation symptoms (Kleno and Wolkoff, 2004).

Irritative effects in experimental animals exposed to oxidation products of ozone and unsaturated hydrocarbons,

including limonene, were recently described in Wolkoff et al. (2000), Clausen et al. (2001), Rohr et al. (2002), Wilkins

et al. (2003).

Oxidised Histo-clear solvent (used in some pathology laboratories, containing d-limonene as a replacement for xylene)

produced respiratory irritation requiring medical treatment in a hystology technician (NICNAS, 2002). It was

considered that the autoxidation products might be the cause of the irritation. An increase in nasal or throat irritation

was also reported by 4/12 workers in a facility were a limonene-based cleaner was stored in an uncovered tank

(NICNAS, 2002).

The acute toxicity of d-limonene in rodents is fairly low after oral, intraperitoneal, subcutaneous, and intravenous

administration, based on the magnitude of the LD

values. LD

values were approximately 5 g/kg body weight for the

oral administration of d-limonene or d/l-limonene to rats and for dermal application of d/l-limonene to rabbits and 6

g/kg body weight for oral administration to mice (Tsuji et al., 1974, 1975b; Opdyke, 1978).

Effects observed following the acute exposure of rodents to limonene include increased bile flow at 85 mg/kg body

weight (Kodama et al., 1976), inhibition of S-3-hydroxy-3-methylglutaryl-CoA reductase activity at 409 mg/kg body

weight (Clegg et al., 1980), enzyme induction at 600 and 1200 mg/kg body weight (Ariyoshi et al., 1975), and

decreased motor activity, hypothermia, and potentiation of hexobarbital-induced sleep at 3 ml/kg body weight (Tsuji et

al., 1974).

Studies in guinea-pigs have revealed that air-oxidized d-limonene, but not d-limonene itself, induced contact allergy

(Karlberg et al., 1992).

Summary of short-term exposure effect levels

NOAEL

mg/m

LOAEL

mg/m

Target system;

Critical effect studied

Remarks Study Source

(Organization)

225 450 CNS related symptoms, irritation,

Endpoint: decline in vital capacity

Volunteers, 2h - work load 50W Falk Filipsson et al., 1993

Effects of long-term exposure

No information is available on the chronic health effects of inhalation exposure to d-limonene in humans, and no long-

term inhalation studies have been conducted in laboratory animals. Limonene/ozone reaction products have not yet been

submitted to chronic toxicity studies.

The INDEX project Final report

233

In numerous experimental studies on animals, oral exposure (gavage) to limonene has been shown to affect the liver.

Exposure affects the amount and activity of liver enzymes, liver weight, cholesterol levels and bile flow. However, no

toxic effects on the liver have been reported.The liver effects in animals are thought to be due to physiological

adaptation. Owing to a lack of data on d-limonene exposure in humans, this organ cannot with certainty be stated as the

critical organ in humans. From the data available, it is not possible to identify a NOAEL for these effects.

The only microscopic evidence of compound-related toxicity noted in rats was nephropathy in the males. d-Limonene is

one of a diverse group of hydrocarbons that has been shown to induce a unique syndrome of nephropathy in male rats

following subchronic or chronic exposure. Based on a review of the literature concerning this effect (U.S. EPA, 1991),

EPA's Risk Assessment Forum concluded that nephropathy in male rats that is associated with alpha- 2u-globulin

accumulation in hyaline droplets is not an appropriate endpoint to determine noncancer effects potentially occurring in

humans (U.S. EPA, 1993).

In two subchronic oral exposure studies in rats, the following effect levels were identified. The NOEL, based upon

histopatological examination of the kidneys, was considered to be 5 mg/kg bw/d. The LOEL for increased liver and

kidney weight was 75 mg/kg bw/d. For effects in the liver the NOEL was 10 mg/kg bw/d and the LOAEL was 30

mg/kg bw/d (Ariyoshi et al., 1975; Webb et al., 1989).

Based on available data, food is believed to be the principal source of exposure (96%) to limonene; the contribution

from ambient air is approximately 4%. To calculate a tolerable intake for humans, the animal study was chosen in

which effects on the liver were observed at the lowest exposure level (Webb et al., 1989). In this study, gavage

administration of d-limonene (5 days/week for 13 weeks) to rats caused increased relative liver weight at 30 and 75

mg/kg body weight per day. The NOEL for the liver was considered to be 10 mg/kg body weight per day. Using

uncertainty factors of 10 for intraspecies differences and 10 for interspecies differences, a tolerable intake for ingestion

of d-limonene by humans of 0.1 mg/kg body weight per day may be calculated from the NOEL. This value is of a

similar magnitude as the estimated daily US consumption of d-limonene of 0.27 mg/kg body weight per day (Flavor

and Extract Manufacturers Association, 1991). U.S. FDA approves d-limonene for use as a food additive.

Matthys et al. (2000) observed that “Myrtol standardized” (Gelomyrtol forte), a phytotherapeutic extract (distillate)

consisting mainly of three monoterpenes: d-limonene, (+)alpha-pinene, and 1,8-cineole, is a well-evidenced alternative

to antibiotics for acute bronchitis treatment and was considered safe and tolerable in 170 patients at daily doses of 4 x

300 mg for 2 weeks.

Occupational exposure standards for limonene are summarised in Table 5.1. Supporting documentation for the exposure

standard adopted by the American Industrial Hygiene Association (AIHA) states that the 167 mg/m

(30 ppm) 8-hr

TWA was set to protect against liver effects seen in male mice and reduced survival in female rats in a 2-yr NTP study

(AIHA, 1993). In Germany no MAK (maximum workplace concentration) has been set for limonene as it was

considered that insufficient information was available.

Table 5.1: National occupational exposure standards for limonene

Country 8-h TWA

mg/m

STEL

mg/m

Year adopted Reference

Norway 140 - 1999 RTECS, 2001

Sweden 140 280 1990 Karlberg and Lindell, 1993

Finland 140 280 not known Finnish Institute of Occupational Health, 2001

AIHA 167 - 1993 AIHA, 1993

TWA: time-weighted average

STEL: short-term (15 min) exposure limit

U.S. EPA deemed the database at the time of review (1993) to be insufficient to derive an inhalation RfC according to

the Interim Methods for Development of Inhalation Reference Concentrations (U.S. EPA, 1990).

Carcinogenic and genotoxic effects

There is inadequate evidence in humans for the carcinogenicity of d-limonene.

There is sufficient evidence in experimental animals for the carcinogenicity of d-limonene. Overall evaluation: In

making its overall evaluation of the carcinogenicity to humans of d-limonene, the IARC Working Group concluded that

d-limonene produces renal tubular tumors in male rats by a non-DNA reactive alpha-2-globulin associated response.

Therefore, the mechanism by which d-limonene increases the incidence of renal tubular tumors in male rats is not

The INDEX project Final report

234

relevant to humans. d-Limonene is not classifiable as to its carcinogenicity to humans (Group 3) (IARC, 1999).

The few available data indicate that d-limonene and its 1,2-epoxide metabolite are not genotoxic. Limonene

administered in the diet was not mutagenic in the liver or kidney of male Big Blue rats Turner et. al. (2001).

Interactions with other chemicals

The formation of unidentified strong upper airway irritants in reaction mixtures of terpenes and ozone has been recently

reported. Identified products included formaldehyde and other aldehydes, carboxylic acids, peroxides, which may cause

irritation at low concentrations or odor annoyance due to their usually low odour thresholds (Weschler and Shields,

1997; Wolkoff et al., 1997; 2000).

Maximum irritation of reaction mixtures of limonene and ozone, measured in mice, was observed at low humidity (<2%

RH) and short time (16-30 s) reaction mixtures. Moderate humidity (approximately 32% RH) and longer reaction times

(60-90 s) resulted in significantly less irritation, suggesting that some unidentified intermediates react with water vapor

to give less irritating products (Wilkins et al., 2003). Irritation measured at four ozone concentrations (1, 2, 4 and 7

mg/m

) using low humidity/short time reaction conditions for limonene (280 mg/m

) and isoprene (1400 mg/m

)

revealed that at 1 mg/m

, irritation was at the same level as that for the pure terpenes, indicating that at 1 mg/m

ozone

the combined irritant effect was near the no effect level for the product mixture (Wilkins et al., 2003).

Odour perception

Source: Leffingwell & Associates

d-limonene - fresh citrus, orange-like - odour perception : 1.1 mg/m

(200 ppb)

l-limonene - harsh, turpentine-like, lemon note - odour perception : 2.8 mg/m

(500 ppb)

Summary of Limonene Dose Response Assessment

________________________________________________________________________________________________

Exposure other than inhalation: Oral: About 96% of limonene absorbed by humans originate from food. Generally

recognized as safe in food by U.S.FDA. Dermal: Potential hazard to the general population may derive from skin

irritancy and sensitisation from use of consumer products For contact allergy, air-oxidized limonene seems to play a

major role than limonene itself.

Toxicokinetics: ~65% net uptake of the inhaled amount, rapid distribution and high affinity for adipose tissue

Health effect levels of short- and long-term exposure

NOAEL

mg/m³

LOAEL

mg/m³

Target system; critical effects Remarks Study Source; Ref.

Value mg/m³

Short-term exposure

225 450 CNS related symptoms, irritation, Endpoint:

decline in vital capacity

Volunteers, 2h - work

load 50W

Falk Filipsson et al.,

1993

Long-term exposure (no studies have been found in literature)

Carcinogenicity: Inadequate evidence in humans; sufficient evidence in animals, with mechanism not relevant to

humans.

Genotoxicity: The few available data indicate that d-limonene and its 1,2-epoxide metabolite are not genotoxic.

Odour threshold: 2.15 mg/m

(Devos); d-limonene (orange-like): 1.1 mg/m

; l-limonene (turpentine-like) : 2.8 mg/m

(Leffingwell & Associates)

Susceptible population:

Remarks: RD

: 6 g/m

; In addition to known irritant reaction products (formaldehyde, acrolein, methacrolein, methyl

vinylketone), one or more strong airway irritant(s) of unknown structure(s) could be formed following the reaction

between limonene and ozone or reactive radicals.

The INDEX project Final report

235

6. Risk Characterization

Health hazard evaluation of long-term exposure

Effect level - mg/m

Assessment

factor

mg/m

Toxicological endpoint

Long-term

Exposure Limit

Human

LOAEL

Volunteers

450 1000

abc

0.45

CNS related symptoms, irritation, Endpoint:

decline in vital capacity

extrapolation from subacute to chronic exposure (10);

not considering a NOAEL (10)

intraspecies variability (10);

Relevance of EU-population exposure to limonene

Population based studies

derived

Margin of

safety

(MOS)

Description (Study, Year) N 0.45 mg/m

(90

)

Athens (Expolis, 96-98) 42 < 11 (3)

Basel (Expolis, 96-98) 47 < 49 (14)

Helsinki (Expolis, 96-98) 188 < 33 (6)

Milan (Expolis, 96-98) 41 < 14 (3)

Oxford (Expolis, 98-00) 40 < 43 (11)

Prague (Expolis, 96-98) 46 < 23 (8)

England 28d-TWA (BRE, 97-99) 796 < 56 (11)

Avg MOS: 33 (8)

< out of the evaluation range (i.e. <5% of the environments investigated)

Result

An attempt has been done in deriving an exposure limit (EL) for long-term effects associated with limonene exposure

by refering to a study on volunteers exposed at sub-acute (2 hours) inhalation doses. When comparing this EL (450

µg/m

) with the results from seven indoor surveys it is concluded that no neurologic effects would be expected at

background limonene levels encountered in European homes, with median (90

percentile) levels at least 10 (3) times

lower than the proposed EL. It is assumed that at 10-fold the level set as the EL, irritation could be expected following

acute exposure. Due to its widespread use as a flavouring agent in numerous consumer products, short-term exposures

at levels in the order of some mg/m

could not be excluded, although significative exposure data are lacking.

The INDEX project Final report

236

alpha-Pinene

Synonyms: 2,6,6-Trimethylbicyclo(3.1.1)-2-hept-2-ene

2,6,6-Trimethylbicyclo(3.1.1)hept-2-ene

2-Pinene

CAS Registry Numbers: 80-56-8

Molecular Formula: C

1. Compound identification

Alfa-pinene is a colourless liquid with a characteristic odour of pine. It is a naturally occurring terpene, which is emitted

by trees, fruits, grasses, bushes, fungi, herbs, and flowers and it is a constituent of many common oils. Alfa-pinene will

exist solely as a vapour in the atmosphere. It is used in aerosol paint concentrates, cleaning and sanitation products,

paints and varnish removers, waterproofing compounds solvent, and flavouring. It is also emitted to indoor air from

commonly used wooden furniture and waxes (HSDB 2003, Maroni et al 1995).

Alpha and beta pinene (C

) are principal constituents of turpentine. Turpentine is a thin volatile essential oil

) obtained by steam distillation from the wood or the exudate of pine trees. Main uses of turpentine include as a

solvent thinner for paint, varnishes, and lacquer, and as a vehicle for paints (professional painters, private individuals).

Turpentine is also used in perfumery, sprays, and deodorizers. The composition of wood turpentine varies with species,

location, and season. Steam-distilled (wood) turpentine produced in the United States is made up primarily of

alphapinene (75 to 85%) with varying amounts of beta-pinene (up to 3%), camphene (4 to 15%), limonene (dipentene, 5

to 15%), 3-carene, and terpinolene (percentages not provided) (Chinn, 1989). In a Swedish study (Falk Filipsson, 1996)

volunteers were exposed to turpentine vapour with a composition measured in air: 54% alpha-pinene, 11% beta-pinene

and 35% 3-carene.

Exposure to alpha-pinene in the general population may occur by inhalation, by dermal contact of consumer products

and by ingestion of foods.

2. Physical and Chemical properties

Molecular weight (g/mol) 136.23

Melting point (°C) -55

Boiling point (°C) 156

Density (at 20 oC, 1 atm) 0.859

Relative density (air =1) 4.7

Solubility: Insoluble in water, soluble in alcohol, chloroform,

ether, and glacial acetic acid

Conversion factors at 20 °C and 760 mm Hg:

1 ppb = 5.654 µg/m

1 µg/m

= 0.177 ppb

Sources: Verschueren 2001, HSDB 2003

The INDEX project Final report

237

3. Indoor Air Exposure assessment

Terpene/ozone reaction products

It has been proposed that reactions between unsaturated volatile compounds (e.g. limonene, a-pinene, styrene) and

ozone or hydroxyl (OH) radicals produce chemically reactive products more likely to be responsible for eye and airway

irritation than the chemically non-reactive VOCs usually measured indoors. Many of the known reaction products are

oxygenates (e.g. formaldehyde and other aldehydes, carboxylic acids, peroxides) which may cause irritation at low

concentrations or odour annoyance due to their usually low odour thresholds (Weschler and Shields, 1997; Wolkoff et

al., 1997; 2000).

Indoor air and exposure concentrations

Exposures to alpha-Pinene in indoor environments did not follow the same increasing trend from north to south as many

other VOC compounds do (Figure 3.1). Average indoor concentrations were at about the same level in Northern and

Central Europe, being typically 15 – 17 µg/m

(EXPOLIS 2002, Edwards et al 2001)). Slightly lower levels were

measured in Southern Europe, being there about 10-17 µg/m

. Maroni et al (1995) concluded, as a summary of several

studies, a typical mean concentration being 10 µg/m

for alpha-Pinene in indoor air. The results presented by Kostiainen

et al (1995) in Helsinki, Finland showed similar mean and median indoor concentrations, 9.3 µg/m

and 7.7 µg/m

respectively. Wolkoff et al (2000) reviewed alpha-Pinene concentrations in major indoor air studies. Mean

concentrations ranged from 10 µg/m

to 40 µg/m

and maximum concentrations from 8 µg/m

to 250 µg/m

. Rehwagen

et al reported 4-week mean concentration of 25 µg/m

and a maximum value of 393 µg/m

in Leipzig, Germany.

Interesting detail in this study was the increasing trend of terpenes between 1996 and 1999 on the contrary to other

VOCs, which showed a decreasing trend. Explanations to this trend for alpha-Pinene could not be identified.

Average personal exposures to alpha-Pinene ranged from 7 µg/m

to 18 µg/m

(Figure 3.2, Jantunen et al 1999,

Hoffman et al 2000). All over Europe average population exposures tend to be lower than respective indoor

concentrations, suggesting the presence of remarkable indoor sources of alpha-Pinene. Exposures to alpha-Pinene in

Europe were considerably higher than those measured in TEAM studies in the USA, median daytime exposures ranging

from 1.0 µg/m

to 2.8 µg/m

(Wallace et al 1996).

100

0 1020304050607080

Indoor concentration (µg/m

)

Cumulative frequency (%)

Ath

Bas

Hel

Mil

Oxf

Pra

Figure 3.1. Cumulative frequency distributions of indoor air concentrations of alpha-Pinene in Athens (Ath, n=42),

Basel (Bas, n=47), Helsinki (Hel, n=188), Milan (Mil, n=41) Oxford (Oxf, n=40) and Prague (Pra, n=46) (EXPOLIS

2002).

The INDEX project Final report

238

100

0 1020304050607080

Exposure (µg/m

)

Cumulative frequency (%)

Ath

Bas

GeS

Hel

Oxf

Pra

Figure 3.2. Cumulative frequency distributions of 48-hour personal exposure concentrations of alpha-Pinene in Athens

(Ath), Basel (Bas), Helsinki (Hel), Oxford (Oxf) and Prague (Pra) (EXPOLIS 2002), and 1-week mean exposures of the

German Survey GerES II (GeS) (Hoffman et al 2000).

4. Toxicokinetics

Absorption

The toxicokinetics of alpha-pinene was studied in volunteers experimentally exposed to the vapour of the pure

compound and of turpentine with composition in air: 54% alpha-pinene, 11% beta-pinene and 35% 3-carene (Falk et

al., 1990a; Falk Filipsson, 1996). In the first study eight healthy males were exposed to 0, 10, 225, or 450 mg/m

(+)-

alpha-pinene or 450 mg/m

(-)-alpha-pinene for 2 hr in an inhalation chamber. During exposure they exercised on a

cycle ergometer at the rate 50 watts. Average pulmonary uptake of (+)-alpha-pinene and (-)-alpha-pinene amounted to

59% (62%, 66%, and 68% respectively for turpentine vapour constituents in Falk Filipsson, 1996) of the exposure

concentration. The blood:air partition coefficient for alpha-pinene is 15. Absolute uptake increased linearly with

concentration. Blood alpha-pinene concentration increased rapidly at first then tapered off. Mean blood concentration at

the end of exposure were linearly related to inhaled concentration.

Distribution

The solubility of monoterpenes in olive oil is high; the oil:air partition coefficient for alpha-pinene is 2900. This should

imply accumulation in adipose tissue (Falk et al., 1990b). In rats, terpenes accumulate in peripheral fat, kidneys, and the

brain (Savolainen and Pfaffli, 1978; Sperling et al. 1967).

Metabolism and elimination

Slow metabolism and renal elimination of monoterpenes were suggested after a patient acutely poisoned by pine oil was

studied (Koppel et al., 1981). Up to 4% of the total uptake of alpha-pinene after human exposure by inhalation was

eliminated in the urine as cis- and transverbenol (Levin et al. 1992). Both the respiratory elimination of alpha-pinene

and the urinary excretion of verbenols after inhalation of alpha-pinene enantiomers were studied (same exposure

conditions as in Falk et al., 1990a). Respiratory elimination of both pinene enantiomers was similar; at a concentration

of 450 mg/m

, 7.7% of the total uptake of (+)alpha-pinene and 7.5% of the total uptake of (-)alpha-pinene was

eliminated. Urinary excretion of verbenol 4 hr after exposure to (+)alpha-pinene ranged from 1.7% at 450 mg/m

3.8% at a dose of 10 mg/m

. Urinary excretion of (-)alpha-pinene was similar. A semilogarithmic plot of the excretion

data suggested the existence of more than one rate constant for the elimination of (+)alpha-pinene and (-)alpha-pinene.

Most of the verbenols were eliminated within 20 hr after a 2 hr exposure. The renal excretion of unchanged alpha-

pinene was less than 0.001%. The determination of urinary verbenols may be useful as a biological exposure index for

exposure to terpenes.

The INDEX project Final report

239

Elimination of alpha-pinene from the blood was triphasic (Falk et al., 1990a). Half times for elimination of inhaled (+)-

alpha-pinene from the blood during the three phases were 4.8, 39, and 695 minutes. Elimination half times for (-)-alpha-

pinene were 5.6, 40, and 555 minutes.

5. Health effects

Effects of short-term exposure

Most of the studies identified relative to α-pinene were associated with human responses to one or several

monoterpenes. Acute toxic effects of α-pinene are stated as similar to those resulting from turpentine exposures

(Budavari, 1996; NIEHS, 2002).

Pulmonary function and subjective ratings of discomfort were described in two studies (Falk et al., 1990a; Falk

Filipsson, 1996), whereby eight healthy male volunteers were exposed for 2 hr in an inhalation chamber (working load

50W with a cycle ergometer) to 0, 10, 225, or 450 mg/m

(+)-alpha-pinene or 450 mg/m

(-)-alpha-pinene and to 450

mg/m

turpentine vapour (composition in air: 54% alpha-pinene, 11% beta-pinene and 35% 3-carene) , respectively.

Five among eight subjects complained of eye, nose and throat irritation, after being exposed to 450 mg/m

alpha-pinene.

No exposure related changes in lung function were seen. At the concentration tested the capacity of the liver to

metabolize alpha-pinene was not exceeded (Falk et al., 1990a). Subjects experienced discomfort in the throat and

airways during exposure to turpentine vapour (450 mg/m

) and airway resistance was increased after the end of

exposure (Falk Filipsson, 1996).

In male and female volunteers (mean age: 35 years) exposed to 0 or 450 mg/m

of a 10:1:5 mixture of alpha- and beta-

pinene and 3-carene (a synthetic turpentine) for 12 hr, 4 times during a 2 week period, an acute alveolar cellular

inflamatory reaction was observed. Exposure did not significantly alter bronchial hyperreactivity to methacholine

(Bingham et al., 2001).

Mølhave et al (2000) concluded that similarly to other 3 terpenes ((+)3-carene, (-)limonene, (rac)alpha-terpineol) α-

pinene can probably be ruled out as cause of acute eye irritation indoors. After applying eye goggles, too few subjects

reported eye-irritation for α-pinene to allow estimates of a threshold of this compound, having much less irritative

potency than n-butanol, 3-carene, and limonene.

No acute effects on forced vital capacity or forced expiratory volume (1s) were detected after personal exposure to 10-

214 mg/m

monoterpenes (alpha-pinene, beta-pinene and delta 3-carene) of 38 joinery workers (Eriksson et al., 1997)

and to 11-158 mg/m

terpenes of 48 sawmill workers (Eriksson et al., 1996). A decrease in carbon monoxide lung

diffusing capacity after a workshift was detected. Workers with ≥5 years of sawmill employment showed a higher

reactivity to methacholine than those with < 5 years. Eye irritation increased during a workday (Eriksson et al., 1996).

For single monoterpenes, mixtures of monoterpenes and for turpentine the Swedish occupational exposure limit (OEL)

value is 150 mg/m

. In U.S. no occupational exposure limits have been established for alpha-Pinene. For turpentine,

workplace exposure limits for 8-h workshift are 557 mg/m

(OSHA), 557 mg/m

(NIOSH), and 111 mg/m

(ACGIH).

As in humans, turpentine acts as a CNS depressant, with symptoms progressing from lethargy, prostration, and

convulsions to death in animals (Bystrom, 2000). Acute toxicity in animals included irritation of the skin, eyes, nose,

and mucous membranes. Signs and symptoms of acute toxicity are CNS depression and increased respiration rate with a

decrease in tidal volume. Major systemic effects include kidney and bladder injury, and are thought to be attributable to

induction of an acute inflammatory response (Key et al., 1977; cited by Baxter, 2001).

In mice exposed by inhalation to turpentine, the RD

(the concentration that causes a 50% decrease in respiratory

frequency) of turpentine (6500 mg/m

) was the same order of magnitude as those of α-pinene (5900 mg/m

) and β-

pinene (7100 mg/m

) (Kasanen et al., 1999; Kasanen et al., 1998). The effect on breathing was most likely due to

sedation or anesthesia since no histological observations were reported in the exposed animals.

The INDEX project Final report

240

Summary of short-term exposure effect levels

NOAEL

mg/m

LOAEL

mg/m

Target system;

Critical effect studied

Remarks Study Source

(Organization)

225 450 eyes, nose and throat irritation Volunteers, 2h - work load 50W Falk et al., 1990b

UBA, 2003

GV II: 2;

GV I: 0.2

UBA, 2003

Effects of long-term exposure

No chronic exposure studies were identified for both α-pinene and turpentine.

Forty-eight subjects exposed to terpenes (mean air concentration 258 mg/m

) and 47 unexposed subjects, all employed

at sawmills, were studied with regard to symptoms and pulmonary function. Dyspnoea and chest oppression were

significantly increased in the exposed subjects compared to the unexposed controls. A reduced FEV1, on spirometry

and an increased CV% and slope of the alveolar plateau (phase III) on single breath nitrogen washout were seen on

Monday morning before exposure to terpenes. There was no correlation between exposure time (duration of

employment) and lung function impairment. A day of industrial exposure to terpenes caused no further change in any

lung function variable. The unexposed controls showed normal spirometry and nitrogen washouts. The findings indicate

a slight stable lung function impairment of an obstructive nature which does not necessarily undergo further

deterioration with increased duration of exposure (Hedenstierna et al., 1983).

Prolonged exposure to terpenes may result in allergic contact dermatitis (Brun, 1975; Dooms-Goossens et al., 1977) or

chronic impairment of lung function (Hedenstierna et al., 1983).

Chronic effects associated with occupational exposures to turpentine include cerebral atrophy, behavioral changes,

anemia and bone marrow damage, glomerulonephritis, and dermatitis. Urinary disturbances, albuminuria, and urinary

casts were observed in workers exposed to paints and varnishes. However, renal damage associated with occupational

exposures to turpentine was transient and reversible (NIEHS, 2002).

Inhalation of turpentine (4000 mg/m

), four hours per day, 45 or 58 days failed to result in any hematological changes

in guinea pigs, though slight changes in the liver and moderate scattered tubular degeneration in the kidneys were

observed. Subchronic inhalation of turpentine (1670 mg/m

) in rats resulted in the accumulation of α-pinene in brain

and perinephric fat. At higher concentrations (estimated at 5000 to 10,000 mg/m

) for up to 293 hours over a period of

time ranging to 14 months provided no histological evidence of nephritis or chronic Bright’s disease. There were foci of

pneumonitis and extensive lung abscesses in many of these animals (NIEHS, 2002).

Carcinogenic and genotoxic potential

The International Agency for Research on Cancer has not evaluated the carcinogenicity of this chemical. The US

National Toxicology Program has not listed this chemical in its report on carcinogens. The American Conference of

Governmental Industrial Hygienists has designated α-pinene as not classifiable as a human carcinogen (A4). No human

information is available and the limited amount of animal information available is inconclusive.

Toxicity associated with turpentine oleoresin [CAS RN 9005-90-7] includes the development of benign tumors after

chronic exposures (Budavari, 1996).

A case-control study of workers in particle-board, plywood, sawmill, and formaldehyde glue factories demonstrated a

statistically significant association between chronic exposure (longer than 5 years) to terpenes (the principal component

of turpentine) and the development of respiratory tract cancers (HSDB, 1989).

α-Pinene was not found to be mutagenic, assayed on TA98 or TA100 in the presence of S9 (Rockwell et.al., 1979), in a

screening assay using the Ames test (Florin et. al., 1980) or using a battery of bacterial test strains (Connor et. al.

1985).

The INDEX project Final report

241

Interactions with other chemicals

It has recently been proposed that reactions between α-pinene (or other unsaturated volatile compounds) and ozone or

reactive radicals produce chemically reactive products more likely to be responsible for eye and airway irritation than

the chemically non-reactive VOCs usually measured indoors (Weschler and Shields, 1997; Wolkoff et al., 1997; 2000).

Irritative effects in experimental animals exposed to these oxidation products, were also described in Clausen et al.

(2001), Rohr et al. (2002) and Wilkins et al. (2003).