This report contains the collective views of an international group of experts and does not necessarily represent the decisions or the stated policy of the United Nations Environment Programme, the International Labour Organization, or the World Health Organization.

First draft prepared by G. Long and M.E. Meek, Health Canada, Ottawa, Canada, and P. Cureton, Environment Canada, Ottawa, Canada

Published under the joint sponsorship of the United Nations Environment Programme, the International Labour Organization, and the World Health Organization, and produced within the framework of the Inter-Organization Programme for the Sound Management of Chemicals.

World Health Organization

Geneva, 2002

The International Programme on Chemical Safety (IPCS), established in 1980, is a joint venture of the United Nations Environment Programme (UNEP), the International Labour Organization (ILO), and the World Health Organization (WHO). The overall objectives of the IPCS are to establish the scientific basis for assessment of the risk to human health and the environment from exposure to chemicals, through international peer review processes, as a prerequisite for the promotion of chemical safety, and to provide technical assistance in strengthening national capacities for the sound management of chemicals.

The Inter-Organization Programme for the Sound Management of Chemicals (IOMC) was established in 1995 by UNEP, ILO, the Food and Agriculture Organization of the United Nations, WHO, the United Nations Industrial Development Organization, the United Nations Institute for Training and Research, and the Organisation for Economic Co-operation and Development (Participating Organizations), following recommendations made by the 1992 UN Conference on Environment and Development to strengthen cooperation and increase coordination in the field of chemical safety. The purpose of the IOMC is to promote coordination of the policies and activities pursued by the Participating Organizations, jointly or separately, to achieve the sound management of chemicals in relation to human health and the environment.

WHO Library Cataloguing-in-Publication Data

Acrylonitrile.

(Concise international chemical assessment document ; 39)

1.Acrylonitrile - toxicity 2.Risk assessment

3.Environmental exposure 4.Occupational exposure

I.International Programme on Chemical Safety II.Series

ISBN 92 4 153039 1        (NLM Classification: QV 633)

ISSN 1020-6167

The World Health Organization welcomes requests for permission to reproduce or translate its publications, in part or in full. Applications and enquiries should be addressed to the Office of Publications, World Health Organization, Geneva, Switzerland, which will be glad to provide the latest information on any changes made to the text, plans for new editions, and reprints and translations already available.

©World Health Organization 2002

Publications of the World Health Organization enjoy copyright protection in accordance with the provisions of Protocol 2 of the Universal Copyright Convention. All rights reserved.

The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the Secretariat of the World Health Organization concerning the legal status of any country, territory, city, or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.

The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.

The Federal Ministry for the Environment, Nature Conservation and Nuclear Safety, Germany, provided financial support for the printing of this publication.

Printed by Wissenschaftliche Verlagsgesellschaft mbH, D-70009 Stuttgart 10

Concise International Chemical Assessment Documents (CICADs) are the latest in a family of publications from the International Programme on Chemical Safety (IPCS) — a cooperative programme of the World Health Organization (WHO), the International Labour Organization (ILO), and the United Nations Environment Programme (UNEP). CICADs join the Environmental Health Criteria documents (EHCs) as authoritative documents on the risk assessment of chemicals.

International Chemical Safety Cards on the relevant chemical(s) are attached at the end of the CICAD, to provide the reader with concise information on the protection of human health and on emergency action. They are produced in a separate peer-reviewed procedure at IPCS. They may be complemented by information from IPCS Poison Information Monographs (PIM), similarly produced separately from the CICAD process.

CICADs are concise documents that provide summaries of the relevant scientific information concerning the potential effects of chemicals upon human health and/or the environment. They are based on selected national or regional evaluation documents or on existing EHCs. Before acceptance for publication as CICADs by IPCS, these documents undergo extensive peer review by internationally selected experts to ensure their completeness, accuracy in the way in which the original data are represented, and the validity of the conclusions drawn.

The primary objective of CICADs is characterization of hazard and dose–response from exposure to a chemical. CICADs are not a summary of all available data on a particular chemical; rather, they include only that information considered critical for characterization of the risk posed by the chemical. The critical studies are, however, presented in sufficient detail to support the conclusions drawn. For additional information, the reader should consult the identified source documents upon which the CICAD has been based.

Risks to human health and the environment will vary considerably depending upon the type and extent of exposure. Responsible authorities are strongly encouraged to characterize risk on the basis of locally measured or predicted exposure scenarios. To assist the reader, examples of exposure estimation and risk characterization are provided in CICADs, whenever possible. These examples cannot be considered as representing all possible exposure situations, but are provided as guidance only. The reader is referred to EHC 170¹ for advice on the derivation of health-based guidance values.

While every effort is made to ensure that CICADs represent the current status of knowledge, new information is being developed constantly. Unless otherwise stated, CICADs are based on a search of the scientific literature to the date shown in the executive summary. In the event that a reader becomes aware of new information that would change the conclusions drawn in a CICAD, the reader is requested to contact IPCS to inform it of the new information.

The flow chart shows the procedures followed to produce a CICAD. These procedures are designed to take advantage of the expertise that exists around the world — expertise that is required to produce the high-quality evaluations of toxicological, exposure, and other data that are necessary for assessing risks to human health and/or the environment. The IPCS Risk Assessment Steering Group advises the Co-ordinator, IPCS, on the selection of chemicals for an IPCS risk assessment, the appropriate form of the document (i.e., EHC or CICAD), and which institution bears the responsibility of the document production, as well as on the type and extent of the international peer review.

The first draft is based on an existing national, regional, or international review. Authors of the first draft are usually, but not necessarily, from the institution that developed the original review. A standard outline has been developed to encourage consistency in form. The first draft undergoes primary review by IPCS and one or more experienced authors of criteria documents to ensure that it meets the specified criteria for CICADs.

The draft is then sent to an international peer review by scientists known for their particular expertise and by scientists selected from an international roster compiled by IPCS through recommendations from IPCS national Contact Points and from IPCS Participating Institutions. Adequate time is allowed for the selected experts to undertake a thorough review. Authors are required to take reviewers’ comments into account and revise their draft, if necessary. The resulting second draft is submitted to a Final Review Board together with the reviewers’ comments.

A consultative group may be necessary to advise on specific issues in the risk assessment document.

The CICAD Final Review Board has several important functions:

• to ensure that each CICAD has been subjected to an appropriate and thorough peer review;
• to verify that the peer reviewers comments have been addressed appropriately;
• to provide guidance to those responsible for the preparation of CICADs on how to resolve any remaining issues if, in the opinion of the Board, the author has not adequately addressed all comments of the reviewers; and
• to approve CICADs as international assessments.

Board members serve in their personal capacity, not as representatives of any organization, government, or industry. They are selected because of their expertise in human and environmental toxicology or because of their experience in the regulation of chemicals. Boards are chosen according to the range of expertise required for a meeting and the need for balanced geographic representation.

Board members, authors, reviewers, consultants, and advisers who participate in the preparation of a CICAD are required to declare any real or potential conflict of interest in relation to the subjects under discussion at any stage of the process. Representatives of nongovernmental organizations may be invited to observe the proceedings of the Final Review Board. Observers may participate in Board discussions only at the invitation of the Chairperson, and they may not participate in the final decision-making process.

This CICAD on acrylonitrile was prepared jointly by the Environmental Health Directorate of Health Canada and the Commercial Chemicals Evaluation Branch of Environment Canada based on documentation prepared concurrently as part of the Priority Substances Program under the Canadian Environmental Protection Act (CEPA). The objective of assessments on Priority Substances under CEPA is to assess potential effects of indirect exposure in the general environment on human health as well as environmental effects. Data identified as of the end of 31 May 1998 (environmental effects) and April 1998² (human health effects) were considered in this review. Other reviews that were also consulted include US EPA (1980, 1985), IPCS (1983), ATSDR (1990), IARC (1999), and EC (2000). Information on the nature of the peer review and availability of the source document (Environment Canada & Health Canada, 2000) is presented in Appendix 1. Information on the peer review of this CICAD is presented in Appendix 2. This CICAD was approved as an international assessment at a meeting of the Final Review Board, held in Geneva, Switzerland, on 8–12 January 2001. Participants at the Final Review Board meeting are listed in Appendix 3. The International Chemical Safety Card on acrylonitrile (ICSC 0092), produced by the International Programme on Chemical Safety (IPCS, 1993), has also been reproduced in this document.

Acrylonitrile (CAS No. 107-13-1) is a volatile, flammable, water-soluble liquid at room temperature. The large majority of acrylonitrile in Canada is used as a feedstock or chemical aid in the production of nitrile-butadiene rubber and in acrylonitrile-butadiene-styrene and styrene-acrylonitrile polymers. Estimated world capacity in 1993 was about 4 million tonnes. Major areas of production are the European Union (>1.25 million tonnes per year), the USA (approximately 1.5 million tonnes per year), and Japan (approximately 0.6 million tonnes per year).

Acrylonitrile is released into the environment primarily from chemical production and the chemical and plastic products industries (>95% in the sample country). There are no known natural sources. Acrylonitrile is distributed largely to the environmental compartments to which it is principally released (i.e., air or water), with movement to soil, sediment, or biota being limited; reaction and advection are the major removal mechanisms. In limited surveys in the country on which the sample risk characterization is based (i.e., Canada), acrylonitrile has been detected in the general environment only in the vicinity of industrial sources.

Occupational exposure to acrylonitrile occurs during production and its use in the manufacture of other products; potential for exposure is greater in the latter case, where the compound may not be as easily contained. Based on recent data for countries in the European Union, time-weighted-average (TWA) exposures are 0.45 ppm (<1 mg/m³) during production and 1.01 ppm (<2.2 mg/m³) in various end uses.

Acrylonitrile is rapidly absorbed via all routes of exposure and distributed throughout examined tissues. There is little potential for significant accumulation in any organ, with most of the compound being excreted primarily as metabolites in the urine within the first 24–48 h following administration. Available data are consistent with conjugation to glutathione being the major detoxification pathway, while oxidation to 2-cyanoethylene oxide is considered an activation pathway.

Available data from studies in animals indicate that acrylonitrile is a skin, respiratory, and severe eye irritant. Acrylonitrile may cause allergic contact dermatitis, but the available data are inadequate to assess its sensitization potency. With the exception of developmental toxicity, for which effects (fetotoxic and teratogenic) have not been observed at concentrations that were not toxic to the mothers, available data on other non-neoplastic effects in experimental animals are inadequate to characterize exposure–response. In the few studies in human populations in which non-neoplastic effects of acrylonitrile have been systematically investigated, only acute dermal irritation has been reported consistently.

Based on studies in animals, cancer is the critical end-point for effects of acrylonitrile on human health. A range of tumours in rats — including those of the central nervous system (brain and/or spinal cord), ear canal, gastrointestinal tract, and mammary glands — has been consistently observed following both ingestion and inhalation. In almost all adequate bioassays, there have been reported increases in astrocytomas of the brain and spinal cord, which are rarely observed spontaneously; these have occurred at highest incidence consistently across studies. Increases have been statistically significant, and there have been clear dose–response trends. Tumours have sometimes been reported at non-toxic doses or concentrations and at periods as early as 7–12 months following onset of exposure. Tumours have also been observed in exposed offspring of a multigeneration reproductive study at 45 weeks.

Increases in cancer incidence have not been consistently observed in available epidemiological studies. However, the meaningful quantitative comparison of the results of these investigations with those of studies in animals is precluded by inadequate data on mode of induction of brain tumours, relative paucity of data on exposure of workers in the relevant investigations, and the wide range of the confidence limits on the standardized mortality ratios for cancers of possible interest in the epidemiological studies.

In numerous studies on the genotoxicity of acrylonitrile involving examination of a broad spectrum of end-points both in vitro, with and without metabolic activation, and in vivo in mice and rats, the pattern of results has been quite mixed, while the metabolite, cyanoethylene oxide, is mutagenic. Although direct evidence is not available, based on available data, it is reasonable to assume that induction of tumours by acrylonitrile involves direct interaction with genetic material. The weight of evidence for other potential modes of induction of tumours is inadequate. Acrylonitrile or its epoxide can react with macromolecules.

Cancer is considered the critical end-point for quantification of exposure–response for risk characterization for acrylonitrile. The lowest tumorigenic concentration (TC₀₅, the concentration that causes a 5% increase in tumour incidence over background) (human equivalent value) was 2.7 ppm (6.0 mg/m³) for the combined incidence of benign and malignant tumours of the brain and/or spinal cord in female rats exposed by inhalation. This equates to a unit risk of 8.3 × 10^–3 per mg/m³.

Although limited, available data are consistent with air being the principal medium of exposure of the general population to acrylonitrile; intake from other media is likely to be negligible in comparison. The focus of the human health risk characterization is populations exposed through air in the vicinity of industrial sources. Based on the margins between carcinogenic potency and limited available data on predicted and measured concentrations of acrylonitrile primarily in the vicinity of point sources in the sample risk characterization, risks in the vicinity of industrial point sources are >10^–5.

In the sample environmental risk characterization, levels in treated industrial wastewaters are less than the estimated no-effects value (ENEV) for the most sensitive aquatic organism, and predicted maximum levels (near a chemical industry processing plant) are less than the ENEV for the most sensitive terrestrial organism.

Acrylonitrile is also known as acrylic acid nitrile, acrylon, carbacryl, cyanoethylene, fumigrain, propenenitrile, 2-propenenitrile, propenoic acid nitrile, propylene nitrile, VCN, ventox, and vinyl cyanide. Its Chemical Abstracts Service (CAS) number is 107-13-1, its molecular formula, C₃H₃N, and its relative molecular mass, 53.06. The molecular structure of acrylonitrile is presented in Figure 1.

The physical and chemical properties of acrylonitrile are presented in Table 1. At room temperature, acrylonitrile is a volatile, flammable, colourless liquid with a weakly pungent odour (IPCS, 1983). Acrylonitrile has two chemically active sites, at the carbon–carbon double bond and at the nitrile group, where it undergoes a wide variety of reactions. It is a polar molecule because of the presence of the cyano (CN) group. It is soluble in water (75.1 g/litre at 25 °C) and miscible with most organic solvents. The vapours are explosive, with cyanide gas being produced.

Acrylonitrile may polymerize spontaneously and violently in the presence of concentrated caustic acid, on exposure to visible light, or in the presence of concentrated alkali (IPCS, 1983). Hence, it is stored accordingly, often as an acrylonitrile–water formulation that acts as a polymerization inhibitor (Kirk et al., 1983). Spontaneous polymerization during storage and transport can also be prevented by the addition of an inhibitor, typically hydroquinone methyl ether (NICNAS, 2000).

The most common method for analysis of acrylonitrile is gas chromatography. Method S156 of the US National Institute for Occupational Safety and Health (NIOSH, 1978, 1994) specifies sampling with activated charcoal sorption tubes, rinsing with methanol, and subsequent analysis by gas chromatography with a nitrogen–phosphorus detector. The working range of this method is 0.5–31 ppm (1–68 mg/m³) for a 15-litre sample. The estimated limit of detection is 0.02 ppm

(0.04 mg/m³). The US Occupational Safety and Health Administration specifies a similar method of sampling with charcoal tubes, desorption with acetone, and subsequent analysis with gas chromatography using a nitrogen–phosphorus detector. The detection limit for this method is 0.01 ppm (0.026 mg/m³) (OSHA, 1982, 1990). Health and Safety Executive (1993, 2000) specifies additional methods using porous polymer adsorption tubes and thermal desorption with gas chromatographic analysis.

A method has been developed for monitoring of exposure to acrylonitrile by determination of the adduct, N-(2-cyanoethyl)valine, formed in the reaction of acrylonitrile with the N-terminal group of haemoglobin (Bergmark et al., 1993; Osterman-Golkar et al., 1994; Tavares et al., 1996). The method is based on a modified Edman procedure and detection with selected ion monitoring by gas chromatography–mass spectrometry. The limit of detection is about 0.1–1 pmol/g globin (Tavares et al., 1996; Licea Perez et al., 1999).

Data on sources and emissions primarily from the source country of the national assessment on which the CICAD is based (i.e., Canada) are presented here as an example. Sources and patterns of emissions in other countries are expected to be similar, although quantitative values may vary.

Acrylonitrile is not known to occur naturally, and there are no known reactions that could lead to in situ formation of this substance in the atmosphere (Grosjean, 1990a).

The total release of acrylonitrile in Canada in 1996 was 19.1 tonnes (97.3% to air and 2.7% to water) (Environment Canada, 1997). The major source of releases was the organic chemicals industry (97.4%) (namely, the chemicals and chemical products industries and the plastic products industries), while municipal wastewater treatment facilities accounted for 2.6% of releases. Although accounting for at most 1% of the releases to air from chemical industries in Canada, incineration of sewage sludge represents another potential source of acrylonitrile release to air. Use of acrylonitrile polymers as conditioners for wastewater treatment is another potential source, although this is also considered to be minor in relation to industrial sources in Canada.

There is also potential for long-range transport of acrylonitrile (up to 2000 km from its source) based on its half-life in air of between 55 and 96 h (see Table 1).

Acrylonitrile has also been used in some countries as a pesticide. Its registration as a fumigant for stored grain in Canada ceased in 1976 (J. Ballantine, personal communication, 1997).

Environmental tobacco smoke is a potentially important source of acrylonitrile indoors (Miller et al., 1998).

World production of acrylonitrile exceeded 3.2 million tonnes in 1988, with later production increasing slowly (IARC, 1999). The estimated world capacity in 1991 was 4.2 million tonnes, while world demand in 1993 was 3.846 million tonnes (PCI, 1994). Major areas of production are the European Union (>1.25 million tonnes per year), the USA (approximately 1.5 million tonnes per year), and Japan (approximately 0.6 million tonnes per year). The large majority of acrylonitrile is used as a feedstock or chemical aid in the production of nitrile-butadiene rubber (68% of 1994 imports in the sample country) and in acrylonitrile-butadiene-styrene and styrene-acrylonitrile polymers (30% of 1994 imports in the sample country).

Acrylonitrile emitted to air reacts primarily with photochemically generated hydroxyl radicals (·OH) in the troposphere (Atkinson et al., 1982; Edney et al., 1982; Munshi et al., 1989; US DHHS, 1990; Bunce, 1996). The atmospheric half-life, based on hydroxyl radical reaction rate constants, is calculated to be between 4 and 189 h (Callahan et al., 1979; Cupitt, 1980; Edney et al., 1982; Howard, 1989; Grosjean, 1990b; Kelly et al., 1994). Modelling of environmental partitioning (section 5.5) is based on a mean half-life for acrylonitrile in air of 55 h.

The reaction of acrylonitrile with ozone and nitrate is slow, because of the absence of chlorine and bromine atoms in the molecule, and is not likely to constitute a major route of degradation (Bunce, 1996).

The reaction of hydroxyl radicals with acrylonitrile yields formaldehyde and, to a lesser extent, formic acid, formyl cyanide, carbon monoxide, and hydrogen cyanide (Edney et al., 1982; Spicer et al., 1985; Munshi et al., 1989; Grosjean, 1990a).

Acrylonitrile in water can be biodegraded by acclimatized microorganisms or volatilized (Going et al., 1979). In water, half-lives of 30–552 h are estimated based on aqueous aerobic biodegradation (Ludzack et al., 1961; Going et al., 1979; Howard et al., 1991). Modelling of environmental partitioning (section 5.5) is based on a mean half-life for acrylonitrile in water of 170 h (7 days). The half-life based on volatilization is 1–6 days (Howard et al., 1991). The hydrolysis of acrylonitrile is slow, with half-lives under acidic and basic conditions of 13 and 188 years, respectively (Ellington et al., 1987).

Acrylonitrile has an initial inhibitory effect on activated sludge systems and other microbial populations and does not meet the criteria of Organisation for Economic Co-operation and Development (OECD) Test Method 301C for ready biodegradability (Chemicals Inspection and Testing Institute of Japan, 1992; AN Group, 1996; BASF AG, 1996). However, acrylonitrile will be extensively degraded (95–100%) following a short acclimation period if emitted to wastewater treatment plants (Tabak et al., 1980; Kincannon et al., 1983; Stover & Kincannon, 1983; Freeman & Schroy, 1984; Watson, 1993).

Acrylonitrile is biodegraded in a variety of surface soils (Donberg et al., 1992) and by isolated strains of soil bacteria and fungi (Wenzhong et al., 1991). Concentrations of acrylonitrile up to 100 mg/kg were degraded in under 2 days (Donberg et al., 1992). Similar breakdown by microbial populations present in sediment is likely (DMER & AEL, 1996; EC, 2000). Results of experimental studies (Zhang et al., 1990) or soil sorption coefficients calculated by quantitative structure–activity relationships (Koch & Nagel, 1988; Walton et al., 1992) or based on water solubility (Kenaga, 1980) indicate little potential for adsorption of acrylonitrile to soil or sediments.

A half-life of acrylonitrile in soil of 6–7 days has been reported (Howard et al., 1991; Donberg et al., 1992) (see Table 1). Based on biodegradability and the soil partition coefficient (EC, 1996), the half-life of acrylonitrile in soil was classified in the category of 300 days (EC, 2000). Modelling of environmental partitioning (section 5.5) is based on a mean half-life for acrylonitrile in soil of 170 h (7 days). The half-life in the oxic zone of sediment can be assumed to be similar.

Bioaccumulation of acrylonitrile is not anticipated, given experimentally derived values of the octanol/water partition coefficient (log K_ow) ranging from -0.92 to 1.2 (mean 0.25) (Collander, 1951; Pratesi et al., 1979; Veith et al., 1980; Tonogai et al., 1982; Tanii & Hashimoto, 1984; Sangster, 1989) and a log bioconcentration factor (log BCF) of 0 calculated from the water solubility of acrylonitrile (EC, 2000).

Log BCF values were 0.48–1.68 in bluegill (Lepomis macrochirus) (Barrows et al., 1980) and rainbow trout (Oncorhynchus mykiss) (Lech et al., 1995). The experimentally derived log BCF of 1.68 reported by Barrows et al. (1980) in whole-body tissue of bluegill may be an overestimate, due to uptake of ¹⁴C-labelled degradation products in addition to acrylonitrile and to cyanoethylation of macromolecules (EC, 2000).

Fugacity modelling was conducted to characterize key reaction, intercompartment, and advection (movement out of a compartment) pathways for acrylonitrile and its overall distribution in the environment. A steady-state, non-equilibrium model (Level III fugacity model) was run using the methods developed by Mackay (1991) and Mackay & Paterson (1991). Assumptions, input parameters, and results are presented in DMER & AEL (1996) and summarized here. Values for input parameters were as follows: molecular mass, 53.06 g/mol; water solubility, 75.5 g/litre; vapour pressure, 11.0 kPa; log K_ow, 0.25; Henry’s law constant, 11 Pa·m³/mol; half-life in air, 55 h; half-life in water, 170 h; half-life in soil, 170 h; half-life in sediments, 550 h. Modelling was based on an assumed default emission rate of 1000 kg/h into a region of 100 000 km², which includes a surface water area (20 m deep) of 10 000 km². The height of the atmosphere was set at 1000 m. Sediments and soils were assumed to have an organic carbon content of 4% and 2% and a depth of 1 cm and 10 cm, respectively. The estimated percent distribution predicted by this model is not affected by the assumed emission rate.

Modelling indicates that when acrylonitrile is continuously discharged into a specific medium, most of it (84–97%) can be expected to be present in that medium (DMER & AEL, 1996). More specifically, Level III fugacity modelling by DMER & AEL (1996) predicts that:

• when released into air, the distribution of mass is 92.8% in air, 6.4% in water, 0.8% in soil, and 0.0% in sediment;
• when released into water, the distribution of mass is 2.5% in air, 97.3% in water, 0.0% in soil, and 0.1% in sediment; and
• when released into soil, the distribution of mass is 4.4% in air, 11.9% in water, 83.7% in soil, and 0.0% in sediment.

The major removal mechanisms in air, water, and soil are reaction within the medium and, to a lesser degree, advection and volatilization. Abiotic and biotic degradation in the various compartments result in low persistence overall and little, if any, bioaccumulation.

Owing to the paucity of data on concentrations of acrylonitrile in environmental media, fugacity modelling with version 4 of the ChemCAN3 model (Mackay et al., 1995) was also conducted with the conservative assumption that all known releases in 1996 (Environment Canada, 1997) in Canada occurred in southern Ontario. Release to air was considered to be approximately 19 tonnes per year, with simultaneous release to water of 0.53 tonnes per year. Since the half-life of acrylonitrile in air is the major determinant of its fate in the environment, the model was run using the minimum, median, and maximum half-life values (4, 55, and 189 h) under summer, winter, and year-round conditions. Modelling predicted distribution primarily to air (41.9–78.1%) and water (21.6–57.9%).

Data on concentrations in the environment primarily from the source country of the national assessment on which the CICAD is based (i.e., Canada) are presented here as a basis for the sample risk characterization. Patterns of exposure in other countries are expected to be similar, although quantitative values may vary. Detection of acrylonitrile in the general environment is restricted primarily to the vicinity of industrial sources.

6.1.1 Ambient air

Maximum predicted rates of emission of acrylonitrile during any half-hour period were 0.003, 0.018, and 0.028 g/s for stacks 14, 17, and 11 m high, respectively, near the site of the largest user in Canada (a Sarnia, Ontario, plant), based on dispersion modelling conducted in 1998 (H. Michelin, personal communication, 1999). Predicted concentrations at 11, 25, 41, and 1432 m from the stacks under the assumption that inversion occurs just above stack height and the plume is therefore forced to the ground were 6.6, 2.2, 0.4, and 0.1 µg/m³. Predicted concentrations at 11, 35, 41, and 3508 m under the assumption of close to stable or neutral atmospheric conditions were 9.3, 2.9, 0.6, and 0.1 µg/m³. Accuracy testing indicates that the model may overpredict actual values by as much as 2 orders of magnitude.

In six samples taken on 2 different days in the vicinity of a nitrile-butadiene rubber production plant in Sarnia, Ontario, 5 m outside the company fence line, 2 m above ground, and directly downwind of the stacks, acrylonitrile was not detected (detection limit 52.9 µg/m³) (B. Sparks, personal communication, 1997; M. Wright, personal communication, 1998).

Levels of acrylonitrile in ambient air sampled for 6 days near a chemical manufacturing plant in Cobourg, Ontario, ranged from 0.12 to 0.28 µg/m³. Measurements from stacks of the facility in 1993 ranged from <251 to 100 763 µg/m³ (Ortech Corporation, 1994). The concentration at point of impingement estimated on the basis of dispersion modelling of these data was 1.62 µg/m³.

At six urban stations in Ontario in 1990, concentrations of acrylonitrile in 10 of 11 samples were below the detection limit of 0.0003 µg/m³. In this study, the maximum and only detectable concentration of acrylonitrile was 1.9 µg/m³ in one sample (OMOE, 1992a).

Levels of acrylonitrile were <0.64 µg/m³ in all seven samples of ambient air taken in the industrialized area of Windsor, Ontario, in August 1991 (Ng & Karellas, 1994).

Ambient air samples were collected from downtown (n = 16) and residential (n = 7) areas of Metropolitan Toronto, Ontario, during a personal exposure pilot survey. The air samples were obtained at 1.5 m above ground for 12 consecutive hours. Acrylonitrile was not detected (detection limit 0.9 µg/m³) in any sample analysed (Bell et al., 1991).

Air samples were collected within the inhalation zone by a personal unit for 1–2 h while the participants were commuting to and from work (n = 19) and while they were spending the noon-hour period (n = 8) in downtown Toronto, Ontario, from June to August 1990. Acrylonitrile was not detected (detection limit 0.9 µg/m³) in any sample analysed. Acrylonitrile was also not detected (detection limit 0.9 µg/m³) in four special composite samples collected during the same study; the first two samples were collected while the participants were attending meetings, the third was collected while the participants were at a barbecue, and the fourth was an overall composite sample of the afternoon and morning commutes and the overnight residential indoor air (Bell et al., 1991).

6.1.2 Indoor air

Environmental tobacco smoke appears to be a source of acrylonitrile in indoor air (California Air Resources Board, 1994).

Acrylonitrile was not detected in samples collected overnight (duration up to 16 h) from June to August 1990 in four different residences near Toronto, Ontario (detection limit 0.9 µg/m³) (Bell et al., 1991).

6.1.3 Surface water and groundwater

Acrylonitrile has been detected only in water associated with industrial effluent; it has not been detected in ambient surface water in Canada (detection limit 4.2 µg/litre).

Of the effluents sampled from five companies using acrylonitrile and discharging to the environment in 1989–1990 in Ontario, the compound was detected in 12 of 256 samples (OMOE, 1993). Daily concentrations ranged from 0.7 to 3941 µg/litre; annual site averages ranged from 2.7 to 320 µg/litre. Intake water at the 26 organic chemical manufacturing plants sampled over the same period did not contain detectable amounts of acrylonitrile (207 samples; detection limit 4.2 µg/litre) (OMOE, 1992b). Biological treatment reactors have recently been introduced for two of the five companies with remaining commercial activity involving acrylonitrile, and levels from both sites are below 4.2 µg/litre (Y. Hamdy, personal communication, 1998).

In a large study of Canadian municipal water supplies in 1982–1983, acrylonitrile was not detected in any of the 42 raw (and 42 treated) water samples from nine municipalities on the Great Lakes (detection limit 5 µg/litre) (Otson, 1987). Acrylonitrile was not detected (detection limit 2.1 µg/litre) in groundwater samples downgradient of a wastewater treatment pond at an Ontario chemical industry site (Environment Canada, 1997).

6.1.4 Drinking-water

Acrylonitrile was monitored in municipal water supplies at 150 locations in Newfoundland, Nova Scotia, New Brunswick, and Prince Edward Island over the period 1985–1988. It was detected at a trace concentration (0.7 µg/litre) in only one sample of treated water in Nova Scotia in June 1988 (detection limit 0.5–1.0 µg/litre) (Environment Canada, 1989a,b,c,d).

Acrylonitrile was not identified in treated (or raw) water at facilities near the Great Lakes in 1982–1983 (n = 42; detection limit 5 µg/litre during the initial sampling and <1 µg/litre during later sampling after the technique was modified) over three sampling periods (Otson, 1987). Analyses were by gas chromatography–mass spectrometry.

6.1.5 Soil and sediment

Significant concentrations of acrylonitrile are not expected in soil or sediment based on the release patterns and the environmental partitioning, behaviour, and fate of the substance (section 5.3).

Significant levels of acrylonitrile have not been detected in Canadian soils. Levels in 18 soil samples at an Alberta chemical blending plant were below the detection limit of 0.4 ng/g (G. Dinwoodie, personal communication, 1993). Significant quantities of acrylonitrile in soil at a LaSalle, Quebec, chemical industrial site have not been identified since regular monitoring began at the site in 1992 (Environment Canada, 1997).

Data on levels of acrylonitrile in sediment have not been identified.

6.1.6 Food

Acrylonitrile may potentially be introduced into foodstuffs from acrylonitrile-based polymers used in food packaging. Page & Charbonneau (1983) measured concentrations of acrylonitrile in five types of food packaged in acrylonitrile-based plastic containers, purchased from several stores in Ottawa, Ontario. Average concentrations of acrylonitrile (measured in three duplicate samples of each food type by gas chromatography with a nitrogen–phosphorus-selective detector) ranged from 8.4 to 38.1 ng/g.

A survey of food packed in acrylonitrile-based plastics containing up to 2.6 mg acrylonitrile/kg was conducted in Ottawa, Ontario. The samples represented five food companies and a variety of luncheon meats, including mock chicken, ham, salami, pizza loaf, and several types of bologna. Acrylonitrile was not identified (detection limit 2 ng/g). Analyses were by gas chromatography, with nitrogen–phosphorus-selective detection (Page & Charbonneau, 1985).

6.1.7 Multimedia study

In a multimedia study carried out for Health Canada (Conor Pacific Environmental & Maxxam Ltd., 1998), exposure to several volatile organic chemicals, including acrylonitrile, was measured for 50 participants across Canada. Thirty-five participants were randomly selected from the Greater Toronto Area in Ontario, six participants from Liverpool, Nova Scotia, and nine from Edmonton, Alberta. For each participant, samples of drinking-water, beverages, and indoor, outdoor, and personal air were collected over a 24-h period. Acrylonitrile was not detected in air (detection limit 1.36 µg/m³), water (detection limit 0.7 ng/ml), beverages (detection limit 1.8 ng/ml), or food (detection limit 0.5 ng/g).

Point estimates of average daily intake (per kilogram body weight) were developed for the sample country (i.e., Canada), based on the few monitoring data available and reference values for body weight, inhalation volume, and amounts of food and drinking-water consumed daily by six age groups (Environment Canada & Health Canada, 2000). Similar estimates were developed based on the results of the ChemCAN3 fugacity modelling (Environment Canada & Health Canada, 2000). In view of the limitations of the data on which these estimates are based, however (i.e., lack of detection of acrylonitrile in most media in which it was monitored or fugacity modelling), these estimates are primarily useful as a basis for identification of principal routes and media of exposure.

Although it is uncertain, based on this limited information, air (ambient and indoor) is likely the principal medium of exposure. Intakes from food and drinking-water are likely to be negligible in comparison. This is consistent with the physical/chemical properties of acrylonitrile, which has moderate vapour pressure and a low log K_ow, and the results of fugacity modelling (section 5.5). On the basis of the estimates derived above, intake from ambient and indoor air ranges from 96% to 100% of total intake.

Exposures from ambient air may be substantially higher for populations in the vicinity of point sources. Data presented above (see section 6.1.1) on concentrations in the vicinity of point sources indicate that populations in the area might be exposed to levels of acrylonitrile in the range of tenths of µg/m³ (Ng & Karellas, 1994; Ortech Corporation, 1994). Additional data from the USA indicate that levels vary considerably in the vicinity of various point sources (Health Canada, 2000).

Limitations of the data preclude development of meaningful probabilistic estimates of exposure to acrylonitrile in the general population.

Exposure to acrylonitrile may occur during its production and its use in the manufacture of other products; potential for exposure is greatest in factories where acrylonitrile is used to make other products, where it may not be as easily contained (Sax, 1989). IARC (1999) indicates that approximately 35 000 workers in Europe and as many as 80 000 workers in the USA are potentially exposed to acrylonitrile. These individuals include acrylic resin makers, synthetic organic chemists, pesticide workers, and rubber, synthetic fibre, and textile makers. The primary routes of potential exposure in the occupational environment are inhalation and dermal.

Based on data collected in 1995 for countries in the European Union (EC, 2000), 8-h time-weighted-average (TWA) exposures for production and various end uses were as follows: production, <0.45 ppm (1 mg/m³); fibre, <1.01 ppm (<2.2 mg/m³); latex, <0.10 ppm (<0.22 mg/m³); acrylonitrile-butadiene-styrene polymer, <0.40 ppm (<0.88 mg/m³); and acrylamide, <0.20 ppm (<0.44 mg/m³).

For six European producers of acrylonitrile, average personal monitoring levels at the workplace varied from <0.12 to 0.49 ppm (<0.26 to 1.1 mg/m³); the maximum recorded level was 5.5 ppm (12.1 mg/m³). For production of acrylonitrile fibres, the average personal monitoring levels were from <0.26 to 0.43 ppm (<0.57 to 0.95 mg/m³), with a maximum value of 3.6 ppm (7.9 mg/m³). For production of acrylonitrile-butadiene-styrene polymers, the average levels for personal monitoring were 0.08–0.3 ppm (0.18–0.66 mg/m³), with a maximum recorded value of 8.6 ppm (19.0 mg/m³). The higher levels in use were consistent with production of acrylonitrile being initially in a closed system, while manufacture of, for example, acrylonitrile-butadiene-styrene polymers is carried out in a partially closed system, with local exhaust ventilation and emission (EC, 2000).

Recent information on the occupational exposure scenario for Australia (NICNAS, 2000) correlates well with the above data. Australia imports approximately 2000 tonnes of acrylonitrile per year, 70% of which is used for the manufacture of styrene-acrylonitrile polymer, which is further compounded into plastic resins. The remainder is used in the manufacture of latex polymers (polymers dispersed in water) for adhesive and coating applications. Of 187 breathing-zone air samples collected in 1991–1999 during normal operations, 68% were <0.1 ppm (<0.22 mg/m³), 95% <0.5 ppm (<1.1 mg/m³), and 97% <1 ppm (<2.2 mg/m³), expressed as 8-h TWAs. Personal exposure levels were slightly higher in latex than in styrene-acrylonitrile polymer and plastic resin manufacturing plants.

Surveys of full-shift personal exposures in four US acrylonitrile production plants have also been reported (Zey et al., 1989, 1990a,b; Zey & McCammon, 1990). Mean 8-h TWA personal exposures for monomer production operators were 1.1 ppm (2.4 mg/m³) or less from about 1978 to 1986, with some individual TWA levels up to 37 ppm (82 mg/m³). In three of these plants, levels for maintenance employees averaged below 0.3 ppm (0.7 mg/m³), but in one plant, the average TWA exposure for these workers was 1 ppm (2.2 mg/m³). Average 8-h TWA exposures for loaders of acrylonitrile into tank trucks, rail cars, or barges varied from 0.5 to 5.8 ppm (1.1 to 12.8 mg/m³). For some of the higher values for production and maintenance workers and loaders in these plants, respirator use was noted. Although there were several changes introduced to reduce exposure levels, no trends over the years were noted.

In three US fibre plants for which data for full-shift personal samples were available between the years 1977 and 1986, the average 8-h TWA exposures were between 0.3 and 1.5 ppm (0.7 and 3.3 mg/m³), based on nearly 3000 individual samples. The dope (viscous pre-fibre solution) and spinning operators had exposures averaging 0.4–0.9 ppm (0.9–2.0 mg/m³). The lower exposure occurred in the plant that dried the polymer before the spinning operation, resulting in a lower monomer content in the polymer. The other plant had a continuous wet operation without the drying stage. Exposure of maintenance workers averaged 0.2 ppm (0.4 mg/m³). Tank farm operators, who are also likely to unload acrylonitrile monomer from trucks, rail cars, or barges, had homogeneous exposure levels (0.5 ppm [1.1 mg/m³]) across plants (IARC, 1999).

Haemoglobin adduct measurement is an extremely sensitive method for monitoring exposure to acrylonitrile. The level of N-(2-cyanoethyl)valine reflects exposure during a 4-month period (i.e., the life span of the erythrocytes) prior to blood sampling. Licea Perez et al. (1999) reported levels of 0.76 ± 0.36 pmol/g globin in 18 non-smokers (who claimed that they were not exposed to environmental tobacco smoke). Adduct levels in smokers range from 8 to a few hundred pmol/g globin and are related to cigarette consumption. Adduct levels in 10 smoking mothers (92.5–373 pmol/g globin) and their newborns (34.6–211 pmol/g globin) were strongly correlated and demonstrated that transplacental transfer of acrylonitrile occurs (Tavares et al., 1996). Adduct levels ranging from 20 to 66 000 pmol/g have been observed in occupationally exposed workers (Bergmark et al., 1993; Tavares et al., 1996; Thier et al., 1999). Polymorphism with respect to the glutathione transferases GSTTI and GSTMI had little effect on adduct levels (Thier et al., 1999; Fennell et al., 2000).

Based on studies conducted primarily in laboratory animals, acrylonitrile is rapidly absorbed by all routes of administration and distributed throughout examined tissues. In inhalation studies with volunteers, 50% of acrylonitrile was absorbed (Jakubowski et al., 1987). However, there appears to be little potential for significant accumulation in any organ, with most of the compound being excreted primarily as metabolites in urine in the first 24–48 h following administration (Kedderis et al., 1993a; Burka et al., 1994).

Acrylonitrile is metabolized primarily by two pathways: conjugation with glutathione to form N-acetyl-S-(2-cyanoethyl)cysteine and oxidation by cytochrome P-450 to form remaining urinary metabolites (Langvardt et al., 1980; Geiger et al., 1983; Fennell et al., 1991; Kedderis et al., 1993a) (Figure 2). Oxidative metabolism of acrylonitrile leads to the formation of 2-cyanoethylene oxide, which is either conjugated with glutathione (Fennell & Sumner, 1994; Kedderis et al., 1995) to form a series of metabolites including cyanide and thiocyanate or directly hydrolysed by epoxide hydrolase (Borak, 1992; Kim et al., 1993). Recent data indicate that cytochrome P4502E1 is the sole P-450 catalysing the oxidation of acrylonitrile (Sumner et al., 1999).

Available data³ are consistent with conjugation with glutathione being the major detoxification pathway of acrylonitrile, while the oxidation of acrylonitrile to 2-cyanoethylene oxide can be viewed as an activation pathway, producing a greater proportion of the total metabolites in mice than in rats. Available data also indicate that there are route-specific variations in metabolism (Lambotte-Vandepaer et al., 1985; Tardif et al., 1987). Based on studies in which 2-cyanoethylene oxide has been administered, there is no indication of preferential uptake or retention in specific organs, including the brain (Kedderis et al., 1993b).

Liver microsomes from rats, mice, and humans produced 2-cyanoethylene oxide at a greater rate than lung or brain microsomes, indicating that the liver is the major site of formation in vivo of 2-cyanoethylene oxide (Roberts et al., 1989; Kedderis & Batra, 1991). Studies in subcellular hepatic fractions indicate that there is an active epoxide hydrolase pathway for 2-cyanoethylene oxide in humans, which is inactive, although inducible, in rodents (Kedderis & Batra, 1993). Studies with inhibitory antibodies in human hepatic microsomes indicate that the 2E1 isoform of cytochrome P-450 is primarily involved in epoxidation of acrylonitrile (Guengerich et al., 1991; Kedderis et al., 1993c).

A physiologically based pharmacokinetic model has been developed and verified for the rat (Gargas et al., 1995; Kedderis et al., 1996), and work is under way to scale it to humans. In a recent, although incompletely reported, study, Kedderis (1997) estimated in vivo activity of epoxide hydrolase in humans based on the ratio of epoxide hydrolase to P-450 activity in subcellular hepatic fractions multiplied by the P-450 activity in vivo. Human blood to air coefficients for acrylonitrile and 2-cyanoethylene oxide have also been recently determined, although incompletely reported at present (Kedderis & Held, 1998). Research is in progress to determine partition coefficients for other human tissues.

The acute toxicity of acrylonitrile is relatively high, with 4-h LC₅₀s ranging from 140 to 410 ppm (300 to 900 mg/m³) (Knobloch et al., 1971, 1972) and oral LD₅₀s ranging from 25 to 186 mg/kg body weight (Maltoni et al., 1987). Dermal LD₅₀ values for various species were in the range of 148–693 mg/kg body weight, with the rat being most sensitive (BUA, 1995). Signs of acute toxicity include respiratory tract irritation and central nervous system dysfunction, resembling cyanide poisoning. Superficial necrosis of the liver and haemorrhagic gastritis of the forestomach have also been observed following acute exposure (Silver et al., 1982).

Acrylonitrile-induced neurotoxicity following acute exposure via inhalation or ingestion has been described as a two-phase phenomenon. The first phase, which occurs shortly after exposure and is consistent with cholinergic overstimulation, has been likened to toxicity caused by acetylcholinesterase inhibition. Cholinomimetic signs in rats exposed to acrylonitrile have included vasodilation, salivation, lacrimation, diarrhoea, and gastric secretion. These effects are maximal within 1 h of dosing. The second phase of toxicity is delayed by 4 h or more and includes signs of central nervous system disturbance, such as trembling, ataxia, convulsions, and respiratory failure (TERA, 1997).

Available data indicate that acrylonitrile is a skin, respiratory, and severe eye irritant. It induces skin sensitization in guinea-pigs, but available data are inadequate to assess its sensitization potency.

8.2.1 Skin irritancy

Identified data on irritancy to the skin are restricted to three studies in which acrylonitrile was applied to the shaved skin of rabbits (McOmie, 1949; Zeller et al., 1969; Vernon et al., 1990). In early periods following administration (e.g., 15 min), slight local vasodilation and oedema were observed; at longer periods (20 h), effects were more severe, with necrosis reported.

8.2.2 Eye irritancy

Identified investigations of eye irritancy are restricted to principally early unpublished studies (McOmie, 1949; BASF, 1963; Zeller et al., 1969; DuPont, 1975). Results of these investigations were consistent with those reported in the most recent study conducted by DuPont (1975), in which 0.1 ml of undiluted acrylonitrile was placed in the right conjunctival sac of each of two albino rabbits. After 20 s, the treated eye of one rabbit was washed with tap water for 1 min. Acrylonitrile produced moderate corneal opacity, moderate iritis, and severe conjunctival irritation in the unwashed treated eye. In the washed eye, there was slight temporary corneal opacity, transient, moderate iritic congestion, and moderate conjunctival irritation. These effects were not completely reversible in the unwashed eye; by washing the eye, the effects were considerably lessened, as was the duration of these ocular effects.

8.2.3 Respiratory tract irritancy

While not examined directly, in repeated-dose toxicity studies, there has been evidence of irritant effects on the upper respiratory tract, including nasal discharge in rats acutely exposed (Food and Drug Research Laboratories, 1985) and rhinitis and hyperplastic changes in the nasal mucosa following chronic exposure (Quast et al., 1980b).

8.2.4 Sensitization

In a guinea-pig maximization test, animals challenged with 0.5% and 1.0% acrylonitrile had a 95% positive sensitization rate. Exposure to 0.2% on challenge caused an 80% sensitization rate (Koopmans & Daamen, 1989).

Available short-term inhalation studies are restricted to a few investigations involving administration of single dose levels and, for one, examination of clinical signs only. Exposure–response has not, therefore, been well characterized. There were effects on biochemical parameters, clinical signs, and body weight, although no histopathological effects on principal organs, following exposure of rats to 130 ppm (280 mg/m³) acrylonitrile (Gut et al., 1984, 1985).

In short-term studies by the oral route, effects on the liver, adrenal, and gastric mucosa have been observed, with effects on the gastric mucosa occurring at lowest doses in all studies in which they were examined. Effects on the adrenal cortex observed in short-term repeated-dose toxicity studies from one laboratory have not been noted in longer-term investigations in animals exposed to higher concentrations. In investigations by Szabo et al. (1984), effects on the non-protein sulfhydryl content in gastric mucosa and hyperplasia in the adrenal cortex have been reported at levels as low as 2 mg/kg body weight per day administered by drinking-water and gavage, respectively, for 60 days. Effects on hepatic glutathione were also observed by these authors at similar doses administered by gavage but not in drinking-water (2.8 mg/kg body weight per day for 21 days), although Silver et al. (1982) noted only slight biochemical effects but no histopathological effects in the liver at doses up to 70 mg/kg body weight per day (drinking-water, 21 days). Significant increases in proliferation in the forestomach but no changes in the liver or glandular stomach have been observed at 11.7 mg/kg body weight (Ghanayem et al., 1995, 1997).

Gastric lesions in the rat have been accompanied by a decrease of gastric reduced glutathione concentration. It has been suggested that depletion and/or inactivation of critical endogenous sulfhydryl groups cause configurational changes of cholinergic receptors and increase agonist binding affinity, which may lead to gastric mucosal erosion (Ghanayem et al., 1985; Ghanayem & Ahmed, 1986).

Effects of pretreatment with inducers of the mixed-function oxidase system or antioxidants on toxicity in short-term studies have been consistent with metabolism to the epoxide 2-cyanoethylene oxide being the putatively toxic metabolic pathway (Szabo et al., 1983).

Results of identified subchronic toxicity studies are limited to an early 13-week inhalation study in rats and dogs that has not been validated (IBT, 1976) and a preliminary brief report of the results of a 13-week National Toxicology Program (NTP) gavage study in mice.⁴ Lack of validation and inadequate detail limit the utility of these studies for hazard evaluation or characterization of dose–response.

Data on effects of long-term exposure to acrylonitrile are currently restricted to those conducted in rats, although an NTP bioassay in mice is under way (NTP, 1998). In the descriptions of the following studies, tumour types are reported as described by the authors. However, it should be noted that the histopathology of the tumours may be unclear (see footnote in section 8.5.2).

8.5.1 Inhalation

Quast et al. (1980b) conducted a bioassay in which Sprague-Dawley (Spartan substrain) rats (100 per sex per group) were exposed by inhalation to average concentrations of 0, 20, or 80 ppm (0, 44, or 176 mg/m³) acrylonitrile 6 h/day, 5 days/week, for 2 years. Non-neoplastic histopathological changes related to the treatment were present in the nasal turbinates and the central nervous system of both males and females. In the brain, the changes were characterized by focal gliosis and perivascular cuffing at the highest concentration. The inflammatory changes in the nasal turbinates were considered to be due to acrylonitrile irritation. These effects were not observed at 20 ppm (44 mg/m³), and this dose is considered as a no-observed-effect level (NOEL) for inflammatory changes in the nasal turbinates. There was an early onset of chronic renal disease in the group exposed to 20 ppm (44 mg/m³) based upon histopathological examination. The renal effect was not apparent at the high dose because of early mortality. The chronic renal disease, which is commonly observed in older rats of this strain, was considered a secondary effect caused by increased water intake, although there was no pair-fed control study, and clinical analyses were inadequate to confirm the cause. In males, mortality at 80 ppm (176 mg/m³) was consistently and significantly increased from days 211–240 to the end of the experiment. Similar findings were noted for females beginning at days 361–390.

In both sexes, there was an increase in the combined incidence of malignant and benign tumours of the brain and spinal cord (Table 2) and benign and malignant tumours of the Zymbal gland at the high dose. In males, the combined incidence of benign and malignant tumours of the small intestine and the tongue was increased at the high dose. The incidence of adenocarcinoma of the mammary gland was increased at the high dose in females (Quast et al., 1980b).

Although Maltoni et al. (1977) reported an increased incidence of tumours in the mammary gland, forestomach, and skin in Sprague-Dawley rats exposed to up to 40 ppm (88 mg/m³) for 52 weeks, low concentrations of acrylonitrile, short exposure time, and small group size (n = 30) limit the sensitivity of the study. In a follow-up study (Maltoni et al., 1987, 1988), 54 female Sprague-Dawley rat breeders and offspring were administered 60 ppm (132 mg/m³) by inhalation for 4–7 h/day, 5 days/week. The breeders and some of the offspring were exposed for 104 weeks, and the remaining offspring were exposed for 15 weeks only. The non-neoplastic treatment-related changes included slight, but significant, increases in the incidence of encephalic glial cell hyperplasia and dysplasia in offspring exposed for 104 weeks. The incidence of various tumours was increased in the exposed offspring, both males and females. These included mammary gland tumours in females, Zymbal gland tumours in males, extrahepatic angiosarcoma in both males and females, hepatomas in males, and encephalic gliomas in both males and females. The most pronounced acrylonitrile-related tumour was encephalic glioma (in control and exposure groups, respectively: 2/158 and 11/67 in males; 2/149 and 10/54 in females) in the offspring treated with acrylonitrile for 104 weeks.

8.5.2 Drinking-water

Quast et al. (1980a) administered acrylonitrile in drinking-water to Sprague-Dawley rats for 2 years at dose levels of 0, 35, 100, or 300 mg/litre. There was treatment-related hyperplasia and hyperkeratosis of the squamous epithelium of the forestomach in females at all dose levels and in males at 100 and 300 mg/litre. In the brain of females, there was a significantly increased incidence of focal gliosis and perivascular cuffing in the 35 and 100 mg/litre groups. Tumours (including astrocytomas) were observed as early as 7–12 months in females in the high-dose group; in other dose groups, tumours appeared initially in the 13- to 18-month period. In both males and females, the combined incidence of benign and malignant tumours of the brain and spinal cord was significantly increased in a dose-related manner at all levels of exposure (Table 3).

In a study conducted by Bio/Dynamics Inc. (1980a), Sprague-Dawley rats were administered acrylonitrile at dose levels of 0, 1, or 100 mg/litre in drinking-water for 19 and 22 months. Non-neoplastic effects included increased weight of kidney and testes. The concentration of 1 mg/litre can be considered as a NOEL and 100 mg/litre as a lowest-observed-adverse-effect level (LOAEL) for non-neoplastic effects. In high-dose males, increased incidences of squamous cell carcinoma of the stomach and carcinoma of the Zymbal gland were observed. In high-dose females, astrocytoma of the brain and carcinoma of the Zymbal gland were increased. At the high dose, there was an increased cumulative incidence of astrocytoma of the brain, carcinoma of the Zymbal gland, and papilloma/carcinoma of the stomach in both sexes. In females, the incidence of astrocytoma of the spinal cord was significantly increased at the high dose. The spinal cord tissue of the males was not examined. Dose spacing was poor in this study.

These results are consistent with those of a second bioassay by Bio/Dynamics Inc. (1980b), in which exposure–response was better characterized. Fischer 344 rats (200 per sex, control group; 100 per sex per dose group) were administered acrylonitrile in drinking-water for approximately 2 years. The dose levels were 0, 1, 3, 10, 30, and 100 mg acrylonitrile/litre (0, 0.1, 0.3, 0.8, 2.5, and 8.4 mg/kg body weight per day for males and 0, 0.1, 0.4, 1.3, 3.7, and 10.9 mg/kg body weight per day for females, as reported by US EPA, 1985). Serial sacrifices were conducted at 6, 12, and 18 months (20 per sex per control group and 10 per sex per treated group). To ensure at least 10 rats per sex per group for histopathological evaluation, all females were sacrificed at 23 months, owing to low survival. The males were continued on test until the 26th month.

The consistently elevated mortality in the highest dose groups was primarily a consequence of tumours. Other changes observed primarily in the highest exposure group included consistently lower body weights in females and males and consistent reduction in haemoglobin, haematocrit, and erythrocyte counts in females throughout the study. A decrease in water intake was also observed, while food consumption was comparable for all groups (Bio/Dynamics Inc., 1980b).

An increase in the relative organ weights of the liver and kidney was noted at the highest dose levels; however, the mean absolute weights for these organs were either comparable to those in the controls or only slightly increased. At terminal sacrifice, the absolute liver and heart weights were elevated in females exposed to 30 mg/litre, but body weight was comparable to that in controls. The LOAEL is considered 100 mg/litre, the lowest-observed-effect level (LOEL), 30 mg/litre, and the NOEL for non-carcinogenic effects, 10 mg/litre. In both males and females, the incidences of astrocytoma of the brain (Table 4) and of carcinoma of the Zymbal gland were significantly increased at the two highest dose levels (Bio/Dynamics Inc., 1980b).

In a multigeneration reproductive study, 0, 100, or 500 mg acrylonitrile/litre (0, 14, or 70 mg/kg body weight per day; Health Canada, 1994) was administered in drinking-water to breeders (F₀) and the offspring of Charles River Sprague-Dawley rats (Litton Bionetics Inc., 1980). Rats of the F_1b generation in the high-exposure group had a significantly increased incidence of astrocytomas and Zymbal gland tumours. For control, low-exposure, and high-exposure groups, the incidence of astrocytomas was 0/20, 1/19, and 4/17 (P < 0.05), respectively, and the incidence of Zymbal gland tumours was 0/20, 2/19, and 4/17 (P < 0.05), respectively. The tumour incidence was low, but the exposure and observation period (approximately 45 weeks) was also relatively short. Not all tissues were examined histopathologically.

More recently, Bigner et al. (1986) observed neuro-oncogenic effects in Fischer 344 rats administered 0, 100, or 500 mg acrylonitrile/litre in drinking-water (0, 14, and 70 mg/kg body weight per day; Health Canada, 1994). Each exposure group consisted of 50 male and 50 female rats. A fourth group of 300 rats (147 males, 153 females) was exposed to 500 mg acrylonitrile/litre. Although the protocol of the study indicated that rats were exposed for their lifetime, results were presented for an 18-month observation period. There was a dose-related significant reduction in body weight in both males and females at 500 mg/litre. In rats exposed for 12–18 months, neurological signs such as decreased activity, paralysis, head tilt, circling, and seizures were observed in the 100 and 500 mg/litre groups. In control, low-exposure, and two high-exposure groups, the incidence of neurological signs was 0/100, 4/100, 16/100, and 29/300, respectively. Based on histopathological examination of 215 animals in the 500 mg/litre group, there were 49 primary brain tumours, which were difficult to classify.⁵ Other tumours frequently observed included Zymbal gland tumours, forestomach papillomas, and subcutaneous papillomas. No further details, however, were presented. The authors reported that the increase in incidence of the primary brain tumour in the highest exposure group was significant (P-values were not reported, data were poorly presented). Other end-points were not examined. The results are inadequate, therefore, for establishing effect levels for non-neoplastic effects or for characterizing exposure–response for tumours.

Gallagher et al. (1988) investigated the carcinogenicity of acrylonitrile administered via drinking-water at 0, 20, 100, or 500 mg/litre (approximately 0, 2.8, 14, and 70 mg/kg body weight per day; Health Canada, 1994) to male Sprague-Dawley rats (20 per group) for 2 years. There was no survival in the 500 mg/litre exposure group at 2 years. Ingestion of acrylonitrile at concentrations up to and including 100 mg/litre did not increase mortality. There was a significant increase in Zymbal gland tumours at 500 mg/litre (0/18, 0/20, 1/19, and 9/18 [P < 0.005] in control, low-, mid-, and high-dose groups, respectively). No increase in tumours of other organs including brain was observed, although four rats developed papillomatous proliferation of the epithelium of the forestomach in the high-exposure group.

8.5.3 Gavage

Groups of 100 male and female Sprague-Dawley rats were exposed in a Bio/Dynamics Inc. (1980c) study to acrylonitrile in water by intubation at 0, 0.1, or 10 mg/kg body weight per day for 20 months. The non-neoplastic effects in the high-dose group included higher mortality (both sexes), decreased body weight (males), and increased relative liver weight (males). The dose of 10 mg/kg body weight per day is a LOAEL, based upon decreased body weight and increased liver to body weight ratio in males, with 0.1 mg/kg body weight being the NOEL. In both sexes at the high dose, there was an increased incidence of astrocytoma of the brain, squamous cell carcinoma of the Zymbal gland, and papilloma/carcinoma of the stomach.

Maltoni et al. (1977) exposed 40 Sprague-Dawley rats of each sex by gavage to acrylonitrile in olive oil at 0 or 5 mg/kg body weight per day, 3 days/week for 52 weeks. In females, the incidence of mammary gland carcinomas was 7/75 and 4/40 in control and exposed groups, respectively; the incidence of forestomach epithelial tumours was 0/75 and 4/40 in control and exposed groups, respectively. However, a high spontaneous incidence of mammary gland tumours in this strain of rats, the single dose level, and the short duration of exposure limit contribution of the study to characterization of exposure–response.

8.6.1 In vitro studies

In the Salmonella assay, acrylonitrile has induced reverse mutations in strains TA1535 (Lijinsky & Andrews, 1980), TA1535, and TA100 (Zeiger & Haworth, 1985), but only when hamster or rat S9 was present. Weak positive results were also reported in several Escherichia coli strains in the absence of metabolic activation (Venitt et al., 1977).

In mammalian cells, acrylonitrile induced hprt mutations in human lymphoblasts without metabolic activation (Crespi et al., 1985), but not in Chinese hamster V79 cells (Lee & Webber, 1985). In several studies, acrylonitrile was positive at the TK locus in mouse lymphoma L5178 TK+/- cells, either with or without rat S9 (Amacher & Turner, 1985; Lee & Webber, 1985; Myhr et al., 1985; Oberly et al., 1985), and in mouse lymphoma P388F cells with metabolic activation (Anderson & Cross, 1985). It was also mutagenic at the TK locus in human lymphoblasts with metabolic activation (Crespi et al., 1985; Recio & Skopek, 1988).

Acrylonitrile induced structural chromosomal aberrations either with or without metabolic activation in Chinese hamster ovary cells (Danford, 1985; Gulati et al., 1985; Natarajan et al., 1985) and without metabolic activation in Chinese hamster lung cells (Ishidate & Sofuni, 1985). Results for sister chromatid exchanges in Chinese hamster ovary cells and human lymphocytes both with and without metabolic activation are mixed (Brat & Williams, 1982; Perocco et al., 1982; Gulati et al., 1985; Natarajan et al., 1985; Obe et al., 1985; Chang et al., 1990).

Results of in vitro assays for DNA single strand breaks (Bradley, 1985; Lakhanisky & Hendrickx, 1985; Bjorge et al., 1996) and DNA repair (unscheduled DNA synthesis) (Perocco et al., 1982; Glauert et al., 1985; Martin & Campbell, 1985; Probst & Hill, 1985; Williams et al., 1985; Butterworth et al., 1992) were mixed but more commonly negative in a range of cell types from rats and humans, with and without activation. Cell transformation in mouse and hamster embryo cells has also been investigated, with mixed results (Lawrence & McGregor, 1985; Matthews et al., 1985; Sanner & Rivedal, 1985; Abernethy & Boreiko, 1987; Yuan & Wong, 1991).

Binding of 2-cyanoethylene oxide to nucleic acids has also been reported in in vitro studies at high concentrations (Hogy & Guengerich, 1986; Solomon & Segal, 1989; Solomon et al., 1993; Yates et al., 1993, 1994⁶). The formation of acrylonitrile–DNA adducts is increased substantially in the presence of metabolic activation. Under non-activating conditions involving incubation of calf thymus DNA with either acrylonitrile or 2-cyanoethylene oxide in vitro, 2-cyanoethylene oxide alkylates DNA much more readily than acrylonitrile (Guengerich et al., 1981; Solomon et al., 1984, 1993). Incubation of DNA with 2-cyanoethylene oxide yields 7-(2-oxoethyl)-guanine (Guengerich et al., 1981; Hogy & Guengerich, 1986; Solomon & Segal, 1989; Solomon et al., 1993; Yates et al., 1993, 1994) as well as other adducts. Compared with studies with rat liver microsomes, little or no DNA alkylation by acrylonitrile was observed with rat brain microsomes (Guengerich et al., 1981). DNA alkylation in human liver microsomes was much less than that observed with rat microsomes (Guengerich et al., 1981); although there was no glutathione S-transferase activity in cytosol preparations from human liver exposed to acrylonitrile, there was some activity for 2-cyanoethylene oxide (Guengerich et al., 1981).

8.6.2 In vivo studies

Limitations of the few in vivo studies conducted in which the genotoxicity of acrylonitrile has been investigated preclude definitive conclusions. Data from these studies are also inadequate for characterization of dose–response for comparison between studies or with the cancer bioassays.

Exposure to acrylonitrile in drinking-water resulted in increased frequency of mutants at the hprt locus in splenic T-cells (Walker & Walker, 1997). Five female F344 rats were exposed to 0, 33, 100, or 500 mg/litre (0, 8, 21, or 76 mg/kg body weight per day; Health Canada, 1994) in drinking-water for up to 4 weeks and serially sacrificed throughout exposure and up to 8 weeks post-exposure. At 4 weeks post-exposure, the average observed mutant frequency in splenic T-cells was increased in a dose-related manner (significant at the two highest doses).

Results of a range of assays for structural chromosomal aberrations, micronuclei in bone marrow, and micronuclei in peripheral blood cells have been negative or inconclusive, although there was no indication in the published accounts of three of the four studies that the compound reached the target site. These include studies in Swiss (Rabello-Gay & Ahmed, 1980), NMRI (Leonard et al., 1981), and C57B1/6 (Sharief et al., 1986) mice and a collaborative study following exposure by multiple routes in mice and rats (Morita et al., 1997).

Results of dominant lethal assays were inconclusive in mice (Leonard et al., 1981) and negative in rats (Working et al., 1987).

In assays for unscheduled DNA synthesis in rats, results were positive only for the liver (Hogy & Guengerich, 1986), equivocal in lung, testes, and gastric tissues (Ahmed et al., 1992a,b; Abdel-Rahman et al., 1994), and, notably, negative in the brain (Hogy & Guengerich, 1986). In these studies, however, unscheduled DNA synthesis was measured by liquid scintillation counting to determine [³H]thymidine uptake in the cell population, which does not discriminate between cells undergoing repair and those that are replicating. Results for unscheduled DNA synthesis in rat liver and spermatocytes were negative when [³H]thymidine uptake in individual cells was determined by autoradiography, which eliminates replicating cells from the analysis (Butterworth et al., 1992).

Urine from acrylonitrile-exposed rats and mice was also mutagenic in Salmonella typhimurium following intraperitoneal administration of acrylonitrile to rats and mice (Lambotte-Vandepaer et al., 1980, 1981). In both species, mutagenic activity occurred without activation. Mutagenic activity was also observed in urine of rats administered acrylonitrile by stomach intubation (Lambotte-Vandepaer et al., 1985). Thiocyanate, hydroxyethylmercapturic acid, and cyanoethylmercapturic acid were not believed to be responsible for urinary mutagenicity.

In in vivo studies in F344 rats administered 50 mg acrylonitrile/kg body weight intraperitoneally, 7-(2-oxoethyl)-guanine adducts were detected in liver (Hogy & Guengerich, 1986). Incorporation of acrylonitrile into hepatic RNA was observed following intraperitoneal administration to rats (Peter et al., 1983). However, no DNA adducts were detected in the brain, which is the primary target for acrylonitrile-induced tumorigenesis, in this or a subsequent study in which F344 rats received 50 or 100 mg acrylonitrile/kg body weight by subcutaneous injection (Prokopczyk et al., 1988). In contrast, in three studies from one laboratory, exposure of SD rats to 46.5 mg [¹⁴C]acrylonitrile/kg body weight (50 µCi/kg body weight) resulted in apparent binding of radioactivity to DNA from liver, stomach, brain (Farooqui & Ahmed, 1983), lung (Ahmed et al., 1992a), and testicles (Ahmed et al., 1992b). In each tissue, there was a rapid decrease in radioactivity of DNA samples collected up to 72 h following treatment.

It is not clear why acrylonitrile–DNA binding was detected in the brain in these studies and not in those by Hogy & Guengerich (1986) or Prokopczyk et al. (1988). The DNA isolation protocols and method for correcting for contaminating protein in the DNA sample used by Hogy & Guengerich (1986) may have allowed a more stringent determination of DNA-bound material. Alternatively, the methods used to achieve greater DNA purity might have caused the loss of adducts or inhibited the recovery of adducted DNA; more likely, however, 7-oxoethylguanine and cyanoethyl adducts are of little consequence in the induction of acrylonitrile-induced brain tumours. Indeed, investigation of the role of cyanohydroxyethylguanine and other adducts in the induction of these tumours seems warranted.

Consistent effects on the reproductive organs of male or female animals have not been observed in repeated-dose toxicity and carcinogenicity studies conducted to date. In a specialized investigation in CD-1 mice exposed by gavage, however, degenerative changes in the seminiferous tubules and associated decreases in sperm counts were observed at 10 mg/kg body weight per day (NOEL = 1 mg/kg body weight per day) (Tandon et al., 1988). Although epididymal sperm motility was reduced in a 13-week study with B6C3F₁ mice, there was no dose–response and no effect upon sperm density at doses up to 12 mg/kg body weight per day by gavage, although histopathological results were not reported (Southern Research Institute, 1996). In a three-generation study in rats exposed via drinking-water (14 or 70 mg/kg body weight per day), adverse effects on pup survival and viability and lactation indices were attributed to maternal toxicity (Litton Bionetics Inc., 1980).

In two studies by inhalation, developmental effects (fetotoxic and teratogenic) were not observed at concentrations that were not toxic to the mothers (Murray et al., 1978; Saillenfait et al., 1993). In the investigation in which concentration–response was best characterized (four exposure concentrations and controls with 2-fold spacing: 0, 12, 25, 50, and 100 ppm [0, 26.4, 55, 110, and 220 mg/m³]), the LOEL for maternal toxicity and for fetotoxicity was 25 ppm (55 mg/m³); the NOEL was 12 ppm (26.4 mg/m³) (Saillenfait et al., 1993).

Similarly, in two studies by the oral route, developmental effects have not been observed at doses that were not also toxic to the mothers (lowest reported effect level in the mothers was 14 mg/kg body weight per day) (Murray et al., 1978; Litton Bionetics Inc., 1980). Reversible biochemical effects on the brain but not functional neurological effects were observed in offspring of rats exposed to 5 mg/kg body weight per day (a dose that did not impact on body weight of the dams); dose–response was not investigated in this study (Mehrotra et al., 1988).

In recently published studies in rats exposed by inhalation to 25 ppm (55 mg/m³) acrylonitrile and above for 24 weeks, there were partially reversible time- and concentration-dependent reductions in motor and sensory conduction (Gagnaire et al., 1998).

In the few identified investigations of the immunological effects of acrylonitrile, effects on the lung following inhalation (Bhooma et al., 1992) and on the gastrointestinal tract following ingestion (Hamada et al., 1998) have been observed at concentrations and doses at which histopathological effects have also been observed. The effects on the lungs included an increase in the level of procoagulant activity in alveolar macrophages (Bhooma et al., 1992). The effects on the gastrointestinal tract consisted of an increase in the number of IgA-producing cells in the duodenum, jejunum, and ileum, an increase in the number of cells in S-phase in the duodenum and ileum, and a decrease in the response of splenocytes to mitogens (Hamada et al., 1998).

8.9.1 Cancer

Results of the few identified investigations in which the relative potency of acrylonitrile was compared with that of cyanoethylene oxide are consistent with the oxidative pathway of metabolism being critical in genotoxicity. In an assay with two strains of S. typhimurium, cyanoethylene oxide was mutagenic without activation, whereas acrylonitrile required activation (Cerna et al., 1981). In one study, cyanoethylene oxide was approximately 15-fold more mutagenic than acrylonitrile at the TK locus in cultured human lymphoblastoid cells (Recio & Skopek, 1988). In vitro, the formation of DNA adducts at high unphysiological concentrations is increased substantially in the presence of metabolic activation. Under non-activating conditions, cyanoethylene oxide alkylates DNA much more readily than acrylonitrile (Guengerich et al., 1981; Solomon et al., 1984, 1993).

Data on the binding of acrylonitrile to DNA are presented in section 8.6. However, available data are inadequate to implicate a particular adduct in the induction of acrylonitrile-induced brain tumours.

There are some suggestions from in vitro studies reported as abstracts that free radicals (·OH, O₂·) and hydrogen peroxide may be directly implicated in the oxidation of acrylonitrile and DNA damage. Formation of free radicals may be partially related to the release of cyanide or other mechanisms responsible for cellular and DNA damage (Ahmed et al., 1996; Ahmed & Nouraldeen, 1996; El-zahaby et al., 1996; Mohamadin et al., 1996).

In more recent investigations, the results of which have been presented incompletely at this time, Prow et al. (1997) reported that acrylonitrile inhibited gap junctional intercellular communication in a rat astrocyte cell line in a dose-dependent manner, possibly through an oxidative stress mechanism. Similarly, Zhang et al. (1998) assayed acrylonitrile with Syrian hamster embryo cells, with and without an antioxidant, and concluded that oxidative stress contributed to morphological transformation in the cells. Jiang et al. (1998) assayed acrylonitrile with a rat astrocyte cell line and reported oxidative damage (indicated by the presence of 8-hydroxy-2'-deoxyguanosine) at all concentrations tested.

Jiang et al. (1997) exposed male Sprague-Dawley rats to 0 or 100 mg acrylonitrile/litre in drinking-water for 2 weeks. End-points examined were levels of glutathione and reactive oxygen species in brain and liver, presence of 8-hydroxy-2'-deoxyguanosine (indicative of oxidative DNA damage) in several tissues, and determination of activation of NF-_KB (a transcription factor strongly associated with oxidative stress). Glutathione in brain was decreased. (Whysner et al. [1998a] reported no effects upon concentrations of glutathione in brain of male Sprague-Dawley rats exposed to 3, 30, or 300 mg acrylonitrile/litre in drinking-water for 3 weeks.) Reactive oxygen species were increased 4-fold in brain. Levels of 8-hydroxy-2'-deoxyguanosine were increased 3-fold in the brain. Activation of NF-_KB was also observed in the brain.

In recently conducted studies, levels of 8-oxodeoxyguanosine in the brain of rats exposed to acrylonitrile in drinking-water in each of the three following protocols have been examined (Whysner et al., 1997, 1998a):

• In male Sprague-Dawley rats exposed for 21 days to 0, 3, 30, or 300 mg/litre, there was a significant increase in 8-oxodeoxyguanosine in brain nuclear DNA at the two highest doses. In the liver, nuclear DNA 8-oxodeoxyguanosine concentrations were significantly increased at the two highest doses (Whysner et al., 1998a). In a bioassay with comparable dose levels, the incidence of brain and/or spinal cord tumours was significantly increased in male Sprague-Dawley rats exposed to 35 mg acrylonitrile/litre (3.4 mg/kg body weight per day) and higher for 2 years (Quast et al., 1980a).
• In male F344 rats exposed for 21 days to 0, 1, 3, 10, 30, or 100 mg/litre, there were no significant differences between groups for 8-oxodeoxyguanosine in the brain (Whysner et al., 1998a).
• In male Sprague-Dawley rats exposed for up to 94 days to 0 or 100 mg/litre, concentrations of 8-oxodeoxyguanosine in the brain were significantly increased after 3, 10, and 94 days of exposure (Whysner et al., 1998a). In the 2-year drinking-water bioassay with male Sprague-Dawley rats (Quast et al., 1980a), the incidence of brain and/or spinal cord tumours was significantly increased at 100 mg/litre (8.5 mg/kg body weight per day).

The end-point for which changes were consistently observed in male Sprague-Dawley rats was the induction of oxidative DNA damage, including the accumulation of 8-oxodeoxyguanosine in the brain. The authors drew correlations between these results and the incidence of brain/spinal cord tumours that had been reported in carcinogenicity bioassays in which male Sprague-Dawley rats were exposed to acrylonitrile via drinking-water.

Increased levels of 8-oxodeoxyguanosine occur only in the anterior portion of the brain, which contains rapidly dividing glial cells (Whysner et al., 1998b).

8.9.2 Neurotoxicity

Neurotoxicity following inhalation by male Sprague-Dawley rats of 0, 25, 50, or 100 ppm (0, 55, 110, or 220 mg/m³) acrylonitrile for 6 h/day, 5 days/week, for 24 weeks, followed by 8 weeks of recovery, was reported by Gagnaire et al. (1998). Body weight at the high dose was significantly reduced throughout exposure. Clinical observations at the mid and high doses included wet fur and hypersalivation during exposure. The authors noted that signs were similar to those associated with acute acetylcholine-like toxicity. There were time- and concentration-dependent reductions in motor conduction velocity, sensory conduction velocity, and amplitude of sensory action potential in tail nerve, which were partially reversible after 8 weeks of recovery (LOEL = 25 ppm [55 mg/m³]). The protocol did not include histological examination.

In case reports of acute intoxication, effects on the central nervous system characteristic of cyanide poisoning and effects on the liver, manifested as increased enzyme levels in the blood, have been observed. There have also been reports that acrylonitrile is a skin irritant and skin sensitizer, the latter based on patch testing of workers (Balda, 1975; Bakker et al., 1991; EC, 2000; Chu & Sun, 2001).

In the few studies in which non-neoplastic effects of acrylonitrile have been systematically investigated, only acute dermal irritation has been reported consistently. In a cross-sectional investigation of workers exposed in acrylic fibre factories to approximately 1 ppm (2.2 mg/m³), there was no consistent evidence of adverse effects based on examination of a wide range of clinical parameters, including liver function tests (Muto et al., 1992). However, there was an increase in subjective symptoms of acute dermal irritation, consistent with observations in another cohort of acrylic fibre manufacturing workers (Kaneko & Omae, 1992).

In a cross-sectional investigation of a smaller group of workers producing acrylic textile fibres for which quantitative data on exposure were not reported, there was no evidence of induction of hepatic cytochrome P-450 or genotoxicity of urine (Borba et al., 1996).

Although there was some evidence in primarily early limited studies of excesses of lung cancer (Thiess et al., 1980), "all tumours" (Zhou & Wang, 1991), and colorectal cancer (Mastrangelo et al., 1993), such excesses have not been confirmed in well conducted and well reported recent investigations in four relatively large cohorts of workers (Benn & Osborne, 1998; Blair et al., 1998; Swaen et al., 1998; Wood et al., 1998). Indeed, there is no consistent, convincing evidence of an association between exposure to acrylonitrile and cancer of a particular site that fulfils, even in part, traditional criteria for causality in epidemiological studies.

Benn & Osborne (1998) reported results of a historical cohort study carried out in 2763 workers. Vital status was traced from 1978 through 1991, and follow-up was virtually complete. Only for lung cancer was there any indication of excess risk, and that was in the more highly exposed jobs and among the very young; however, based upon detailed analyses, there was no consistent support for the hypothesis of a causal relationship between exposure to acrylonitrile and lung cancer. Limited information on exposure levels, questionable quality of source records of work histories, and use of national rather than local rates for computing standardized mortality ratios (SMRs) compromise the inferences that can be drawn.

Swaen et al. (1998) conducted a historical cohort study in 2842 workers occupationally exposed to acrylonitrile. Extensive industrial hygiene assessments were conducted to quantify past exposure to acrylonitrile. The follow-up between 1979 and 1995 was 99.6% complete; in 99.3% of cases, the cause of death could be ascertained. Selected cause-specific SMRs in the exposed cohort were as follows: lung, 109.8 (95% confidence interval [CI] = 81–146, number of observed cases [Obs] = 47); colon, 126.0 (95% CI = 58–239, Obs = 9); leukaemia, 266.7 (95% CI = 54–390, Obs = 5); prostate, 83.3 (95% CI = 22–213, Obs = 4); bladder, 97.9 (95% CI = 20–286, Obs = 3); brain, 173.9 (95% CI = 64–378, Obs = 6). Some of these results are suggestive of excess risks, but the evidence of excess for any specific type of cancer is not strong. The data appear to have been well collected and analysed, and follow-up was good. While this is a reasonably large study, the power to detect dose–response relationships is still limited. Smoking was not considered.

The largest and most statistically powerful of the recent cohort studies was that conducted by Blair et al. (1998), which included 25 460 workers from eight plants producing and using acrylonitrile. Estimates of exposure to acrylonitrile were based on information from production procedures at the plants, interviews with management and labour, monitoring data from the companies, and monitoring conducted by the investigators specifically for the study. Vital status was successfully determined for 96% of the cohort. There was a total of 545 369 person-years of follow-up, of which one-third was in unexposed workers. This study had the most extensive exposure assessment protocol and the most extensive data on smoking. The mortality follow-up was complete, and the observation period was lengthy. There was a small (non-significant) excess of lung cancer in the highest quintile of cumulative exposure (relative risk = 1.5; 95% CI = 0.9–2.4), but no exposure–response trend. The exposure categories were as follows:

It should be noted that the power to detect moderate excesses was small for some sites (stomach, brain, breast, prostate, lymphatic/haematopoietic) because of small numbers of expected deaths.

Wood et al. (1998) reported a historical cohort mortality and cancer incidence study in 2559 workers who had been occupationally exposed to acrylonitrile. There was a total of 75 009 person-years of mortality observation among the study subjects. There were no significantly elevated SMRs for specific cancer sites. This study was carried out in a company with good-quality information on work practices and industrial hygiene, complete work records, and good follow-up for mortality and cancer incidence. The latter is a unique advantage of this data set. The data appear to have been well collected and analysed. Limitations include the lack of data on workers’ race and smoking habits. Based upon a comparison with US rates, there was a 40% all-cause deficit and 34% cancer deficit seen in the Waynesboro cohort, which suggests a lack of reporting of cause of death. While this is a reasonably large study, power to detect dose–response relationships is still limited. While considerable effort was made to gather exposure data, there was no monitoring of acrylonitrile before 1975, and data for early years were inferred. An additional strength of this study is the number of person-years at the high level of exposure.

The toxicity of acrylonitrile has been examined in a wide range of aquatic organisms, while the data set on the toxicity of acrylonitrile to terrestrial organisms is more limited.

The data set for acrylonitrile includes a wide range of information on short- and long-term toxicity in 34 species of fish, amphibians, aquatic invertebrates, and algae, although none complies totally with the requirements of OECD or similar test guideline protocols, and volatility was often not adequately addressed (Environment Canada, 1998). Below, a brief summary of the key studies carried out in general compliance with current OECD testing protocols and/or where concentrations were measured or could be adequately adjusted is presented. Primary studies selected include those where concentrations have been measured in static or static-renewal tests or flow-through tests with five turnovers per day (Henderson et al., 1961; Bailey et al., 1985; V. Nabholz, personal communication, 1998).

T.D. Sabourin (personal communication, 1987) determined the ratio of flow-through to static concentrations at the 96-h period to be 0.23. Therefore, studies with 96-h end-points can be adjusted by multiplying the reported concentration by 0.23, although the evidence provided by these studies is considered secondary. Tests done under static conditions or those with nominal concentrations only at a time period different from 96 h are considered as supporting evidence only.

Of the freshwater studies, there are five studies on five fish species and one study with an amphibian that are considered to provide primary data (Henderson et al., 1961; Sloof, 1979; Analytical BioChemistry Laboratories, 1980a; Bailey et al., 1985; Zhang et al., 1996). In addition to these, there is secondary evidence (adjusted concentrations) from studies with six fish, seven invertebrate, and one plant species. In these studies, a variety of end-points was examined, including survival, growth, respiration, and mobility at exposure durations ranging from 24 to 840 h (1–35 days). The remainder of the studies (i.e., where there was no replication of doses or other limitations such as lack of aeration) were considered as supporting evidence only.

The 96-h LC₅₀s for freshwater fish range from 10 to 20 mg/litre (nominal) (Henderson et al., 1961; Analytical BioChemistry Laboratories, 1980b; Zhang et al., 1996). Reported 48-h LC₅₀s range between 14.3 and 33.5 mg/litre. At 840 h, the LC₅₀ for fathead minnow (Pimephales promelas) was 0.89 mg/litre (Analytical BioChemistry Laboratories, 1980a).

Based on the primary evidence, the most sensitive aquatic end-point was that for long-term exposure of the frog Bufo bufo gargarizans in its early life stage (Zhang et al., 1996). Three-day-old tadpoles were exposed for 28 days in a flow-through system with four turnovers per day. The most sensitive end-point was foreleg growth, where the lower and upper chronic limits around the 28-day EC₅₀ were 0.4 mg/litre and 0.8 mg/litre, respectively. The 96-h and 48-h EC₅₀s for immobility were 11.59 mg/litre and 14.22 mg/litre, respectively.

The effect of acrylonitrile on the growth (length and wet weight) and mortality of the early life stage (<18-h-old eggs) of the fathead minnow (Analytical BioChemistry Laboratories, 1980a) in a flow-through system with more than 5.5 turnovers per day has been examined. Mean measured concentrations were 98% of nominal. The most sensitive end-point in the study was the 840-h (35-day) lowest-observed-effect concentration (LOEC) for weight (20% reduction in wet weight) at 0.44 mg/litre; the corresponding no-observed-effect concentration (NOEC) was 0.34 mg/litre. For mortality, the 840-h NOEC (LC₁₅) was 0.44 mg/litre, and the LOEC (LC₄₆) was 0.86 mg/litre.

Henderson et al. (1961) reported mortality of fathead minnow exposed to acrylonitrile in a flow-through system in which solutions were renewed every 100 min. Test durations were 24, 48, 72, and 96 h and 5, 10, 15, 20, 25, and 30 days (720 h). Effects ranged from the 24-h LC₅₀ of 33.5 mg/litre through decreasing concentrations to the most sensitive end-point in the study, the 720-h LC₅₀ at 2.6 mg/litre.

Sloof (1979