Benzene - PSL1

3.0 Assessment of "Toxic" under CEPA

As described in the Introduction, the following assessment will consider the entry of benzene to the environment, the exposure of humans and other biota to benzene, and potential harmful effects in humans and other biota.

3.1 Entry

Benzene enters into the Canadian environment primarily through atmospheric releases. Approximately 34 150 tonnes are released yearly to the atmosphere. The major source of release is from combustion of gasoline and diesel fuels, which together account for more than 76% of total atmospheric releases. Light-duty vehicles alone account for 61 % of total releases. Benzene is released to the soil from spills, leaking underground storage tanks, and in leachate from contaminated waste disposal sites. Release to water occurs through spills and discharge of contaminated effluents. Benzene has been measured in Canada in the atmosphere and in certain samples of drinking water, surface water, groundwater, industrial effluents, and leachate from waste disposal sites.

3.2 Exposure

Benzene does not persist in water or soil because it biodegrades and volatilizes rapidly to the atmosphere. It also does not persist in the atmosphere because it undergoes rapid photo-oxidation. Benzene does not appreciably absorb ultraviolet light at wavelengths passing through the upper atmosphere, or infrared radiation at wavelengths of 7 to 13 µm.

Airborne concentrations of benzene in rural areas of Canada are generally below 1.2 µg/m³. Mean concentrations at urban sites have ranged from 1.2 to 14.6 µg/m³, with an overall mean concentration of 4.4 µg/m³ and a maximum 24-hour average recorded at one site of 41.9 µg/m³.

Concentrations of benzene in Canadian surface waters are generally less than 1 µg/L. The mean concentration in untreated water measured in one study was 2 µg/L. The highest reported mean concentration of benzene in effluents has been 65.3 µg/L, measured at an outfall from an organic chemicals industry.

Accumulation of benzene is not expected to be important in any terrestrial or aquatic organism and there are no reports indicating any significant bioconcentration in organisms or biomagnification in the food chain. The main route of exposure for terrestrial biota is, therefore, inhalation rather than exposure via the food chain.

Estimated average daily intakes (on a body weight basis) of benzene from environmental media by various age groups in the general population in Canada are presented in Table 6. These estimates are based on mean concentrations of benzene found in environmental media. Elevated exposure resulting from spills, contaminated groundwater supplies, or other localized conditions were not considered. Ambient air is the main source of exposure to benzene for the general human population, with estimated intakes ranging from 1.3 to 3.0 µg/(kg b.w.·day). Automobile-related activities are estimated to contribute an additional 0.7 to 0.9 µg/(kg b.w.·day), while use of household products, indirectly estimated from the difference between the concentration of benzene in outdoor and indoor air in the homes of nonsmokers, is estimated to increase intake by 0.4 to 0.6 µg/(kg b.w. ·day). Estimated intake from food and drinking water is considerably less, ranging from 0.02 to 0.07 µg/(kg b.w.·day) and 0.02 to 0.06 µg/(kg b.w.·day), respectively. Total daily intake from these sources for five different age groups in the general population is estimated to range from 2.1 to 3.2 µg/(kg b.w.·day). Cigarette smoking may contribute an additional 26 to 33 µg/(kg b.w.·day) to the daily intake of benzene, while passive smoking may contribute 0.9 to 1.3 µg/(kg b.w.·day).

Table 6 Estimated Daily In take of Benzene by Canadians

Medium	Estimated Intake (g/(kg b.w.·day)
	Age (years)
	0 to 0.5a	0.5 to 4 b	5 to 11c	12 to 19d	20 to 70e
Ambient airf	1.5	1.7	2.0	1.7	1.3
Drinking Waterg	0.02	0.06	0.04	0.02	0.02
Foodh	0.07	0.06	0.05	0.03	0.02
Automobile-related activitiesi	-	-	-	0.9	0.7
Household productsj	0.5	0.05	0.6	0.5	0.4
Total Intake	2.1	2.3	2.7	3.2	2.4
Cigarette smoking
Activek	-	-	-	33.0	26.0
Passivel	1.0	1.2	1.3	1.1	0.9

1. Assumed to weigh 6 kg, breathe 2 m³ of air per day, and drink 0.1 L of water per day (Environmental Health Directorate, 1988).
2. Assumed to weigh 13 kg, breathe 5 m³ of air per day, and drink 0.8 L of water per day (Environmental Health Directorate, 1988).
3. Assumed to weigh 27 kg, breathe 12 m³ of air per day, and drink 1.1 L of water per day (Environmental Health Directorate, 1988).
4. Assumed to weigh 55 kg, breathe 21 m³ of air per day, and drink 1.1 L of water per day (Environmental Health Directorate, 1988).
5. Assumed to weigh 70 kg, breathe 20 m³ of air per day, and drink 1.5 L of water per day (Environmental Health Directorate, 1988).
6. Based on an average concentration of benzene in ambient air of 4.4 µg/m³ (Dann, 1991).
7. Based on a concentration of benzene in drinking water of 1.0 µg/L, determined from data in Otson et al. (1982) and provincial monitoring programs.
8. Estimate for adult intake of benzene in food from Holliday and Park (1989), based on the assumption that benzene in food is in equilibrium with air, an average airborne level of benzene of 4 ppb (12.8 µg/m³), and estimated partition coefficients for the components of the diet; estimates for other age groups modelled after Holliday and Park (1989). Food consumption patterns obtained from Nutrition Canada (1977).
9. Based on estimated intake of 40 µg/day while travelling in an automobile, and 10 µg/day while refuelling at a self-service gas station (Wallace, 1989).
10. Based on estimation by Holliday and Park (1989) of the contribution to indoor air concentration of benzene of 2µ g/m³ from household products, indirectly determined from the differences in reported concentrations in indoor air in the homes of non-smokers and the corresponding outdoor levels in the TEAM study (Wallace, 1989; Wallace et al., 1987) (Holliday and Park, 1989) and an average of 17 hours per day spent indoors (Environmental Health Directorate, 1988).
11. Based on estimated intake of 1800 µg/day through cigarette smoking (Wallace, 1989a).
12. Based on average additional indoor air concentration of 3 µg/m³ of benzene due to tobacco smoke in homes of smokers (Wallace, 1989).

3.3 Effects

3.3.1 Human Health

On the basis of available data, carcinogenicity is potentially the most sensitive endpoint for the assessment of "toxic" to humans for benzene under CEPA. In numerous case studies, and in the majority of epidemiological studies conducted to date, associations between leukemia and exposure to benzene in occupationally exposed populations have been observed (see Tables 3 and 4). In addition, there was a clear exposure-response relationship in the population for which exposure has been the most extensively characterized (Rinsky et al., 1987). Benzene has also been consistently clastogenic in occupationally exposed populations, inducing both structural and numerical chromosomal aberrations in human lymphocytes (Agency for Toxic Substances and Disease Registry, 1989; Occupational Safety and Health Administration, 1987). In recent studies, benzene has also been carcinogenic in two species of experimental animals, inducing a wide variety of tumours following inhalation (Table 1) and ingestion (Table 2). Available data on the mechanisms of action of benzene also indicate that induction of leukemia by this compound is biologically plausible. Benzene has been classified, therefore, in Group I ("Carcinogenic to Man") of the classification scheme developed by the Bureau of Chemical Hazards for use in the derivation of the "Guidelines for Canadian Drinking Water Quality" (Health and Welfare Canada, 1989b).

3.3.2 Environment

For aquatic biota, the leopard frog was the most sensitive organism identified in long-term tests. The reported LC₅₀ was 3.7 mg/L for continuous 9-day exposure of the embryo-larval stages.

Rainbow trout was the most sensitive aquatic species in acute tests, with a 96-hour LC₅₀ of 5.3 mg/L for juveniles.

Acute effects have been reported for terrestrial invertebrates and plants for concentrations of benzene in air greater than 10 000 mg/m³. Data on effects resulting from chronic exposure are not available.

The effect levels reported for laboratory mammals are considered to be relevant for wild mammals. The inhalation LC₅₀ for rats exposed to benzene for 7 hours was 32 500 mg/m³. The concentration observed to cause immunological changes in laboratory rats was 32 mg/m³.

Since benzene does not appreciably absorb radiation at wavelengths from 7 to 13 µm, it is not associated with global warming. Because benzene is not halogenated and is of low persistence in the environment, it is not associated with depletion of stratospheric ozone.

3.4 Conclusions

Benzene is used in Canada in a variety of applications that lead to its entry into the Canadian environment. This entry results in measurable concentrations of benzene in the various media to which humans and other organisms may be exposed.

3.4.1 Effects on the Environment (Paragraph 11(a))

The most sensitive response reported for exposure to benzene in an aquatic organism is a 9-day LC₅₀ of 3.7 mg/L for the leopard frog, the most sensitive aquatic species in chronic or subchronic studies. This value can be multiplied by a factor of 0.05 to convert the LC₅₀ to a chronic no-observed-effect concentration (NOEC) for a non-persistent, non-bioaccumulative substance and to account for differences in species sensitivity and extrapolation from laboratory to field conditions. This yields an estimated effects threshold of 185 µg/L for long-term exposure. The highest reported mean concentration of benzene in ambient freshwater in Canada is 2 µg/L; this is 1850 times lower than the LC₅₀ for the leopard frog and 93 times lower than the estimated effects threshold. Therefore, benzene is not considered to be "toxic" to freshwater organisms exposed to ambient surface water.

The most sensitive acute response reported for exposure to benzene in an aquatic organism is a 96-hour LC₅₀ of 5.3 mg/L for the rainbow trout. This value can be multiplied by a factor of 0.1 to account for differences in species sensitivity and extrapolation from laboratory to field conditions. This yields an estimated effects threshold of 530 µg/L for short-term exposure. The highest reported mean concentration of benzene in undiluted effluents is 65.3 µg/L; this is 81 times lower than the LC₅₀ for rainbow trout and 8 times lower than the estimated effects threshold for short-term exposure. Therefore, benzene is not considered to be "toxic" to freshwater organisms exposed under conditions approximating a worst-case scenario.

Acute effects have been reported for terrestrial invertebrates, plants, and laboratory mammals at benzene concentrations in air greater than 10 000 mg/m³. The highest 24-hour average concentration measured in cities is 41.9 µg/m³, which is almost 240 000 times lower than the effects threshold of 10 000 mg/m³. The concentration at which immunological changes were noted in rats under conditions of long-term exposure is 32 mg/m³; other effects, including neurological and behavioural changes, occurred at concentrations at least ten times higher. The average concentration of benzene reported in rural areas (1.2 µg/m³) is 26 667 times lower than the effects threshold of 32 mg/m³. Benzene is therefore not considered to be "toxic" to populations of wild mammals and other terrestrial biota as a result of exposure by inhalation.

Benzene is of low acute oral toxicity to mammals (LD₅₀ of 3306 mg/kg b.w. for rats). Given the ability of most organisms to metabolize or excrete benzene and benzene's low potential for bioaccumulation, wild mammals are not likely to be exposed to deleteriously high concentrations of benzene in food.

Therefore, on the basis of available data, benzene is not considered to be "toxic" as defined under Paragraph 11(a) of CEPA.

3.4.2 Effects on the Environment on which Human Life Depends (Paragraph 11(b))

Benzene will not contribute directly to global warming because of its short residence time in the troposphere and because it does not appreciably absorb radiation within the critical wavelengths (7 to 13 µm). Benzene is not expected to contribute to depletion of stratospheric ozone because of its short persistence in the atmosphere and non-halogenated nature. Benzene is not suspected of being associated with other known direct effects on the environment on which human life depends.

Therefore, on the basis of available data, benzene is not considered to be "toxic" as defined under Paragraph 11(b) of CEPA.

3.4.3 Effects on Human Life or Health (Paragraph 11(c))

Benzene has been classified in Group I ("Carcinogenic to Man") of the classification scheme developed by the Bureau of Chemical Hazards for use in the derivation of the "Guidelines for Canadian Drinking Water Quality" (Health and Welfare Canada, 1989b), based on its documented carcinogenicity in humans and experimental animals.

For compounds classified in Group I, where data permit, the estimated total daily intake or concentrations in relevant environmental media are compared to quantitative estimates of carcinogenic potency (expressed as the concentration or dose that induces a 5% increase in the incidence of, or mortality due to relevant tumors) in order to characterize risk and provide guidance for further action under the Act. Issues critical to the quantitative assessment of carcinogenic potency are discussed briefly in the following text. A more extensive discussion of these issues is presented in the Supporting Document.

It has been hypothesized that there may be a threshold for the development of leukemia in humans resulting from exposure to benzene. This is based on the supposition that leukemia results from progression of a precursor condition such as pancytopenia, for which there may be a threshold. However, available data in humans and experimental animals are insufficient to firmly support a relationship between pancytopenia or other precursor damage to bone marrow and benzene-induced leukemia. It is generally presumed, therefore, that there is an exposure-response relationship between induction of leukemia and exposure to benzene even at low levels.

The data considered most relevant to the quantification of the carcinogenic potency of benzene are those obtained in epidemiological studies in humans. There is considerable uncertainty in the extrapolation to humans of exposure-response relationships obtained in studies in animals, based on available information on the pharmacokinetics and metabolism of benzene. The toxicity of benzene is believed to be due to a metabolite or metabolites; however, though the principal routes of metabolism appear to be similar in all species studied, there are considerable differences in the contribution made by each pathway. There is also a paucity of information on the metabolism of benzene in the species of interest, i.e., humans. Moreover, there is evidence in three species of experimental animals, including primates, that the proportion of putative toxic metabolites formed decreases with increasing exposure.

The study considered most suitable for estimating the leukemogenic potency of benzene is that of Rinsky et al. (1987). In this study, the largest number of observed deaths due to leukemia was reported in an exposed population for which there was sufficient information on exposure to benzene to serve as a basis for quantitative risk assessment. In addition, benzene was the only hematotoxic solvent to which employees in this cohort were exposed in the workplace. Although the numbers of observed and expected cases of leukemia reported in the published account of this study, in which workers were followed up to 1981, were rather small (Rinsky et al., 1987), there have been additional deaths due to this cause in the most recent follow-up of a portion of this cohort (to December 1987), which has not been published (Rinsky, 1991). Moreover, there was a strongly positive trend in mortality due to leukemia with increasing cumulative exposure. In the nested case-control analysis, the average duration of exposure was longer for cases than controls (8.7 versus 2.6 years).

The type of leukemia most commonly associated with occupational exposure to benzene is acute myelogenous leukemia. However, persons with chronic myelogenous leukemia may suffer a terminal "blast crisis", with transformation to acute myelogenous leukemia. The cause of the death could subsequently be recorded as acute myelogenous leukemia (Robbins and Angell, 1971; Stewart, 1991). Though the clinical presentation of chronic myelogenous leukemia is different than that of the acute variety, owing to the occasional difficulty in distinguishing the cause of death from the two disease states on death certificates and since, to date, only two of the nine cases in the cohort studied by Rinsky et al. (1987) were chronic or unspecified myelogenous leukemias, the quantitative assessment of potency would ideally include estimates based on acute myelogenous leukemia, and acute, unspecified and chronic myelogenous leukemias combined. Since data on the background rates of chronic and unspecified leukemia were not available, however, and there is lack of convergence of the maximum likelihood estimation procedure, potency estimates based only on acute myelogenous leukemia are presented here.

Although there were four deaths in the pliofilm cohort due to this cause (versus one expected), multiple myeloma is not included as an endpoint in the quantitative assessment of potency, since it is not possible on the basis of available data to conclude unequivocally that multiple myeloma is causally related to benzene exposure.

Attempts to quantify exposure of the workers in the cohort examined by Rinsky et al. (1987) have been relatively extensive (Rinsky et al., 1987; Crump and Allen, 1984; Paustenbach et al., 1991). However, the estimates of exposure for different job categories vary considerably among the different authors. These variations are principally a result of differences in the methods used to extrapolate from existing data to fill gaps. The extent of consideration of factors, such as peak and dermal exposures, the quality of earlier monitoring data, the effect of modifications to ventilation systems, and extended work weeks during the war, has also contributed to variations in exposure estimates.

Although the exposure estimates developed by Paustenbach et al. (1991) are based on additional information which was not available to either Crump and Allen (1984) or Rinsky et al. (1987), it has not been possible to independently estimate exposure of the workers in the critical study. Estimates of cancer potency presented here are based on the exposure estimates of Crump and Allen (1984), owing to the lack of availability of sufficient data for those of Paustenbach et al. (1991) and Rinsky et al. (1987).

The age-specific death rate for acute myelogenous leukemia was assumed to be a linear-quadratic function of the total biologically effective dose, which is additive to the death rate for the general population assumed not to be exposed to benzene. The total biologically effective dose is based on the assumption that there is a lag between the time of exposure and the time of onset of acute myelogenous leukemia. This has been modelled using a gamma density function.

The increase in probability of death due to constant lifetime exposure to benzene has been determined assuming a constant exposure for a period equal to the median survival time of 75 years and that there are no competing causes of death. The concentration that corresponds to a 5% increase in mortality due to acute myelogenous leukemia (toxic dose _0.05 or TD_0.05), based on the data on mortality in the follow-up of the pliofilm cohort to 1981 (Rinsky et al., 1987), estimates of exposure developed by Crump and Allen (1984) and a linear-quadratic model for the exposure-response relationship (Thorslund and Farrar, 1992), is 14.7 x 10³ µg/m³. Based on an average concentration of benzene in ambient air (the principal source of exposure for humans) in Canadian cities of 4.4 µg/m³ (Dann, 1991), the calculated corresponding exposure/potency index for benzene is 3.0 x 10^-4. The priority for further action (i.e., analysis of options to reduce exposure) is, therefore, considered to be high.

If accessible data permit, additional estimates for the carcinogenic potency, based on the partial follow-up of the pliofilm cohort to 1987 and the estimates of exposure developed by Rinsky et al. (1987) and Paustenbach et al. (1991), will be derived and released separately at a later date.

Since on the basis of available data, benzene is classified as carcinogenic to humans (i.e., as a non-threshold toxicant - a substance for which there is considered to be some probability of harm for the critical effect at any level of exposure), it is considered to be "toxic" as defined under Paragraph 11(c) of the Canadian Environmental Protection Act.

This approach is consistent with the objective that exposure to non-threshold toxicants should be reduced wherever possible and obviates the need to establish an arbitrary de minimis level of risk for determination of "toxic" under the Act.

3.4.4 General Conclusions

On the basis of available data, benzene is not considered to be "toxic" as defined under Paragraphs 11(a) and 11(b) of CEPA. Benzene is considered to be "toxic" as defined under Paragraph 11(c) of CEPA.