Registration Dossier

Hazard assessment conclusion:

high hazard (no threshold derived)

Most sensitive endpoint:

carcinogenicity

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

acute toxicity

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

1.8 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL derivation method:

other: No assessment factors required

Overall assessment factor (AF):

1

Dose descriptor:

NOAEC

Value:

1.8 mg/m³

AF for dose response relationship:

1

Justification:

Default factor

AF for differences in duration of exposure:

1

Justification:

Consideration of exposure duration is not relevant for local effects

AF for interspecies differences (allometric scaling):

1

Justification:

Not relevant: the starting point is derived from a human study

AF for other interspecies differences:

1

Justification:

Not relevant: the starting point is derived from a human study

AF for intraspecies differences:

1

Justification:

Not relevant: the starting point is derived from worker monitoring data and is considered to be applicable to all workers

AF for the quality of the whole database:

1

Justification:

An additional factor is not required: comprehensive and high quality database

AF for remaining uncertainties:

1

Justification:

An additional factor is not required: there are no significant remaining uncertainties

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

10 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL derivation method:

other: No assessment factors are required

Overall assessment factor (AF):

1

Dose descriptor starting point:

NOAEC

AF for dose response relationship:

1

Justification:

Default value

AF for interspecies differences (allometric scaling):

1

Justification:

Not relevant: starting point is derived from a human study

AF for other interspecies differences:

1

Justification:

Not relevant: starting point is derived from a human study

AF for intraspecies differences:

1

Justification:

Not relevant: starting point is derived from a study in exposed workers

AF for the quality of the whole database:

1

Justification:

An additional factor is not required: comprehensive and high quality database

AF for remaining uncertainties:

1

Justification:

An additional factor is not required: comprehensive and high quality database

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

1.4 mg/kg bw/day

Most sensitive endpoint:

neurotoxicity

Route of original study:

Oral

DNEL derivation method:

other: The DNEL is based on a human-specific reference dose derived using PBPK modelling. The assessment factors used in this derivation are specific to this endpoint and reflect substance-specific factors

Overall assessment factor (AF):

6

Dose descriptor starting point:

LOAEL

Value:

25 mg/kg bw/day

Modified dose descriptor starting point:

BMDL05

Value:

8.5 mg/kg bw/day

Explanation for the modification of the dose descriptor starting point:

Based on the Kirman (2008) analysis of the study of Gagnaire et al. (1998) which identifies effects on sensory nerve conduction velocity, the relevant endpoint is the human equivalent oral BMDL05 of 8.5 mg/kg bw/d. The human equivalent value is derived using PBPK modelling. This DNEL derivation assumes that dermal absorption is the same as oral absorption (the default assumption).

AF for dose response relationship:

1

Justification:

Default factor

AF for differences in duration of exposure:

2

Justification:

Extrapolation from sub-chronic study to chronic exposure

AF for interspecies differences (allometric scaling):

1

Justification:

Not required; already accounted for in derivation of the human equivalent BMDL05

AF for other interspecies differences:

3

Justification:

PoD derived from a rat study

AF for intraspecies differences:

1

Justification:

Not required

AF for the quality of the whole database:

1

Justification:

An additional factor is not required: comprehensive and high quality database

AF for remaining uncertainties:

1

Justification:

An additional factor is not required: no significant remaining uncertainties

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

acute toxicity

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

sensitisation (skin)

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

sensitisation (skin)

Hazard assessment conclusion:

medium hazard (no threshold derived)

Overview of the toxicology of acrylonitrile

The toxicity of acrylonitrile is complex and has been comprehensively investigated in standard and non-standard investigative studies in experimental animal species. An extensive human epidemiological dataset is also available, which focuses largely on carcinogenicity but also includes investigations of reproductive toxicity, neurotoxicity and irritant effects. Recent laboratory investigations have focussed on elucidating the mode of action for the carcinogenicity of acrylonitrile, which is clearly apparent in rodent studies but for which there is no convincing evidence in studies in exposed workers. The available toxicological dataset for acrylonitrile has been comprehensively reviewed in the EU RAR (2004) which incorporated studies published prior to 2000 and also in an independently peer-reviewed report by The Sapphire Group (2004). It is notable, however, that a number of important mechanistic studies have been published in recent years which have not been included in these reports.

Toxicokinetics

The toxicokinetic behaviour of acrylonitrile has been extensively investigated in studies in experimental animals. The majority of data come from gavage studies in rats; however data are available for inhalation and other exposure routes in the rat and also for other species. Limited human data are also available. A more recent review of the metabolism of acrylonitrile and its potential role in genotoxicity and carcinogenicity is also provided (Albertini et al., 2016).

Absorption

Animal data indicate that acrylonitrile is rapidly and extensively absorbed following inhalation exposure. Retention of >90% is reported; absorption may be biphasic and may be enhanced in glutathione-depleted rats. Oral absorption of 95 -98% is reported, although the rate of absorption by this route may be relatively slower, with Cmax within 3-4 hours. The retention of inhaled acrylonitrile in humans (~50%) may be lower than that reported for rats. The results of a dermal absorption study in rat and human skin membranesin vitroindicate very limited dermal absorption (~1%) and show that the large majority of the applied test material was lost though volatilisation. However a worst case dermal absorption of 100% (i.e. a level equivalent to that of oral absorption is assumed for the purposes of risk assessment.

Distribution

Acrylonitrile is rapidly and evenly distributed; however the ability of its metabolite CEO to bind to cellular macromolecules results in apparent accumulation in erythrocytes and some organs. Retention of approximately 25% of an oral dose has been reported, this is attributable to covalent binding. The extent of binding is increased in glutathione-depleted rats.

Metabolism

Studies in rats demonstrate that acrylonitrile is metabolised via two distinct pathways. Pathway 1 involves the direct conjugation of acrylonitrile with glutathione and can be considered as a detoxification step; the terminal metabolite of this pathway in the rat is 2-cyanomethylmercapturic acid (CMA; also detected in humans). Pathway 2 involves the initial cytochrome P450-mediated oxidation of acrylonitrile to form the reactive epoxide 2-cyanoethylene oxide (CEO), and can be considered as an activation step. CEO can react with glutathione at the 2-position with the eventual formation of the metabolite HMA (2-hydroxyethylmercapturic acid) or, alternatively, may react with glutathione at the 3-position with the formation of cyanide. Cyanide generated by the metabolism of acrylonitrile is detoxified by rhodanese, reacting with endogenous thiosulphate to form thiocyanate. Other metabolites generated in this pathway include thiodiglycolic acid and thionyldiacetic acid. In the rat, the predominant metabolic pathway is dependent on dose level and route of exposure. Pathway 1 is more important for intraperitoneal, intravenous or high dose gavage exposure than for inhalation or low dose gavage exposure, where metabolism via pathway 2 predominated. Data indicate the importance of the first-pass effect and further indicate that pathway 2 is saturable. Evidence from human studies indicates the presence of an additional pathway which involves the hydrolysis of the reactive metabolite CEO by epoxide hydrolase. This pathway is not normally present in rodent species, but is inducible. Acrylonitrile is also metabolised to carbon dioxide (derived from cyanide). The relatively low glutathione transferase activity in humans means that pathway 1 may be of lesser importance; however the existence of the additional epoxide hydrolase pathway in humans means that CEO can be effectively deactivated. Limited data from studies in mice indicate that pathway 2 is more important in this species, with greater excretion of urinary thiocyanate; in vitro data also show a greater formation of CEO formation by mouse microsomes. The greater formation of cyanide in the mouse may account for the greater acute toxicity of acrylonitrile seen in this species.

Excretion

Excretion of metabolites by rats administered acrylonitrile by gavage dosing is rapid (within 24 hours) and largely in the urine. Some excretion as unchanged acrylonitrile (5%), HCN (0.5%) and carbon dioxide (up to 9%) is reported in exhaled air. Faecal excretion is a relatively minor route, accounting for ~3-8% of an administered dose.

PBPK modelling

Kedderis et al. (1996) and Sweeney et al. (2003) have developed PBPK models for the disposition of acrylonitrile and CEO in exposed humans, based on human in vitro data and scaling from a rat model. PBPK modelling analysis suggests no heightened sensitivity of humans in general to toxic effects of acrylonitrile based on pharmacokinetic differences.

Acute toxicity

Acute oral toxicity

Acrylonitrile is acutely toxic to experimental animals by all routes of exposure investigated. The results of a guideline-comparable rat study report an LD50 value of 81 mg/kg bw; other rat studies report value of 72-186 mg/kg bw. Other species investigated show slightly greater sensitivity with LD50 values of 25-48 mg/kg bw in mice, 50-85 mg/kg bw in guinea pigs and 93 mg/kg bw in the rabbit. Acrylonitrile has a harmonised classification for acute oral toxicity in Category 3 (H301: Toxic if swallowed); no change to this classification is proposed.

Acute dermal toxicity

A recent, guideline-compliant acute dermal toxicity study in the rat reports an LD50 value of >200 mg/kg bw, with only one mortality (out of ten rats) at this dose level. Dermal LD50 values are also reported for the rabbit (<200-226 mg/kg bw), rat (148-282 mg/kg bw) and guinea pig (260-690 mg/kg bw). Acrylonitrile has a harmonised classification for acute dermal toxicity in Category 3 (H311: Toxic in contact with skin); no change to this classification is proposed.

Acute inhalation toxicity

A recent, guideline-compliant inhalation toxicity study in the rat reports a 4-hour LC50 value of 2.05 mg/L. Other LC50 values reported for the rat are 2.24 mg/L (1-hour exposure), 3.42 mg/L (1-hour exposure) and 1.21 mg/L (4-hour exposure). The EU RAR reports LC50 values of 0.3-1.21 mg/L for 4-hour exposures in the rat. The mouse and guinea pig are also noted to be more sensitive species. Acrylonitrile has a harmonised classification for acute inhalation toxicity in Category 3 (H331: Toxic by inhalation); no change to this classification is proposed. Data from accidental human exposure also show a potential for toxicity, including lethality.

Mechanism of acute toxicity

The signs of toxicity following acute exposure to acrylonitrile are consistent between species. It has been postulated that the mechanism of toxicity is a consequence of the metabolism of acrylonitrile to liberate cyanide, although there is no evidence for ATP depletion and cholinergic symptoms are also noted (albeit inconsistently). The mechanism of toxicity is therefore likely to be complex. Acrylonitrile is classified for acute toxicity by all routes of exposure; additional classification for specific target organ toxicity (STOT-SE) is not required based on the results of these studies.

Irritation and corrosion

A number of skin irritation and eye irritation studies are available. Studies are of variable design but indicate that acrylonitrile is a skin irritant and severe eye irritant. The animal data are also consistent with incidents involving accidentally exposed workers. Findings from animal studies and human experience also indicate that the substance is a respiratory irritant. Annex VI of the CLP Regulation (1272/2008/EC) implementing GHS in the EU lists acrylonitrile with classification as Skin Irritation Cat 2 (H315: causes skin irritation), eye damage Cat 1 (H318: causes serious eye damage) and STOT SE3 H335 (May cause respiratory irritation). This classification is consistent with the available data and no changes are proposed. The EU RAR also notes isolated reports of corrosive effects of acrylonitrile following accidental human exposure. It concludes, however, that animal data and more recent human experience indicate that while acrylonitrile is irritating to skin, eye and respiratory tract it should not be considered as corrosive and that classification of acrylonitrile as corrosive is not appropriate.

Sensitisation

The results of a guideline-compliant modern Maximisation assay report a positive result for acrylonitrile. This finding is consistent with the existing classification of acrylonitrile as a skin sensitiser and reports of allergic skin reactions in exposed workers. Acrylonitrile is listed on Annex VI of the CLP Regulation with as a skin sensitiser in Category 1 (H317: May cause an allergic skin reaction). Based on the results of the critical study, classification in Catgeory 1B is appropriate. No classification for respiratory sensitisation is proposed in the absence of any evidence that acrylonitrile can cause occupational asthma.

Repeated dose toxicity

The repeated dose toxicity of acrylonitrile has been investigated in high quality studies using oral (gavage or drinking water) administration to rats and mice, and has also been extensively investigated in rats using inhalation exposure, Limited non-standard investigations in a number of other species are also available.

Repeated dose oral toxicity

A 14-week gavage study in the mouse (NTP, 2001) gives a NOAEL of 5 mg/kg bw/d. Mortality was seen at a dose level of 60 mg/kg bw/d, with signs of toxicity and local gastric irritation seen at 40 mg/kg bw/d. A further 13-week gavage study (Serota et al., 1996) did not identify any clearly treatment-related effects at the highest dose level of 12 mg/kg bw/d. A two-year rat drinking water study (Quast et al., 1980; 2002) notes the primary non-neoplastic effects of acrylonitrile exposure to be forestomach irritation (hyperplasia and/or hyperkeratosis) and CNS gliosis with or without perivascular cuffing. Findings were apparent in rats of both sexes and all treatment groups in this study. A LOAEL for this study was therefore the lowest drinking water concentration of 35 ppm, calculated to be equivalent to dose levels of 3.4 and 4.4 mg/kg bw/d acrylonitrile in males and females respectively. A second rat chronic drinking water study (Johannsen & Levinskas, 2002) identified a NOAEL of 3 ppm, equivalent to 0.25 mg/kg bw/d in males and 0.36 mg/kg bw/d in females. The NOAEL of 0.25 mg/kg bw/d from this study is considered to be the critical endpoint for the repeated dose oral toxicity of acrylonitrile.

Repeated dose dermal toxicity

No data are available for the repeated dose toxicity of acrylonitrile by the dermal route; adequate data are available for repeated dose toxicity by the oral and inhalation routes. Testing by the inhalation route is considered to be most relevant (with regard to the likely route of occupational exposure) for a volatile liquid. Based on kinetic considerations, the systemic dermal toxicity of acrylonitrile is not predicted to be fundamentally different to that seen following oral and/or inhalation exposure, therefore specific data for this route are not required. Due to the irritant and sensitising properties of the substance, it is likely that the effects of repeated dose dermal exposure will be dominated by local (site of contact) effects which will severely limit systemic exposure to the substance and consequently limit the relevance of the study. The use of engineering controls and PPE will also minimise dermal exposure to the substance under normal occupational conditions. Testing is therefore not scientifically justified and additionally cannot be supported on grounds of animal welfare.

Repeated dose inhalation toxicity

The repeated dose inhalation toxicity of acrylonitrile has been investigated in a number of species, with the more modern and reliable studies being performed in the rat. The two generation inhalation toxicity study of Nemec et al. (2008) identified local irritation as the critical effect of acrylonitrile. Histopathological effects in the nasal cavity including respiratory/transitional epithelial hyperplasia, sub-acute inflammation, squamous metaplasia, and/or degeneration of the olfactory epithelium were seen at all exposure levels including the lowest level of 5 ppm. Similar findings were apparent in a 2-year inhalation study (Quast et al., 1980), in which local irritant effects were observed in the nasal epithelium (suppurative rhinitis, hyperplasia, focal erosions, and squamous metaplasia of the respiratory epithelium, hyperplasia of the mucus-secreting cells) at the lowest exposure level of 20 ppm. The EU RAR suggests that, as effects were due to local irritancy and the other systemic, non-neoplastic findings in treated rats were secondary to carcinogenicity rather than direct systemic toxicity, the application of an uncertainty factor of five to the LOAEC of 20 ppm could be used to derive a suggested NOAEC of 4 ppm (9 mg/m3). A 12-month inhalation study reported by Maltoni et al. (1977) focussed on neoplasia and provides little useful information on chronic toxicity, however a NOAEC of 5-10 ppm can be estimated for this study. Dudley et al. (1942) report the results of investigations in various species at relatively high exposure levels which identified effects on the nervous system, respiratory tract and kidney. The dog appeared to be the most sensitive species.

Classification for repeated dose toxicity

No classification is proposed for repeated dose toxicity (STOT-RE). Acrylonitrile is classified as acutely toxic by all routes of exposure; effects seen at high levels of exposure in repeated dose studies are manifestations of the same toxicity, therefore additional classification is not required.

Genetic toxicity

The genetic toxicity of acrylonitrile has been extensively investigated over many years in both standard and non-standard investigative studies in vitro and in vivo; in isolated DNA, cell cultures, experimental animal and also in worker exposure studies. The majority of the studies are non-standard and, individually, many may be considered limited in terms of design. As a consequence of the extensive and historical dataset, it is inevitable that there is some inconsistency between the findings of individual studies. Interpretation of the dataset therefore requires consideration of the weight of evidence from the available studies. The acrylonitrile genotoxicity dataset has previously been reviewed by (among others), Whysner (1998), in the EU RAR (2004) and by The Sapphire Group (2004) and subsequently by Strother (2010) and Albertini (2016).

Genetic toxicity in vitro

Acrylonitrile is shown to bind to isolated protein and DNA; the DNA binding is relatively slow, occurring only at conditions which are not representative of biological activity, and is enhanced by metabolic activation. In contrast, DNA binding by the reactive metabolite CEO is much more rapid and is independent of metabolic activation. Similarly, acrylonitrile can produce DNA adducts following prolonged exposure to massive concentrations, whereas CEO produces DNA adducts much more readily. The induction of 8oxoG adducts by acrylonitrile has been demonstrated in rat astrocyte DNA and (at much higher concentrations) in human astrocyte DNA; no evidence of adduct formation was seen in rat hepatocyte DNA. Both acrylonitrile and CEO are reported to cause DNA strand breaks, however the methodology used in these studies had the potential to convert alkali-labile sites to strand breaks. A Rec assay in B. subtilis reports a positive response for acrylonitrile in the presence of metabolic activation; contrasting results in standard and modified Comet assays indicate a lack of direct DNA damage by acrylonitrile. Studies of UDS in vitro have reported positive responses for acrylonitrile where liquid scintillation counting is used to assess UDS, whereas studies using the preferred autoradiography method (which more effectively excludes replicative DNA synthesis), report mostly negative responses even at cytotoxic concentrations of acrylonitrile. A positive response is noted for acrylonitrile and CEO in a UDS assay using human mammary epithelial cells and autoradiography; the positive response with acrylonitrile was seen only at very high concentrations. Positive responses for SCE have been reported in CHO cells (with and without metabolic activation) and in cultured human bronchial cells; negative responses are reported in metabolically competent rat liver cells and in human peripheral blood lymphocytes (both with and without metabolic activation), however it should be noted that this type of change is of unclear toxicological significance.

Acrylonitrile has been investigated in a large number of bacterial mutation assays; positive results are reported in Salmonella strains sensitive to base pair substitution. The results of mutagenicity studies in mammalian cells are varied and there is no consistent association with metabolic activation; some studies report positive results with activation only, others both with and without activation. In mammalian cells, the potential of acrylonitrile to induce clastogenicity has been investigated in human peripheral blood lymphocytes, CHO, CHL and metabolically competent rat liver RL4 cell lines. Many studies have reported positive results for the induction of structural aberrations, with most requiring metabolic activation. There is no evidence for the induction of numerical aberrations.

Genetic toxicity in vivo

Studies of sex-chromosome aneuploidy with acrylonitrile in Drosophila gave positive results; however further studies with this system show that nitriles may induce aneuploidy due to effects on spindle formation. Two studies of heritable SLRL gene mutation with acrylonitrile have been negative. It is notable that concentrations of acrylonitrile used in the Drosophila studies are high compared to those used in mammalian studies. Studies of DNA binding have demonstrated the induction of oxidative damage-specific DNA adducts in the rat, most markedly in the brain. The induction of adducts is associated with lipid peroxidation, ROS generation and reduced levels of glutathione. Several studies have reported the induction of UDS in rats administered acrylonitrile either by intraperitoneal injection or oral gavage. It is notable that all of the positive studies used liquid scintillation counting to measure UDS. One study showed that the induction of UDS was associated with glutathione depletion, was increased by depleting glutathione prior to administration and was inhibited by pre-treatment with sulphydryl compounds. A single study using autoradiography to assess UDS showed a negative response in testis and liver DNA following gavage with acrylonitrile. A single study reports an essentially negative (weak positive) SCE response in mice following the intraperitoneal injection of acrylonitrile. All published and peer-reviewed studies of mutational effects in mammalian studies have reported negative results, including studies of germ cell mutagenicity. These include investigations of both chromosome aberrations and micronucleus inductions in rat and mouse somatic cells and studies of heritable chromosome level changes, such as dominant lethal effects, in both species, and chromosome level changes in murine spermatocytes. These studies were conducted independently by several laboratories, spanned a period of over 20 years, and involved exposure by a variety of routes including inhalation, oral and intraperitoneal injection.

Human data

Studies in acrylonitrile-exposed human (worker) populations have investigated the induction of chromosome aberration, micronuclei induction, DNA strand breaks, sex chromosome aneuploidy and investigation of sperm parameters as markers of effect. The results of studies investigating effects at the chromosomal level are mixed; one study reports a positive finding but with a response pattern which does not indicate exposure to acrylonitrile as the causative factor. Another positive study reporting chromosomal aberrations is confounded by high levels of illness attributable to other chemical exposures. Other studies report a change in the pattern of chromosomal aberration with no overall increase, or do not provide sufficient detail to facilitate interpretation. A study of HPRT mutations in acrylonitrile-exposed workers used a method known to produce artefacts and is therefore not considered to be reliable.

Relevant mechanistic data

The extensive dataset clearly indicates that acrylonitrile is genotoxic in vitro. Acrylonitrile is, however, a weak mutagenic agent. Most positive findings are reported in vitro and involve high exposure levels and test organisms selected for their susceptibility to mutagenesis. In vivo, however, there is a paucity of positive studies. There is only one report of weak gene mutation induction in mice and scattered positive results in humans, many of which can be questioned. The lack of induction of sex-linked recessive lethal mutations in Drosophila is comparable with other weak mutagens whose mutagenic effects are inhibited by effective repair systems. Acrylonitrile has the capacity to induce oxidative stress and oxidative DNA damage. This effect (and the consequent indirect genotoxicity) may account for the genotoxicity profile of acrylonitrile, indicating a threshold response. The threshold is due to the fact that mutagenicity will only be apparent following exposure to sufficient acrylonitrile to result in oxidative stress, generate ROS and overwhelm natural cellular defences. Acrylonitrile can also produce other (non-genotoxic) cellular effects associated with carcinogenesis, including cell transformation and the inhibition of gap junctions. Some of the carcinogenicity of acrylonitrile in rodent studies in vivo is likely to depend on these non-genotoxic effects, which are also likely to be tissue-dependent. A high incidence of spontaneous tumours in a tissue or organ may reflect a high number of spontaneously initiated cells which require needing only promotion to produce tumours.

It is concluded that the genotoxicity profile of acrylonitrile, considered in toto, is most compatible with an indirect genotoxic MoA for cancer induction in rodents, rather than a direct, DNA-reactive MoA. There are exposure levels of acrylonitrile below which no oxidative stress, and no oxidative DNA damage, have been detected. These factors, together with the non-linear tumour dose response observed with acrylonitrile in rodent studies and the absence of a causal link between acrylonitrile exposure and cancer in extensive occupational epidemiology investigations support a view that acrylonitrile is a threshold rodent carcinogen.

Carcinogenicity

The carcinogenicity of acrylonitrile has been investigated in a large number of studies in rats and mice, using oral (gavage, drinking water) and inhalation exposure. The results of the studies indicate that acrylonitrile is a multi-site carcinogen in rodent species.

Study in the mouse

A mouse carcinogenicity study with acrylonitrile was performed by the NTP (2001). Clear evidence of the carcinogenicity was seen in this study in male and female B6C3F1 mice administered acrylonitrile by gavage at dose levels of 0, 2.5, 10 or 20 mg/kg bw/d for 2 years. Increased incidences of forestomach tumours were seen in both sexes at 10 and 20 mg/kg bw/d. Harderian gland hyperplasia was increased in males at 10 mg/kg bw/d; increased incidences of Harderian gland tumours were seen in all treated groups of males and in females at 10 and 20 mg/kg bw/d. Ovarian and bronchioalveolar tumour incidences were increased in females at 10 mg/kg bw/d.

Studies in the rat

Quast et al. (1980, 2002) administered acrylonitrile in drinking water to male and female Sprague-Dawley rats at concentrations of 0, 35, 100 or 300 ppm. Increased tumour incidences were seen in one or more dose levels in the brain, Zymbal’s gland, forestomach, tongue, small intestine and mammary gland. Bigner et al. (1986), in a study specifically designed to investigate the incidence and origin of brain tumours, administered acrylonitrile in the drinking water to male and female F344 rats at concentrations of 100 and 500 ppm. Increased incidences of brain tumours were seen in both treated groups, with increased incidences of tumours in other organs (skin, stomach, Zymbal’s gland) also noted. Although the brain tumours noted in this study morphologically closely resembled astrocytic tumours commonly seen in this rat strain, specific staining did not reveal the presence of GFAP thus indicating a different cellular origin. Gallagher et al. (1988) administered acrylonitrile in the drinking water to male CD rats at concentrations of 0, 20, 100 or 500 ppm for two years. Increased incidences of Zymbal’s gland tumours were seen at 100 and 500 ppm, with forestomach papillomatous changes also noted at 500 ppm and considered likely to be secondary to local irritation. The incidence of tumours in other organs and tissues were not affected by treatment, however the small group size may have limited the power of the study. Johannsen & Levinskas (2002) administered acrylonitrile in the drinking water for approximately 2 years to groups of 100 male and 100 female F344 rats at nominal concentrations of 1, 3, 10, 30 and 100 ppm. The only significant non-neoplastic finding observed histologically was a dose-related increase in hyperplasia and/or hyperkeratosis in squamous cells of the forestomach in male and female rats at concentrations of 3 ppm and higher. This observation correlated with the induction of treatment-related squamous cell tumours (papillomas and carcinomas) of the forestomach seen primarily in rats in these groups. Mammary gland carcinomas were observed only in female groups. Both sexes given 10 ppm acrylonitrile or more had astrocytomas of the brain/spinal cord and adenomas/carcinomas of the Zymbal's gland. In a study by the same authors in Spartan rats, rats administered acrylonitrile continually at concentrations of 0 (controls), 1 or 100 ppm in the drinking water (Johannsen & Levinskas. 2002), treatment-related tumours of the central nervous system (brain, spinal cord), ear canal, and gastrointestinal tract were observed in both sexes administered 100 ppm. The same authors also performed a comparative gavage study in Spartan rats, at dose levels of 0, 0.1 or 10 mg/kg bw/d (doses selected to approximate the same daily intake of acrylonitrile in the drinking water study). In both sexes, treatment-related tumours of the central nervous system (brain, spinal cord), ear canal and gastrointestinal tract, and in females only, the mammary gland were observed in rats administered 10 mg/kg bw/d. In a somewhat less reliable study (Maltoni et al., 1977; 1988), acrylonitrile was administered by gavage to male and female Sprague-Dawley rats at a single daily dose of 5 mg/kg bw three times weekly for 52 weeks. The rats were subsequently observed until spontaneous death occurred. The only increases in incidence of tumours were in the mammary gland and forestomach of female rats.

In an inhalation carcinogenicity study (Quast et al., 1980), male and female Spartan Sprague-Dawley rats were exposed to acrylonitrile by inhalation at concentrations of 0, 20 or 80 ppm, 6 hours/day, 5 days/week for 2 years. Primary treatment-related effects were observed in the nasal turbinate mucosa of all rats examined in the 80 ppm group as well as in most of the rats in the 20 ppm group. The changes in both groups were qualitatively similar but much less severe in the 20 ppm group than in the 80 ppm group. The main tumours observed in rats exposed to acrylonitrile were microscopic brain tumours and Zymbal’s gland tumours. In an additional inhalation study (Maltoni et al., 1977; 1988), male and female Sprague-Dawley rats were exposed to 5, 10, 20 or 40 ppm acrylonitrile for 4 hours/day on five days/week for 12 months. A significant increase in the number and proportion of animals bearing tumours was found in several treated groups, although a strong dose-response relationship was not established. Slight to moderate increases in tumour incidence were observed in the mammary gland, forestomach and CNS, but none of these were statistically significant. The authors suggested that the carcinogenicity of acrylonitrile was influenced by the age of the animals at the start of treatment, and was dependent on the concentration administered and duration of treatment. The results were considered to indicate a borderline carcinogenic effect.

Classification for carcinogenicity

Acrylonitrile has a harmonised classification for carcinogenicity according to the CLP Regulation in Category 1B. However recent reviews of the extensive epidemiology dataset have concluded that there is a lack of association between acrylonitrile exposure and increased cancer risk in exposed workers. Given the consistent and clear observations of a lack of effect in occupationally exposed humans, it would seem more appropriate to reinterpret the hazard as a 'suspected human carcinogen' (CLP/GHS Category 2). The reason for the difference between findings in occupationally exposed humans and the results of the animal studies is unclear, however this could be due to either the MoA in rats not being relevant to humans, or that acrylonitrile carcinogenicity is a threshold effect and human exposure does not exceed this threshold. Acrylonitrile was considered by previous IARC Working Groups in 1979 and 1987. Since the previous evaluations, new data were incorporated into the monograph and taken into consideration in the most recent (1999) evaluation. It was concluded that there was no significant excess risk for any type of cancer when all exposed workers were compared with unexposed, or with an external comparison population.

Mechanism of action

The induction of tumours at only relatively late stages in chronic rodent toxicity studies with acrylonitrile is consistent with a non-genotoxic mode of action. Although research is presently unable to fully define a mode of action for acrylonitrile carcinogenicity, the existence of a threshold principle is entirely plausible based on the existing data. Kirman et al. (2005) were able to show the link between occupational human exposure and the results of the rodent cancer assays by modelling the exposure concentrations based on internal CEO levels. This demonstrated that even the highest occupationally relevant exposure concentrations in humans (which are now no longer permitted since changes in legislation), gave rise to an internal concentration at the very lowest animal exposure levels where significant cancer risk was not apparent in the animals. Whilst a cancer risk in humans at high concentrations cannot presently be entirely ruled out, the occupational exposures presently imposed are clearly below a threshold for cancer. It is notable that at the time of the EU RAR (last data search in 1999), there was little mechanistic evidence to support the threshold carcinogen hypothesis, therefore acrylonitrile was conservatively classed as a non-threshold (direct DNA-acting) carcinogen. As well as recent updates to the extensive epidemiology dataset, numerous recent mechanistic studies have highlighted the link between effects in rat tissue and oxidative stress and the cell transformation capacity of acrylonitrile. Mechanistic data therefore strongly indicate that the carcinogenicity of acrylonitrile occurs through an indirect DNA-reactive mechanism secondary to the induction of oxidative stress in the target tissue, through the (non-genotoxic) transformation of initiated cells, or through a combination of these threshold mechanisms. The limited effect in the developing embryo (compared to direct-acting carcinogens) is also notable.

Klaunig et al. (2010) have proposed a mode of action for the induction of rat glial tumours by acrylonitrile, based upon published results. It is postulated that neoplasia induced by acrylonitrile occurs through non- genotoxic (not directly genotoxic) mechanisms. The key events identified in this mode of action include the induction of oxidative stress (an increase in reactive oxygen species produced and/or a decrease in cellular antioxidants). This may arise through acrylonitrile metabolism, perhaps involving the action of cyanide on the mitochondria. The second key event is the selection of a sub-population of initiated cells and the subsequent proliferation of the resistant cell population through increased DNA synthesis or decreased apoptosis, ultimately leading to neoplastic changes.

The proposed MoA is outlined below:

Step 1: Acrylonitrile exposure and absorption

Step 2: Distribution of acrylonitrile and stable metabolites to glial cells

Step 3: Production of oxidative stress

Step 4a: Induction of oxidative DNA damage resulting in initiated cells

Step 4b: Cell proliferation

Step 5: Clonal expansion of spontaneous or acrylonitrile-induced initiated cells

Step 6: Formation of glial tumours

Toxicity to reproduction

Effects on fertility

The two-generation inhalation study of Nemec et al. (2008) is considered to be key to the assessment of the reproductive toxicity of acrylonitrile, as it includes comprehensive investigation of a number of relevant parameters and uses an appropriate route of exposure. In this study, Sprague-Dawley rats (25/sex/group) were exposed (whole body) by inhalation (6 hours/day) to acrylonitrile vapour at concentrations of 0, 5, 15, 45 or 90 ppm. F0 animals were exposed for 10 weeks prior to mating and throughout mating, gestation and lactation of the subsequent F1 litters. Selected F1 offspring were similarly treated and mated to produce F2 litters. In addition to standard reproductive indices, the study included assessment of oestrus cyclicity and sperm parameters. Post mortem investigations of parental animals included detailed histopathological assessment of the reproductive system and associated organs/tissues, detailed histopathological assessment of brain and nasal tissues. Offspring were additionally investigated for developmental ontogeny. F1 animals exposed to 90 ppm acrylonitrile showed excessive toxicity; therefore this exposure level was not investigated further. Mortality was unaffected by exposure. Systemic toxicity in exposed adult rats was limited to reduced weight gain and food consumption and increased liver weights at 45 and 90 ppm. Local toxicity (nasal irritation) was apparent during and immediately following exposure to 90 ppm; histopathological effects on the nasal tissues consistent with local irritation were also seen in some animals in all exposure groups. There was no evidence of any effect on reproductive parameters, tissues or organs of the reproductive system. Effects on offspring were limited to bodyweight effects at parentally toxic exposure concentrations.

In a three-generation study (Beliles et al., 1980), Sprague-Dawley rats (20 males, 10 females/group) were exposed to acrylonitrile in the drinking water at concentrations of 0, 100 or 500 ppm for 100 days prior to mating and subsequently throughout mating, gestation and lactation of the resulting litters. Two litters were produced per generation and exposure continued over three generations. No effects on fertility were seen in this study. Toxicity (reduced weight gain, food and water consumption) was seen at 500 ppm, with less marked effects also seen at 100 ppm. Offspring toxicity (reduced growth and survival) was also noted at 500 ppm; findings are considered most likely to be secondary to parental toxicity. A one-generation study (Schwetz et al., 1975) did not indicate any effects on fertility or reproduction following the administration of acrylonitrile in the drinking water at concentrations of up to 500 ppm. High dose effects seen in parental animals (reduced bodyweights, food and water consumption) are considered likely to be secondary to the reduced palatability of the drinking water. Gross necropsy revealed gastric haemorrhage at the highest dose level; however findings may be confounded by the presence of nematode infection. Findings in offspring were limited to bodyweight effects. One litter in each of the treatment groups had an indication of short or absent tail; however in the absence of a clear dose-response relationship this finding is not considered to be treatment-related. Other, non-standard published investigative studies have reported adverse effects on sperm parameters in mice and rats following exposure to acrylonitrile by inhalation or gavage. However the reliability and/or relevance of these studies are limited due to the use of very high dose levels or methodological deficiencies. It is notable that effects on sperm parameters were not reported in other, standard studies, including the two-generation inhalation toxicity study of Quast et al., or in the NTP mouse studies. None of the studies available report an adverse effect on fertility or reproduction; findings in offspring were limited to reduced growth and survival secondary to maternal toxicity or attributable to direct exposure to acrylonitrile. Neal et al. (2009) have critically reviewed the data on the reproductive toxicity data on acrylonitrile and conclude that exposure to acrylonitrile is unlikely to be associated with any effects on fertility or reproduction.

Developmental toxicity

Murray et al. (1978) exposed groups of 30 pregnant Sprague-Dawley rats to to 0, 40, or 80 ppm of acrylonitrile on 6 hrs/day by inhalation on GD 6-15 to investigate the effect of maternally inhaled acrylonitrile on embryofoetal development. Maternal toxicity was apparent at the highest exposure level of 80 ppm and was associated with an increased incidence of total foetal malformations in this group. Another inhalation study (Saillenfait et al., 1993) reports maternal toxicity and reduced foetal weight at exposure concentrations of 25, 50 and 100 ppm; no treatment-related effect on malformation incidence was seen in this study. A gavage study (Murray et al., 1976) performed at dose levels of 0, 10, 25 or 65 mg/kg bw/d reports maternal toxicity at 25 and 65 mg/kg bw/d; toxicity at 65 mg/kg bw/d was marked and was associated with increased post-implantation loss, reduced foetal weight and length. No effects were seen in the other treated groups. A higher incidence of short tail was seen in foetuses at 65 mg/kg bw/d, with other abnormalities and skeletal variations also increased in this group. No effects were seen in the lower dose groups.

In a non-standard study designed to investigate biochemical and developmental effects (Mehrotra et al., 1988), pregnant Wistar rats were gavaged with 0 or 5 mg/kg bw/d acrylonitrile on GD 5-21. Treatment with acrylonitrile had no effect on maternal or pup bodyweight, gestation length, litter size, sex, the onset of pinna detachment, eye opening, incisor eruption or fur appearance. Pups showed no abnormalities in the development of righting reflex, cliff avoidance, grip strength, spontaneous motor activity or learning ability. Specific biochemical changes were detected in the brains of offspring at Week 3; however the significance of these effects is unclear. In a non-standard study (Saillenfait & Sabate, 2000) four pregnant Sprague-Dawley rats were exposed to a single high gavage dose of acrylonitrile (100 mg/kg bw) on GD 1 and embryos evaluated on GD 12. Maternal toxicity was evidenced by decreased body weight, tremors, piloerection and prostration. Embryos showed misdirected allantois or caudal extremities in 3 of the 4 litters evaluated. The study is of unconventional design, and only a small number of litters were evaluated. However, the results suggest that single high-dose exposures to acrylonitrile may prove developmentally toxic. The embryotoxic effects were seen in the presence of severe maternal toxicity. Willhite et al. (1981) exposed pregnant hamsters to acrylonitrile by intraperitoneal injection at dose levels of 4.8, 10, 25, 65, 80 or 120 mg/kg bw on GD 8. Signs of maternal toxicity were seen at 80 and 120 mg/kg bw/d, with 100% mortality at the highest dose level. The dose level of 80 mg/kg bw was reported resulted in encephalocoele (7/51 foetuses), rib fusions and bifurcations in many of the offspring. Overall the study suggests that acrylonitrile may have developmental effects in the hamster, but only at dose levels which are maternally toxic. The developmental toxicity studies and relevant findings from the reproductive toxicity studies are reviewed by Neal et al (2009). The authors conclude that, overall, the animal studies indicate that very high exposure levels of acrylonitrile resulting in maternal toxicity result in foetal toxicity and potentially teratogenicity. It is noted that teratogenicity appears to be most likely following oral gavage exposure, which is not a relevant route of exposure for humans and the highest quality studies do not indicate clear evidence of teratogenicity.

Classification for toxicity to reproduction

There is no convincing evidence that acrylonitrile can cause effects on fertility or reproduction; non-standard studies report high dose effects on sperm parameters, however similar findings are not apparent in more reliable standard studies. There is some evidence from animal studies that exposure to high levels of acrylonitrile (sufficient to cause overt maternal toxicity) can cause developmental toxicity. Neal et al. (2009) have reviewed other earlier studies which investigated potential reproduction effects. In the most reliable studies they found some evidence of foetal malformations at high concentrations (80 ppm and above); short tail, missing vertebrate, short trunk etc. One tailless pup was found in the industry sponsored Friedman & Beliles (2002) three generation study, again at the high dose. One pup at the high dose in Nemec et al. (2008) reportedly had a threadlike tail and absence of caudal vertebrae. While the numbers of affected foetuses were not significant in these studies; it is proposed that the effects are recognised in the classification of acrylonitrile as an indication of potential hazard. However given the low level incidences of findings in the studies available, classification of acrylonitrile in CLP/GHS Category 2 (suspected human reproductive toxicant) is appropriate. Classification in this category is proposed as there is some evidence from experimental animals of an adverse effect on development. Numbers of observed malformations were low, just above the historical control population frequency for malformations.

Neurotoxicity

The neurotoxicity of acrylonitrile is recognised, however the mechanism underlying these effects are complex and have not yet been fully elucidated. Postulated mechanisms include the direct neurotoxicity of acrylonitrile, neurotoxicity due to an active metabolite, the liberation of cyanide by the oxidative metabolism of acrylonitrile or an effect on cholinergic transmission. Gagnaire et al. (1998) investigated the neurotoxicity of acrylonitrile in rats exposed either orally or by inhalation. Rats given acrylonitrile orally developed behavioural effects characterised by salivation, locomotor hyperactivity and stereotypy. In the high dose group, animals developed weaknesses in hind limbs associated with decreases in sensory nerve conduction velocity (SCV) and action potential amplitude (ASAP). Rats exposed to acrylonitrile vapour exhibited time- and concentration-dependent decreases in motor nerve conduction velocity (MCV), SCV and ASAP, which were partially reversible after 8 weeks of recovery. Rongzhu et al. (2007) investigated the neurobehavioral effects of acrylonitrile administered to rats in drinking water. Three tests, including the open field test, rotarod test and spatial water maze, were applied to evaluate locomotor activities, motor co-ordination and learning and memory, respectively, prior to initiation of the treatment and at Weeks 4, 8 and 12 post-exposure. There were no consistent changes in the open field test, except for locomotion and grooming episodes. In the rotarod test, acrylonitrile exposure significantly decreased latency to fall in a dose and time-dependent manner. In the spatial water maze test, rats exposed for 12 weeks had significantly more training times and longer escape latencies than control animals. The authors conclude that exposure to acrylonitrile induces neurobehavioral alterations in rats. In a recent review of the neurotoxicity of acrylonitrile, Maurissen (2010) concludes that, while the central nervous system has been identified as a target organ for acrylonitrile in a number of animal studies, closer examination of a number of papers does not confirm the reported effects. It is concluded that the mechanism of acrylonitrile-induced neurotoxicity is still unknown and most likely multifaceted. Campian & Benz (2008) investigated the role of cyanide production in the acute neurotoxicity of acrylonitrile in rats concluded there was no evidence that brain ATP is depleted when respiration ceases in acrylonitrile intoxicated rats.

No classification is proposed for neurotoxicity in the absence of any clear indication of direct neurotoxicity in exposed humans.

Specific investigations

A large number of studies have been performed in order to further elucidate the carcinogenic mode of action of acrylonitrile. Studies have focussed on the induction of oxidative stress and cell transformation. The dataset on cell transformation has also been reviewed in the EU RAR (2004), by the Sapphire Group (2004) and more recently by Albertini (2009). In a study in cultured human astrocytes, Jacob & Ahmed (2003) conclude that a redox imbalance may play a major role in acrylonitrile-induced neurotoxicity, which is indicated by compromised antioxidant defence mechanisms (depletion of GSH, increase in GSSG, inhibition of catalase, increased ROS formation and TNF-α secretion), resulting in oxidative DNA damage. In a series of studies in cultured Syrian Hamster Embryo cells, Zhang et al. (2002) suggest that the induction of oxidative stress by acrylonitrile involves a temporal decrease in antioxidants and increase in xanthine oxidase activity that is mediated by the oxidative metabolism of acrylonitrile. Pu et al. (2009) investigated oxidative stress and DNA damage in rats exposed to acrylonitrile in the drinking water. No significant increase in direct DNA strand breaks was observed in brain and WBC from acrylonitrile-treated rats. However, oxidative DNA damage (fpg comet and 8'hydroxyl-2-deoxyguanosine) in brain and WBC was increased in a dose-dependent manner. In addition, plasma levels of reactive oxygen species (ROS) increased in rats administered acrylonitrile. Dietary supplementation with NAC prevented acrylonitrile-induced oxidative DNA damage in brain and WBC. A slight, but significant, decrease in the GSH:GSSG ratio was seen in brain at acrylonitrile doses of >30 ppm. The authors conclude that the results of this study provide support for a mode of action for acrylonitrile-induced astrocytomas involving the induction of oxidative stress and damage. Significant associations were seen between oxidative DNA damage in WBC and brain, ROS formation in plasma, and the reported tumour incidences. Since oxidative DNA damage in brain correlated with oxidative damage in WBC, the results suggest that monitoring WBC DNA damage maybe a useful tool to assess acrylonitrile-induced oxidative stress in humans. Esmat et al. (2007) demonstrated that acrylonitrile induces oxidative stress in cultured rat glial cells, depleting reduced GSH and causing lipid peroxidation. The anti-oxidant compound N-acetylcysteine was shown to inhibit the oxidative effects of acrylonitrile. In contrast, the administration of dietary antioxidants did not have any notable effect on the levels of cyanoethylvaline globin adducts in female rats administered acrylonitrile in the drinking water for 28 days (Snyder, 2010). A similar study reported by Klaunig & Forney (2010) demonstrated increased 8OHdG formation in rat brain and oxidative DNA damage (but not DNA strand breakage) in peripheral white blood cells. Pu et al. (2009) investigated oxidative stress and DNA damage in rats exposed to acrylonitrile in the drinking water. No significant increase in direct DNA strand breaks was observed in brain and WBC from acrylonitrile-treated rats. However, oxidative DNA damage (fpg comet and 8'hydroxyl-2-deoxyguanosine) in brain and WBC was increased in a dose-dependent manner. In addition, plasma levels of reactive oxygen species (ROS) increased in rats administered acrylonitrile. Dietary supplementation with NAC prevented acrylonitrile-induced oxidative DNA damage in brain and WBC. A slight, but significant, decrease in the GSH:GSSG ratio was seen in brain at acrylonitrile doses of >30 ppm. The authors conclude that the results of this study provide support for a mode of action for acrylonitrile-induced astrocytomas involving the induction of oxidative stress and damage.

The EU RAR (2004) reviews the extensive dataset on the potential of acrylonitrile to cause cell transformation which is informative in relation to the potential of acrylonitrile to cause carcinogenicity by non-genotoxic mechanisms. It concludes that the various studies summarised in this review document indicate that acrylonitrile has the ability to cause cell transformation; it further states that cell transformation appears to be secondary to oxidative damage and is dependent on metabolic activation. The Sapphire Group review (2004) also notes that the available data indicate that acrylonitrile has the ability to cause cell transformation as a consequence of a reduction in cellular antioxidant status. Albertini (2009) notes that the extensive investigations of cell transformation show that the induction of oxidative stress by acrylonitrile is dependent on its metabolism to CEO and cyanide. The report of the North Carolina Science Advisory Board (NCSAB, 2010) also concludes that the mechanistic work by Pu et al. suggests that ROS related to toxicity induced by high doses of acrylonitrile appear to play a critical role in its carcinogenicity in the highly sensitive rat brain. The NCSAB also state there is relevant evidence that acrylonitrile possesses genotoxic and carcinogenic activity by acting indirectly in the production of brain tumours in the rat, and that this may possibly occur via a high-dose mechanism involving oxidative stress and changes in gap junction communication.

Observations in humans

Rongzhu et al. (2005) investigated the neurobehavioural effects of exposure to acrylonitrile in the workplace in Chinese workers. Exposed workers reported increased tension, depression, anger, fatigue and confusion. Performances in the Simple Reaction Time, Digit Span, Benton Visual Retention and Pursuit Aiming II were also poorer among exposed workers compared to unexposed workers. Some of these poor performances in tests were also related to exposure duration. In a review of neurotoxicity by Maurissen (2010), the author does not consider the results of this study to be reliable. In a workplace monitoring study, Muto et al. (1992) report no health effects in workers exposed to a mean concentration of 1.13 ppm. Two studies reporting genotoxic effects in Hungarian workers (Major et al., 1998; 1999) report increased chromosomal aberration in workers exposed to acrylonitrile, however findings are seriously confounded due to the limited exposure data, the presence of liver disease and exposures to other toxic chemicals; no premature centromere division was seen in workers exposed to acrylonitrile. Han et al. (2008) also report a range of health effects including non-specific symptoms such as headache, dizziness, palpitations, chest congestion, insomnia, sore throat and abdominal pain; increased anger state; increase blood pressure, pulse pressure WBC counts and ECG changes. Additional investigations showed a higher incidence of fatty liver and gall bladder cholesterol and significantly increased incidences of chromosomal aberration and micronuclei in the peripheral blood lymphocytes of exposed workers.

Epidemiology studies

Numerous epidemiology studies are available, predominantly investigating the association between occupational exposure to acrylonitrile and cancer incidence/mortality. Only the largest of these was considered by the EU RAR to approach an acceptable statistical power for detection of a significant elevation of cancer risk (12.5% with a power of 80%) associated with exposure. More seriously, accurate quantification of exposure to acrylonitrile was not possible for most of the cohorts studied, and in some cases it is likely that a proportion of the exposed cohort were not, in fact, exposed. The second drawback of the published literature is that many of the cohorts were exposed to levels of acrylonitrile that can be presumed to be higher than current exposure. Thus their direct application to present day workforces is uncertain. To add to these problems, some studies report workers whose exposure was to multiple chemicals, including acrylonitrile, which are known or suspected carcinogens. In addition confounding factors such as cigarette smoking have not been considered in most of these studies. Finally, shortfalls in mortality for acrylonitrile workers, beyond levels explainable by the healthy worker effect, suggest that in some studies findings were incomplete. Notwithstanding these problems there is a large amount of epidemiological information relating to workers exposed to acrylonitrile. Both consistency and inconsistency across the studies can be used as a means for determining causal associations rather than using individual study size. Two meta-analyses exist (Rothman, 1994; Collins & Acquavella, 1998) which endeavour to overcome the power issue by combining consistent studies.

DuPont cohort studies

O'Berg reports two studies of cancer incidence and mortality in DuPont workers. The information in a second study (O'Berg et al., 1985) includes the data from an earlier study (O’Berg, 1980) and therefore supersedes it. The retrospective cohort study of 1345 male employees with potential for exposure to acrylonitrile analysed cancer incidence and mortality from 1956-1976. Expected numbers based on company and national rates were calculated on the basis of person-years. The analyses considered time, payroll classification, occupation, duration of exposure, severity of exposure and latency. Overall, 25 cases of cancer occurred in the exposed cohort compared to with 20.5 expected, based on company rates. Of these, eight were respiratory cancer cases, with 4.4 expected. Excesses were found primarily during the 1970-76 time period among wage roll employees who had worked during plant start-up. A trend toward increased risks was seen with increased duration and severity of exposure. Twenty cancer deaths were found, with 17.4 expected according to company rates. Since many cancer cases were recently diagnosed and were living at the time of publication, it was considered premature to evaluate mortality statistics. The author concluded that the findings, coupled with results from tests in laboratory animals, raise the suspicion that acrylonitrile may be a human carcinogen. The study was updated to 1983 for cancer incidence and to 1981 for mortality. Overall, 43 cancer cases occurred, with 37.1 expected based on company rates. A previously reported excess number of cases of lung cancer remains, but is not as marked (10 observed vs. 7.2 expected). Prostate cancer cases were significantly in excess, with six cases observed and 1.8 expected. Mortality analyses revealed 36 cancer deaths, with 31.6 expected. Of these, 14 were from lung cancer, with 11.6 expected based on company rates. Only one death from prostate cancer occurred, with 1.0 expected. The authors state that prostate cancer excesses have not been reported in any other acrylonitrile studies, and hence their significance is currently difficult to assess.

In a further study (Chen et al., 1987), a cohort of 1083 male employees who had potential for exposure to acrylonitrile between 1944-1970 at a DuPont textile fibre plant were followed to 1981 for mortality and to 1983 for cancer incidence. In total, the 21 cancer deaths were fewer than expected based on either DuPont or US national rates. No significant excesses were seen by primary site. In all, 37 cancer cases occurred as compared with 36.5 expected based on company rates. Five lung cancer cases were observed and 6.9 expected. There were 5 prostate cancer cases as compared with 1.9 expected. Of these, 4 occurred among wage employees during the 1975-1983 period, compared to 0.9 expected. This excess was statistically significant, however there was no effect on mortality. A study by Wood et al. (1998) combines and updates the O’Berg (O’Berg, 1980; O’Berg et al., 1985) and Chen et al. (1987) studies to study the 2559 Orlon male workers from both plants over the time period from 1944 to 1991. This study assessed the risk of cancer mortality and incidence in the cohort with a vital status follow-up 99% complete through 1991. Latency, duration of exposure, highest level of exposure ever experienced, and cumulative exposure were used as indicators of exposure. The average duration of exposure for the workers was 7.6 years with an average cumulative exposure of 57.6 ppm-years. Overall mortality was lower than expected in a comparison with biot the US population and all DuPont employees [454 deaths, standardized mortality ratios (SMR) of 69 and 91, respectively)]. All the cancer death ratios were lower than expected in a similar comparison. The SMR values for specific sites did not differ significantly from the expected values. Mortality from all cancers and from prostate, respiratory, and digestive cancer did not show any significantly associated increases or a consistent pattern suggestive of a dose-response. The cancer morbidity patterns were similarly unremarkable.

The most recent investigation (Symons et al., 2008) updates the mortality experience to 2002 for a cohort of workers exposed to acrylonitrile. SMRs were estimated based on two reference populations: the US population and a regional employee population. Exposure–response analyses were conducted using Cox regression models for cumulative and mean intensity exposure measures. In the cohort of 2548 workers, 839 deaths have occurred with 91 deaths due to respiratory system cancer. Most standardized mortality ratio estimates are at or near no-effects levels. Hazard ratio (HR) estimates indicate no increased mortality risk for respiratory system cancer (adjusted HR 0.96, 95% confidence interval: 0.74, 1.25). The authors conclude that no mortality outcome of a priori interest, principally respiratory system cancer, is associated with increased acrylonitrile exposure among fibre production workers over five decades of follow-up.

NCI cohort study

This study (Blair et al., 1998) was designed to evaluate the relationship between occupational exposure to acrylonitrile and cancer mortality. Workers (18079 white men, 4293 white women, 2191 non-white men, and 897 non-white women) employed in acrylonitrile production or use in the 1950s and up to 1983 were followed to 1989 for vital status and cause of death. Exposure-response relationships were evaluated from quantitative estimates of historical exposures. Tobacco use was determined for a sample of workers to assess potential confounding. Mortality rates between the exposed and unexposed workers in the cohort were compared using the Poisson regression. Analyses by cumulative, average, peak, intensity, duration, and lagged exposure revealed no elevated risk of cancers of the stomach, brain, breast, prostate or lymphatic and hematopoietic systems. Mortality from lung cancer was elevated for the highest quintile of cumulative exposure. When the decile categories were used, the relative risk did not continue to increase at higher levels. Adjustment for cigarette use reduced the risk for lung cancer only slightly. Separate analyses for wage and salaried workers, long-term and short-term workers, fibre and non-fibre plants, and individual plants revealed no clear exposure-response patterns. The authors considered that the results of this study indicate that exposure to acrylonitrile at the levels studied is not associated with an increased relative risk for most cancers of a priori interest. The excess of lung cancer in the highest quintile of cumulative exposure may indicate carcinogenic activity at the highest levels of exposure, but analyses of exposure-response do not provide strong or consistent evidence for a causal association.

UK cohort study

The mortality experience of 2763 men employed between 1950 and 1978 for at least one year at 6 factories involved in the polymerization of acrylonitrile and the spinning of acrylic fibres was followed to the end of 1991 (Benn & Osborne, 1998). Overall, cancer deaths did not exceed the expected numbers. There were, however, excess cancer deaths among the workers in the jobs more highly exposed to acrylonitrile. The excesses did not reach conventional levels of statistical significance apart from an excess of lung cancer among workers under 45 years of age. Detailed analyses provided no consistent support for a causal association between acrylonitrile exposure and carcinogenesis. The limitations of the study, including a lack of information on smoking habits and very limited estimates of acrylonitrile exposure, need to be borne in mind.

Dutch cohort studies

A retrospective cohort study (Swaen et al., 1992) was carried out in The Netherlands to investigate the potential carcinogenic effects in humans of occupational exposure to acrylonitrile. The total study group consisted of 6803 workers from eight chemical plants and one control plant, of whom 2842 had been exposed to acrylonitrile between January 1, 1956 and July 1, 1979 for at least 6 months. All workers were employed by one of eight chemical companies. An extensive review of the available industrial hygiene data was conducted to assess the magnitude of past exposure to acrylonitrile, occurrence of peak exposures, exposure to recognized potential human carcinogens, and respirator use. The total cohort was observed for mortality until January 1, 1988. In collaboration with the Central Bureau of Statistics, the causes of death were traced for the workers who died before 01-01-1988. In the exposed as well as in the non-exposed cohorts the total mortality was lower than expected, based on national mortality statistics. The observed cancer mortality in the exposed cohort was similar to the expected mortality. Specific analyses were carried out to investigate dose-response relationships and latency for total mortality and lung cancer mortality. Overall, the authors concluded that no indications were found for a carcinogenic effect in this cohort of workers exposed to acrylonitrile. In a follow-up study (Swaen et al., 1998), all 6803 workers were followed for mortality until 1 January 1996. The follow-up was almost complete (99.6%) and for 99.3% the cause of death was ascertained. Age distribution, follow-up period, and temporal changes in background mortality rates were adjusted for in calculations of standardised mortality ratios for separate causes of death. Cumulative dose-effect relations were determined for 3 exposure categories and 3 latency periods. The results showed that, although cancer mortality fluctuated slightly, no cancer excess seems related to exposure to acrylonitrile. To further investigate the possible carcinogenic effects of acrylonitrile, the authors updated the follow-up (Swaen et al., 2004). The results show that no cancer excess seems related to exposure to acrylonitrile. This additional follow up of a cohort of 2842 workers exposed to acrylonitrile further supports the notion that occupational exposures to acrylonitrile that have occurred in the past have not noticeably increased workers' cancer mortality rates.

Case-control study

Scelo et al. (2004) investigated the incidence of lung cancer in workers exposed to a number of chemicals in several European countries. They conclude that exposure to acrylonitrile was associated with risk of lung cancer. This study adequately controls for smoking, however the study is limited by the small exposed population investigated. The authors additionally caution that the results could be explained by concomitant exposure to other agents commonly in the plastic or textile industries.

Cancer meta-analyses

Twelve published epidemiologic studies that have reported incidence or mortality experience among workers exposed to acrylonitrile were identified by Rothman (1994). Many of the studies contain limited descriptions of subjects, and most do not have good information on exposure assessment. Many also may have suffered from incomplete follow-up, as evidenced by an overall deficit in the number of deaths observed, compared with the number expected from general population mortality rates. Such problems are not unique to studies on acrylonitrile, and to some extent they reflect the difficulties of conducting retrospective cohort studies. Despite these drawbacks, a simplified meta-analysis of the mortality experience reported for these cohorts revealed little evidence for carcinogenicity. Approximately the same number of cancer deaths was observed as was expected according to general population mortality rates (SMR 1.03, 90% confidence interval 0.92-1.15). The combined information from these studies is insufficient to support confidence about a lack of carcinogenicity at all sites. Nevertheless, despite the flaws in some of the individual studies, the summarised findings offer reassurance that workers exposed to acrylonitrile face no striking increases in mortality for all cancers or for respiratory cancer.

Collins & Acqavella (1998) identified twenty-five epidemiologic studies of acrylonitrile workers were reviewed and subjected to meta-analytic techniques in this study to assess the findings for 10 cancer sites. The analyses indicate that workers with acrylonitrile exposure have essentially null findings for most cancers, including lung [meta-relative risk (mRR) 0.9, 95% confidence interval (95% CI) 0.9-1.11, brain (mRR 1.2, 95% CI 0.8-1.7), and prostate (mRR 1.0, 95% CI 0.7-1.4) cancers. Bladder cancer rates were elevated (mRR 1.8, 95% CI 1 .&3.4), but the excess was not dose-related and was limited to plants with aromatic amines. Therefore, the bladder cancer excess is unlikely to be related to acrylonitrile exposure. Some evidence of publication bias was found in the examined literature, but the bias did not have a significant impact on risk estimates for individual cancers. It was concluded that the available studies do not support a causal relation between acrylonitrile exposure and cancer.

Cole et al. (2008) reviewed epidemiology studies published since 1970 and identified through Ovid and MEDLINE. A total of 26 studies which examined mortality and/or incidence rates among persons with AN exposure were identified. Where cohorts have been updated the most recent data were relied upon but descriptions of the earlier publications are provided for background and rationale. Results are provided for all causes of death and all cancers. Detailed results and discussions are provided for the cancers which have received the most attention and for which some positive results have been reported. These include lung, bladder, prostate and central nervous system cancers. In this review the four most informative cohort studies are evaluated and it is apparent that the results do not support a causal relationship between AN exposure and all cancers or any specific type of cancer. The authors also note that the IARC actually downgraded acrylonitrile from 'probably carcinogenic' to 'possibly carcinogenic to humans', finding that ''the earlier indications of an increased risk among workers exposed to acrylonitrile were not confirmed by the recent, more informative studies''. This review of the epidemiology data is consistent with the conclusions of the earlier IARC review which found no consistent findings of increased cancer risk across studies.

Reviews

The Sapphire Group (2004) conclude that the existing epidemiology data do not support an increased cancer risk from acrylonitrile exposure in exposed workers. The EU RAR (2004) concludes that the excess risk of lung cancer from acrylonitrile exposure, if any, is small. For the less common cancers such as brain and prostate it is only possible to evaluate consistency across the available studies. In doing so a relatively imprecise estimate of risk was found for prostate and brain cancers in acrylonitrile workers and acrylonitrile cannot be completely ruled out as the causes of these cancers. However given all the evidence available, in particular the recent studies, there is little or no evidence to support a causal relationship between acrylonitrile exposure and cancer. The IARC re-evaluation monograph (1999) is notable in the conclusion that there is inadequate evidence in humans for the carcinogenicity of acrylonitrile, indicating that the clear evidence of carcinogenicity seen in the rat is not of relevance to exposed humans. Coggon & Cole (1998) review the available extensive dataset on the carcinogenicity of acrylonitrile. Despite its carcinogenicity in animals, the authors conclude that there is little evidence to suggest that the compound causes cancer in humans. The weight of the evidence available suggests either that acrylonitrile is not a human carcinogen or that it does not produce significant cancer levels at occupationally relevant exposure concentrations. A more recent review by Cole et al concludes that the epidemiology data is consistent with the conclusions of the earlier IARC review which found no consistent findings of increased cancer risk across studies. A recent review by the North Carolina Science Advisory Board (NCSAB, 2010) notes that there is relevant modern mechanistic evidence that acrylonitrile possesses genotoxic and carcinogenic activity by acting indirectly in the production of brain tumours in the rat, possibly via a high-dose mechanism involving oxidative stress and changes in gap junctional communication. The NCSAB do not propose a cancer-based AAL for acrylonitrile due to unconvincing evidence for carcinogenicity in humans the relevance of extrapolating findings from rodent species to humans. Bofetta et al. (2008) cite acrylonitrile as a chemical for which initial apparently positive reports of carcinogenicity have proved likely to be false positives, based on subsequent investigations.

Reproductive and developmental effects

Collins et al. (2003) reviewed four epidemiology studies performed on Chinese workers exposed to acrylonitrile and reporting reproductive and developmental effects following maternal and/or paternal exposure. The results of the studies are consistent for an increased risk of stillbirth, birth defects, miscarriage, infertility and low birthweight. However the absence of details on the timing of exposure and on the levels of personal exposure are considered to limit the value of the studies as a clear temporal link and relationship between exposure level and adverse outcome cannot be concluded. While the individual studies have several strengths and are largely consistent with each other in indicating that exposure to acrylonitrile is associated with various adverse reproductive and perinatal outcomes, there are limitations which make a clear causal assessment difficult. In addition, experimental studies provide little support for biological plausibility. It is concluded that additional studies with better designs are required to further investigate the suggested association between acrylonitrile exposure and adverse reproductive outcome. Neal (2009) concluded that, while some Chinese epidemiological studies reported increased rates of birth defects, stillbirths, premature delivery, infertility, and spontaneous abortions; these findings may be artefacts of differential ascertainment, reporting bias, confounding exposures, or other factors. The epidemiological studies are not sufficiently robust, free of potential confounders, or adequately documented for exposure to be an appropriate basis for risk assessment.

Direct observations

In a study designed primarily to investigate possible biomarkers of acrylonitrile exposure, Jakubowski et al. (1987) report that a single eight hour experimental inhalational exposure to concentrations of 5-10 mg/m3 acrylonitrile produced no subjective symptoms such as headache, nausea, or general weakness described at a similar level of industrial exposure. Similarly Sakurai et al. (1978) report that symptoms of irritancy were associated with exposures well in excess of 5ppm, indicating that levels less than 10ppm did not cause notable irritancy. In a biomonitoring study, Fustinoni (2007) failed to detect acrylonitrile in the urine of ABS moulding workers. Observations of accidental poisoning incidents indicate that acrylonitrile is toxic by the oral, dermal and inhalation routes, that it causes neurotoxic effects (which can be due both to acrylonitrile itself and to the release of cyanide), and that exposure in children may be fatal.

Additional information

Kirman et al. (2005) used the results of PBPK modelling to predict internal doses of CEO in the brain, dose-response data for brain tumour induction in the rat and mechanistic data indicating a non-direct genotoxic mode of action for carcinogenicity to derive benchmark doses for tumour incidence. Using a non-linear model, it was concluded that oral doses below 0.009 mg/kg bw/d and air concentrations below 0.1 mg/m3 are not expected to pose an appreciable risk to human populations exposed to acrylonitrile. The same authors (Kirman et al. (2008) used a PBPK model which describes the toxicokinetics of acrylonitrile and CEO and BMD approaches to derive reference values for non-cancer effects. Irritation and neurological effects were selected as the critical endpoints. From this assessment, sub-chronic and chronic oral RfD values of 0.5 and 0.05 mg/kg bw/d, respectively, were derived. Similarly, sub-chronic and chronic inhalation RfC values of 0.1 and 0.06 mg/m3, respectively, were derived.

Derivation of DNELS for acrylonitrile

A large number of aspects of the toxicity of acrylonitrile need to be taken into account when considering DNEL derivation. Acrylonitrile is a multi-site carcinogen in rodent studies; however comprehensive epidemiological data do not indicate any excess of tumours in exposed workers. Mechanistic data suggest that the MoA underlying rodent carcinogenesis is due to indirect genotoxicity associated with oxidative stress and is therefore assumed to be a threshold mechanism with a non-linear dose-response relationship. The absence of carcinogenicity in exposed workers indicates either that acrylonitrile is a rodent-specific carcinogen or the existence of a similar (threshold) mechanism. Based on the epidemiological data it can be concluded that the current OEL of 2 ppm [4.34 mg/m3] provides sufficient protection to workers with regard to carcinogenicity. Nevertheless, it is recognised that acrylonitrile has a harmonised CLP classification as a carcinogen in Category 1B; therefore derivation of systemic inhalation DNEL values is not possible and a qualitative risk assessment is appropriate with the aim of minimising exposure through the use of appropriately stringent Risk Management Measures (RMMs) and Operational Conditions (OCs). For the general population, direct inhalation exposure to acrylonitrile (as an industrial monomer and intermediate) is highly unlikely. Direct inhalaton exposure may theoretically be relevant for resident in the vicinity of an industrial facility either manufacturing or using acrylonitrile. Long-term systemic effects, based on the potential for carcinogenicity are considered. Kirman et al. (2005) used all the available rat cancer data to determine Human Equivalent Values, based on PBPK modelling of the CEO concentrations. This would be an appropriate assessment for general population systemic exposure.

Acrylonitrile does not have any effect on fertility or reproductive capacity; however there is some evidence for developmental toxicity at relatively high levels of exposure, particularly following gavage dosing. Data clearly show that developmental toxicity (including low incidences of foetal malformation) is only seen at exposure levels also causing overt maternal toxicity. Acrylonitrile does not have a harmonised classification for developmental toxicity; however classification in Category 2 is proposed based on the limited evidence of developmental toxicity at maternally toxic dose levels. Chinese epidemiological studies have reported increased rates of birth defects, stillbirths, premature delivery, infertility and spontaneous abortions associated with exposure to acrylonitrile. These findings may be artefacts of differential ascertainment, reporting bias, confounding exposures, or other factors. The epidemiological studies are not considered to be sufficiently robust, free of potential confounders, or adequately documented for exposure to be an appropriate basis for causal assessment. Based on the multiple animal studies, acrylonitrile is not expected to be a developmental or reproductive toxicant in the absence of significant parental toxicity. Reproductive and developmental toxicity is not therefore a critical concern for acrylonitrile, taking into account the potential for other toxic effects.

Evidence suggests that acrylonitrile may be neurotoxic following acute and repeated exposures. Neurotoxicity is likely to account at least in part for the acute toxicity of acrylonitrile, which has a harmonised CLP classification for acute oral, dermal and inhalation toxicity in Category 3. In the absence of a suitable dose descriptor, systemic DNELs cannot be derived for acute exposures by any route- a qualitative approach is therefore proposed for short-term dermal systemic effects for workers and for the general population and for short-term oral effects in the general population. Acrylonitrile is also classified as a skin irritant and skin sensitiser; therefore appopriate RMMs and the use of protective equipment will minimise the potential for occupational exposure by the dermal route. Acrylonitrile is used only for the manufacture of resins and other polymers used in the manufacture of articles and is not supplied as the unchanged monomer to the general population. There is therefore no potential for the direct oral or dermal exposure of the general population to acrylonitrile. With regard to acute inhalation exposure, the local inhalation DNELs for workers are protective of any potential for neurotoxicity. Acrylonitrile is used only for the manufacture of resins and other polymers used in the manufacture of articles and is not supplied as the unchanged monomer to the general population. There is therefore no potential for the direct inhalation exposure of the general population to acrylonitrile. Kirman et al. (2008) considered the critical non-cancer systemic effect of oral exposure to acrylonitrile to be neurotoxicity; this endpoint is therefore used to derive the long-term systemic dermal DNEL values for workers.

Both animal data (as summarised by Kirman et al., 2008) and human data (including Muto et al., 1992; Jakubowski et al., 1987; Sakurai et al., 1978) indicate that respiratory tract irritation is the critical relevant effect of inhalation exposure to acrylonitrile. Irritation thresholds in humans are identified and are therefore used as the basis for the derivation of local inhalation DNEL values for workers and for the general population. Acrylonitrile is classified as a skin irritant (CLP Category 1) and skin sensitiser (CLP Category 1B). Quantitative local dermal DNELs cannot be derived; therefore a qualitative approach is taken for the risk assessment, involving the use of appropriate RMMs and personal protective equipment for workers. These measures will limit the potential for dermal contact of workers with acrylonitrile, thereby eliminating any concerns over the potential for systemic toxicity resulting from aucte or repeated dermal exposure. Acrylonitrile is used only for the manufacture of resins and other polymers used in the manufacture of articles and is not supplied as the unchanged monomer to the general population. There is no potential for the direct dermal exposure of the general population to acrylonitrile; therefore there is no potential for skin irritation or sensitisation.

-------------------------------------------------------------------------

Worker DNEL derivation:

Inhalation DNELs

Systemic inhalation DNEL (long-term)

Acrylonitrile is classified as a Category 1B carcinogen (H350: may cause cancer) according to CLP and is therefore assigned to the High Hazard category according to the ECHA Practical Guide R15 (2012). While there is clear evidence that acrylonitrile is a multi-site rodent carcinogen in studies including those using the inhalation route of exposure, the comprehensive epidemiological evidence does not show any increase in the incidence of tumours in occupationally exposed workers. Furthermore while evidence also suggests that acrylonitrile may be genotoxic in rodents, consideration of the comprehensive mechanistic dataset indicates that this activity is most likely to be indirect genotoxicity associated with an oxidative stress response. Acrylonitrile is therefore most likely to be a threshold carcinogen in the rodent, which is supported by the lack of carcinogenicity in the epidemiology studies. Nevertheless occupational exposure to acrylonitrile should be kept to a minimum.

Acrylonitrile is also shown to be a respiratory irritant in animal studies and in exposed workers. Respiratory irritation was noted to be the most sensitive effect of exposure in some repeated dose inhalation studies and effects are reported in workers at low exposure concentrations. The long-term local inhalation DNEL proposed for acrylonitrile of 1.8 mg/m3 is considered to be adequate in controlling worker exposure and ensuring that workers are not exposed to concentrations of acrylonitrile likely to cause carcinogenicity.

Based on the High Hazard Category, the following Risk Management Measures (RMMs) and Operational Conditions (OCs) are therefore implemented for acrylonitrile:

Any measure to eliminate exposure should be considered;

Very high level of containment required, except for short term exposures e.g. taking samples;

Design closed systems to allow for easy maintenance;

If possible keep equipment under negative pressure;

Control staff entry to work area;

Ensure that all equipment is well maintained;

Permit to work for maintenance work;

Regular cleaning of equipment and work area;

Management/supervision in place to check that the RMMs in place are being used correctly and OCs followed;

Training for staff on good practice;

Procedures and training for emergency decontamination and disposal;

Good standard of personal hygiene;

Recording of any 'near miss' situations;

The following Personal Protective Equipment (PPE) is appropriate for acrylonitrile:

Substance/task appropriate respirator;

Substance/task appropriate gloves;

Full skin coverage with appropriate barrier material;

Chemical goggles.

Systemic inhalation DNEL (short-term)

Acrylonitrile is classified for acute inhalation toxicity in CLP Category 3 (H331: Toxic if inhaled) based on an LC50 value of 2050 mg/m3 and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). Human exposure data also demonstrate the potential for toxicity (including lethality in a small number of cases) following inhalation exposure. The data are insufficient for the derivation of a short-term systemic inhalation DNEL. It should be noted, however, that respiratory tract irritation is also observed following inhalation exposure to acrylonitrile and the short-term local inhalation DNEL proposed for acrylonitrile of 10 mg/m3 is protective of any potential systemic toxicity.

The following Risk Management Measures (RMMs) and Operational Conditions (OCs) are appropriate for acrylonitrile.

Appropriate containment;

Minimise number of staff exposed;

Segregation of the emitting process;

Effective extraction;

Good standard of general ventilation;

Minimisation of manual phases;

Avoidance of contact with contaminated tools and objects;

Regular cleaning of equipment and work area;

Management/supervision in place to check that the RMMs in place are being used correctly and OCs followed;

Training for staff on good practice;

Good standard of personal hygiene.

The following Personal Protective Equipment (PPE) is appropriate for acrylonitrile:

Substance/task appropriate gloves;

Skin coverage with appropriate barrier material based on potential for contact;

Substance/task appropriate respirator;

Optional face shield;

Eye protection.

Local inhalation DNEL (long-term)

The proposed DNEL is 0.8 ppm (1.8 mg/m3), based on personal sampling in the worker monitoring study of Muto (1992). Correction of the endpoint and the use of assessment factors are not required as the endpoint is derived from a worker monitoring study and the endpoint is considered to be applicable to all workers.

Local inhalation DNEL (short-term)

Acrylonitrile is a respiratory irritant; animal studies and reports of human exposure identify local irritant effects on the respiratory tract as sensitive effects of exposure. The 2004 EU RAR for acrylonitrile details reports of workers with initial symptoms of toxic effects as well as respiratory irritation at concentrations of 16-100 ppm within 20 minutes (Wilson et al., 1948). Jakubowski et al. (1987) report that an exposure to 2.3-4.6 ppm acrylonitrile was without effect in volunteers exposed for consecutive 8-hour periods. The level of 4.6 ppm [10 mg/m3] is therefore proposed as the short-term local inhalation DNEL for workers. While this study was conducted with six volunteers, it is proposed that additional assessment factors are not required as the exposure period of 8 hours considered to be adequately protective of short-term peak exposures to acrylonitrile.

Dermal DNELs

Systemic dermal DNEL (long-term)

No repeated dose dermal toxicity studies are available; however it is known that accidental dermal exposure to acrylonitrile may results in signs of toxicity. Systemic effects following dermal exposure are therefore possible in exposed workers. Kirman (2008) proposed that neurotoxicity was the most sensitive non-cancer systemic end-point identified following oral exposure. Based on the Kirman (2008) analysis of the study of Gagnaire et al. (1998) which identifies effects on sensory nerve conduction velocity, the relevant endpoint is the human equivalent oral BMDL05 of 8.5 mg/kg bw/d. The human equivalent value is derived using PBPK modelling. This DNEL derivation assumes that dermal absorption is the same as oral absorption (the default assumption). Individual ECETOC assessment factors of 3 (for interspecies differences) and 2 (for duration of exposure) result in an overall assessment factor of 6. Applying the overall assessment factor to the corrected starting point (human equivalent value) results in a long-term systemic dermal DNEL for workers of 1.4 mg/kg bw/d. The requirement for a long-term worker dermal DNEL is debatable given the RMM required to avoid contact with a substances known to be a skin irritant and sensitiser; however, it is considered appropriate to establish a systemic dermal DNEL in order to assess risks of possible long term indirect dermal exposure, even where PPE or technical controls may be present.

Systemic dermal DNEL (short-term)

Acrylonitrile is classified for acute dermal toxicity in CLP Category 3 (H331: Toxic in contact with skin) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). Human exposure data also demonstrate the potential for toxicity following dermal exposure. The data are insfficient for the derivation of a short-term systemic dermal DNEL

The following Risk Management Measures (RMMs) and Operational Conditions (OCs) are appropriate for acrylonitrile.

Appropriate containment;

Minimise number of staff exposed;

Segregation of the emitting process;

Effective extraction;

Good standard of general ventilation;

Minimisation of manual phases;

Avoidance of contact with contaminated tools and objects;

Regular cleaning of equipment and work area;

Management/supervision in place to check that the RMMs in place are being used correctly and OCs followed;

Training for staff on good practice;

Good standard of personal hygiene.

The following Personal Protective Equipment (PPE) is appropriate for acrylonitrile:

Substance/task appropriate gloves;

Skin coverage with appropriate barrier material based on potential for contact;

Substance/task appropriate respirator;

Optional face shield;

Eye protection.

Local dermal DNEL (long-term)

Acrylonitrile is classified for skin sensitisation in CLP Category 1B (H317: May cause an allergic skin reaction) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012).

The following Risk Management Measures (RMMs) and Operational Conditions (OCs) are appropriate for acrylonitrile.

Appropriate containment;

Minimise number of staff exposed;

Segregation of the emitting process;

Effective extraction;

Good standard of general ventilation;

Minimisation of manual phases;

Avoidance of contact with contaminated tools and objects;

Regular cleaning of equipment and work area;

Management/supervision in place to check that the RMMs in place are being used correctly and OCs followed;

Training for staff on good practice;

Good standard of personal hygiene.

The following Personal Protective Equipment (PPE) is appropriate for acrylonitrile:

Substance/task appropriate gloves;

Skin coverage with appropriate barrier material based on potential for contact;

Substance/task appropriate respirator;

Optional face shield;

Eye protection.

Systemic local DNEL (short-term)

Acrylonitrile is classified for skin sensitisation in CLP Category 1B (H317: May cause an allergic skin reaction) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012).

The following Risk Management Measures (RMMs) and Operational Conditions (OCs) are appropriate for acrylonitrile.

Appropriate containment;

Minimise number of staff exposed;

Segregation of the emitting process;

Effective extraction;

Good standard of general ventilation;

Minimisation of manual phases;

Avoidance of contact with contaminated tools and objects;

Regular cleaning of equipment and work area;

Management/supervision in place to check that the RMMs in place are being used correctly and OCs followed;

Training for staff on good practice;

Good standard of personal hygiene.

The following Personal Protective Equipment (PPE) is appropriate for acrylonitrile:

Substance/task appropriate gloves;

Skin coverage with appropriate barrier material based on potential for contact;

Substance/task appropriate respirator;

Optional face shield;

Eye protection.

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.1 mg/m³

Most sensitive endpoint:

carcinogenicity

Route of original study:

Oral

DNEL derivation method:

other: Specific factors used following PBPK modelling and the derivation of a human-specific endpoint

Overall assessment factor (AF):

220

Modified dose descriptor starting point:

other: LED05

Value:

21.3 mg/m³

Explanation for the modification of the dose descriptor starting point:

The relevant point of departure is the PBPK-derived human equivalent LED05 for inhalation exposures of 21.3 mg/m3 from the study of Kirman et al. (2005), based on assessment of all available rat 2-year studies.

AF for dose response relationship:

10

Justification:

A 5% response level reflects a significant response, and cannot be treated as a NOAEL for an effect of this severity. To account for the severity of the response used in this dose–response assessment, a value of 10 was used .

AF for differences in duration of exposure:

1

Justification:

Value derived from studies of chronic duration

AF for interspecies differences (allometric scaling):

1

Justification:

Not required: taken into account as part of the PBPK modelling

AF for other interspecies differences:

3.2

Justification:

Factor used to account for toxicodynamic differences

AF for intraspecies differences:

7

Justification:

A factor of 2.2 was combined with the default factor of 3.2 for human variation in toxicodynamics to give an overall value of 7.0

AF for the quality of the whole database:

1

Justification:

High quality database

AF for remaining uncertainties:

1

Justification:

Comprehensive database; no remaining uncertainties

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

acute toxicity

Justification:

Comprehensive database

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.06 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL derivation method:

other: Specific factors were used following PBPK modelling

Overall assessment factor (AF):

10

Dose descriptor:

BMCL10

Value:

0.64 mg/m³

AF for dose response relationship:

1

Justification:

Not required: BMD methodology was used

AF for differences in duration of exposure:

1

Justification:

Not required; PoD based on chronic and sub-chronic data

AF for interspecies differences (allometric scaling):

1

Justification:

Not required: this was taken into account by PBPK modelling

AF for other interspecies differences:

3

Justification:

A factor of 3 was considered to be appropriate for to account for remaining pharmacodynamic differences

AF for intraspecies differences:

3

Justification:

Because the critical effect (nasal lesions) is not expected to depend upon systemic factors that may vary from one individual to another, a value of 3 was considered to be sufficient to address variation in factors affecting nasal tissue dosimetry

AF for the quality of the whole database:

1

Justification:

High quality database

AF for remaining uncertainties:

1

Justification:

Comprehensive database; no remaining uncertainties

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

1 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL derivation method:

ECHA REACH Guidance

Overall assessment factor (AF):

10

Dose descriptor starting point:

NOAEC

AF for dose response relationship:

1

Justification:

Not required

AF for interspecies differences (allometric scaling):

1

Justification:

Not required: based on human data

AF for other interspecies differences:

1

Justification:

Not required: PoD based on human data

AF for intraspecies differences:

10

Justification:

ECHA default factor

AF for the quality of the whole database:

1

Justification:

High quality database

AF for remaining uncertainties:

1

Justification:

Comprehensive database; no significant remaining uncertainties.

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.009 mg/kg bw/day

Most sensitive endpoint:

carcinogenicity

Route of original study:

Oral

DNEL derivation method:

other: Use of specific assessment factors following PBPK modelling

Overall assessment factor (AF):

200

Modified dose descriptor starting point:

other: LED05

Value:

1.7 mg/kg bw/day

Explanation for the modification of the dose descriptor starting point:

The starting point is the PBPK-derived human equivalent LED05 for oral exposures of 1.7 mg/kg bw/d (Kirman et al., 2005), which is based on analysis of all rat 2-year carcinogenicity bioassays

AF for dose response relationship:

10

Justification:

A 5% response level reflects a significant response, and cannot be treated as a NOAEL for an effect of this severity. To account for the severity of the response used in this dose–response assessment, a value of 10 was used .

AF for differences in duration of exposure:

1

Justification:

Value derived from studies of chronic duration

AF for interspecies differences (allometric scaling):

1

Justification:

Not required: taken into account as part of the PBPK modelling

AF for other interspecies differences:

3.2

Justification:

Factor used to account for toxicodynamic differences

AF for intraspecies differences:

6.4

Justification:

A factor of 2.0 was combined with the default factor of 3.2 for human variation in toxicodynamics to give a value of 6.4

AF for the quality of the whole database:

1

Justification:

High quality database

AF for remaining uncertainties:

1

Justification:

Comprehensive database

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

acute toxicity

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

sensitisation (skin)

Hazard assessment conclusion:

medium hazard (no threshold derived)

Most sensitive endpoint:

sensitisation (skin)

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.009 mg/kg bw/day

Most sensitive endpoint:

carcinogenicity

Route of original study:

Oral

DNEL derivation method:

other: Use of specific assessment factors following PBPK modelling

Overall assessment factor (AF):

200

Modified dose descriptor starting point:

other: LED05

Value:

1.7 mg/kg bw/day

Explanation for the modification of the dose descriptor starting point:

Not required: based on oral data

AF for dose response relationship:

10

Justification:

A 5% response level reflects a significant response, and cannot be treated as a NOAEL for an effect of this severity. To account for the severity of the response used in this dose–response assessment, a value of 10 was used .

AF for differences in duration of exposure:

1

Justification:

Value derived from studies of chronic duration

AF for interspecies differences (allometric scaling):

1

Justification:

Not required: taken into account as part of the PBPK modelling

AF for other interspecies differences:

3.2

Justification:

Factor used to account for toxicodynamic differences

AF for intraspecies differences:

6.4

Justification:

A factor of 2.0 was combined with the default factor of 3.2 for human variation in toxicodynamics to yield a value of 6.4

AF for the quality of the whole database:

1

Justification:

High quality database

AF for remaining uncertainties:

1

Justification:

Comprehensive database

Hazard assessment conclusion:

medium hazard (no threshold derived)

Hazard assessment conclusion:

medium hazard (no threshold derived)

General population DNEL values

Acrylonitrile is not a substance available to the general public. It is an industrial monomer. However, there are certain DNEL values which can be considered as useful, in the usual management of this substance, where a safe level for the public may need to be agreed. Acrylonitrile has been shown to be genotoxic in vitro and is a carcinogen in rodent studies, however the genotoxicity profile is most compatible with an indirect genotoxic MoA for cancer induction in rodents, and not a direct, DNA-reactive MoA. There are exposure levels below which no oxidative stress, and no oxidative DNA damage, has been detected in short term animal tests and concentrations in long term animal cancer bioassays at which the level of neoplastic tumours is indistinguishable from the background in the control animals. This would tend to support the non-direct acting mode of action interpretation. In addition, the absence of a causal link between acrylonitrile exposure and cancer in extensive occupational epidemiology investigations, also supports a view that acrylonitrile is a threshold rodent carcinogen. Exposure of the general population to acrylonitrile both directly (from the use of polymers containing unreacted monomer) and indirectly (via the environment) is likely to be negligible and far below the levels encountered by workers, for whom a causal link between acrylonitrile exposure and cancer has not been demonstrated. Kirman et al. (2005) estimated an inhalation reference value for cancer of 0.1 mg/m3, which was not expected to pose an appreciable risk to the exposed human population, based on the PBPK modelling of CEO levels linked to tumour formation in all the available rodent studies. In a study to determine the non-cancer reference values, based on human equivalent inhalation values BMDL10 of 0.64 mg/m3, driven by observed nasal lesions in rats, with a factor of 10 for a safety factor gives a value of 0.06 mg/m3 (0.027 ppm). This is regarded as protective of inhalation irritation and is lower than the proposed cancer reference value.

Inhalation DNEL values

Systemic inhalation DNELs

Long-term systemic inhalation DNEL

Kirman et al. (2005) conducted a cancer dose–response assessment was conducted for acrylonitrile using updated information on mechanism of action, epidemiology, toxicity, and pharmacokinetics. Although more than 10 chronic bioassays indicate that acrylonitrile produces multiple tumors in rats and mice, a number of large, well-conducted epidemiology studies provide no evidence of a causal association between acrylonitrile xposure and cancer mortality of any type. The epidemiological data include early industry exposures that are far higher than occur today and that approach or exceed levels found to be tumorigenic in animals. Despite the absence of positive findings in the epidemiology data, a dose–response assessment was conducted for acrylonitrile based on brain tumour incidence in rats. Mechanistic studies implicate the involvement of oxidative stress in rat brain due to a metabolite (2 -cyanoethylene oxide; CEO), but do not conclusively rule out a potential role for the direct genotoxicity of CEO. A PBPK model was used to predict internal doses (peak CEO in brain) for 12 data sets, which were pooled together to provide a consistent characterisation of the dose–response relationship for brain tumour incidence in the rat. The internal dose corresponding to a 5% increase in extra risk (ED05 = 0.017 mg/L brain) and its lower confidence limit (LED05 = 0.014 mg/L brain) was used as the point of departure (PoD). The weight-of-evidence supports the use of a non-linear extrapolation for the cancer dose–response assessment. A quantitative comparison of the epidemiology exposure–response data (lung and brain cancer mortality) to the rat brain tumour data in terms of internal dose adds to the confidence in the non-linear extrapolation. An uncertainty factor of 220 for the inhalation route was applied to the LED05 to account for interspecies variation, intraspecies variation, and the severity of the response measure. It is concluded that acrylonitrile air concentrations below 0.1 mg/m3 are not expected to pose an appreciable risk to human populations. The relevant point of departure is the PBPK-derived human equivalent LED05 for inhalation exposures of 21.3 mg/m3 from the study of Kirmanet al. (2005), based on all available rat 2 year studies. Using the assessment factor of 220 specified by the authors results in the derivation of a DNEL (inhalation, long-term, systemic) of 0.1 mg/m3. It is notable that the systemic long-term inhalation DNEL value based on the carcinogenicity of acrylonitrile is higher than the local long-term inhalation DNEL value of 0.06 mg/m3 based on the irritant effects of the acrylonitrile derived by Kirman et al. (2008).

Short-term systemic inhalation DNEL

Acrylonitrile is classified for acute inhalation toxicity in CLP Category 3 (H331: Toxic if inhaled) based on an LC50 value of 2050 mg/m3 and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). Human exposure data also demonstrate the potential for toxicity (including lethality in a small number of cases) following inhalation exposure. The data are insufficient for the derivation of a short-term systemic inhalation DNEL. It should be noted, however, that respiratory tract irritation is also observed following inhalation exposure to acrylonitrile and the short-term local inhalation DNEL proposed for acrylonitrile of 3.3 mg/m3 is highly protective of any potential systemic toxicity resulting from inhalation exposure.

Long-term local inhalation DNEL

Kirman et al. (2008) conducted dose-response assessments were conducted for the non-cancer effects of acrylonitrile for the purposes of deriving an inhalation reference concentration (RfC) value. Based upon an evaluation of available toxicity data, the local irritation effects of acrylonitrile were determined to be an appropriate basis for derivation of the reference concentration. Benchmark dose (BMD) methods were used to derive the reference value. Where sufficient information was available, data-derived uncertainty factors were applied to the points of departure determined by BMD methods. From this assessment, a chronic inhalation RfC of 0.06 mg/m3 was derived. Confidence in the reference value derived for acrylonitrile was considered to be high, based upon a consideration of the confidence in the key studies, the toxicity database, dosimetry, and dose-response modeling.

The chronic reference concentration of 0.06 mg/m3 is based on the BMCL10 of 0.64 mg/m3 for nasal lesions in rats (from the studies of Quastet al., 1980 and Nemecet al., 2005) and applying an assessment factor of 10.

Short-term local inhalation DNEL

Jakubowski et al. (1987) reported that an exposure level of 4.6 ppm acrylonitrile was without effect in 6 volunteers exposed for consecutive 8-hour periods. Using an assessment factor of 10 to coverr variation within the wider population, a level of 0.46 ppm [1 mg/m3] is therefore proposed as a DNEL.

Dermal DNEL values

Systemic dermal DNELs

Long-term systemic dermal DNEL

Kirman et al. (2005) performed a cancer dose–response assessment for acrylonitrile using updated information on mechanism of action, epidemiology, toxicity, and pharmacokinetics. Although more than 10 chronic bioassays indicate that acrylonitrile produces multiple tumors in rats and mice, a number of large, well-conducted epidemiology studies provide no evidence of a causal association between acrylonitrile exposure and cancer mortality of any type. The epidemiological data include early industry exposures that are far higher than occur today and that approach or exceed levels found to be tumorigenic in animals. Despite the absence of positive findings in the epidemiology data, a dose–response assessment was conducted for acrylonitrile, based on brain tumors in rats. Mechanistic studies implicate the involvement of oxidative stress in rat brain due to a metabolite (2-cyanoethylene oxide; CEO), but do not conclusively rule out a potential role for the direct genotoxicity of CEO. A PBPK model was used to predict internal doses (peak CEO in brain) for 12 data sets, which were pooled together to provide a consistent characterisation of the dose–response relationship for brain tumour incidence in the rat. The internal dose corresponding to a 5% increase in extra risk (ED05 = 0.017 mg/L brain) and its lower confidence limit (LED05 = 0.014 mg/L brain) was used as the point of departure (PoD). The weight-of-evidence supports the use of a non-linear extrapolation for the cancer dose–response assessment. A quantitative comparison of the epidemiology exposure–response data (lung and brain cancer mortality) to the rat brain tumour data in terms of internal dose adds to the confidence in the non-linear extrapolation. An uncertainty factor of 200 was applied to the LED05 to account for interspecies variation, intraspecies variation, and the severity of the response measure. Accordingly, doses below 0.009 mg/kg bw/d are not expected to pose an appreciable risk to human populations exposed to acrylonitrile. The assessment was performed for oral exposure; however this is also used for demral exposure (as a worst case), assuming that oral and dermal absorption are equivalent.

Short-term systemic dermal DNEL

Acrylonitrile is classified for acute dermal toxicity in CLP Category 3 (H331: Toxic in contact with skin) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). Human exposure data also demonstrate the potential for toxicity following dermal exposure. The data are insfficient for the derivation of a short-term systemic dermal DNEL, therefore a qualitative approach is appropriate.

Local dermal DNELs

Long-term local DNEL

Acrylonitrile is classified for skin sensitisation in CLP Category 1B (H317: May cause an allergic skin reaction) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). A qualitative risk assessment is therefore appropriate.

Short-term local DNEL

Acrylonitrile is classified for skin sensitisation in CLP Category 1B (H317: May cause an allergic skin reaction) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). A qualitative risk assessment is therefore appropriate.

Oral DNEL values

Long-term systemic oral DNEL

Kirman et al. (2005) performed a cancer dose–response assessment for acrylonitrile using updated information on mechanism of action, epidemiology, toxicity, and pharmacokinetics. Although more than 10 chronic bioassays indicate that acrylonitrile produces multiple tumors in rats and mice, a number of large, well-conducted epidemiology studies provide no evidence of a causal association between acrylonitrile exposure and cancer mortality of any type. The epidemiological data include early industry exposures that are far higher than occur today and that approach or exceed levels found to be tumorigenic in animals. Despite the absence of positive findings in the epidemiology data, a dose–response assessment was conducted for acrylonitrile, based on brain tumors in rats. Mechanistic studies implicate the involvement of oxidative stress in rat brain due to a metabolite (2-cyanoethylene oxide; CEO), but do not conclusively rule out a potential role for the direct genotoxicity of CEO. A PBPK model was used to predict internal doses (peak CEO in brain) for 12 data sets, which were pooled together to provide a consistent characterisation of the dose–response relationship for brain tumour incidence in the rat. The internal dose corresponding to a 5% increase in extra risk (ED05 = 0.017 mg/L brain) and its lower confidence limit (LED05 = 0.014 mg/L brain) was used as the point of departure (PoD). The weight-of-evidence supports the use of a non-linear extrapolation for the cancer dose–response assessment. A quantitative comparison of the epidemiology exposure–response data (lung and brain cancer mortality) to the rat brain tumour data in terms of internal dose adds to the confidence in the non-linear extrapolation. An uncertainty factor of 200 was applied to the LED05 to account for interspecies variation, intraspecies variation, and the severity of the response measure. Accordingly, doses below 0.009 mg/kg bw/d are not expected to pose an appreciable risk to human populations exposed to acrylonitrile.

Short-term oral systemic DNEL

Acrylonitrile is classified for acute oral toxicity CLP Category 3 (H301: Toxic if swallowed) and is therefore assigned to the Moderate Hazard category according to ECHA Practical Guide 15 (2012). A qualitative risk assessment is therefore appropriate.

Registration Dossier

Registration Dossier

Data platform availability banner - registered substances factsheets

Diss Factsheets

Acrylonitrile

Toxicological Summary

Administrative data

Workers - Hazard via inhalation route

Systemic effects

Long term exposure

Acute/short term exposure

DNEL related information

Local effects

Long term exposure

DNEL related information

Acute/short term exposure

DNEL related information

Workers - Hazard via dermal route

Systemic effects

Long term exposure

DNEL related information

Acute/short term exposure

DNEL related information

Local effects

Long term exposure

Acute/short term exposure

Workers - Hazard for the eyes

Local effects

Additional information - workers

General Population - Hazard via inhalation route

Systemic effects

Long term exposure

DNEL related information

Acute/short term exposure

DNEL related information

Local effects

Long term exposure

DNEL related information

Acute/short term exposure

DNEL related information

General Population - Hazard via dermal route

Systemic effects

Long term exposure

DNEL related information

Acute/short term exposure

DNEL related information

Local effects

Long term exposure

Acute/short term exposure

General Population - Hazard via oral route

Systemic effects

Long term exposure

DNEL related information

Acute/short term exposure

DNEL related information

General Population - Hazard for the eyes

Local effects

Additional information - General Population