The former OSHA PEL for chloroform was 50 ppm as a ceiling limit. OSHA proposed to revise this limit to 2 ppm, measured over a 60-minute period. This limit was based on the NIOSH (1977p, as cited in ACGIH 1986/Ex. 1-3, p. 130) REL, and NIOSH (Ex. 8-47, Table N6A) has indicated its concurrence with the proposed limit. In the final rule, OSHA is establishing a 2-ppm limit for chloroform, but is expressing this PEL as an 8-hour TWA limit. The ACGIH has established a TLV-TWA of 10 ppm and assigned chloroform an A2 designation. Chloroform is a clear, colorless, nonflammable, volatile liquid with a pleasant odor.
Chloroform is considered by the ACGIH, the United States Environmental Protection Agency (EPA), and the International Agency for Research on Cancer (IARC) as a probable carcinogen in humans. Chloroform is given an overall weight-of-evidence classification of B2 by the EPA and a classification of 2B by IARC. These classifications are based on these organizations' determination that there is sufficient evidence for the carcinogenicity of chloroform in animals and insufficient evidence in humans. The following discussion is based on information from the EPA Health Assessment Document for Chloroform (EPA 1984f/Ex.1-216).
It is currently believed that the carcinogenicity of chloroform results from the formation of reactive metabolites, such as phosgene, that bind to cellular macromolecules. Although there is some evidence to suggest that chloroform is weakly mutagenic, the results of most mutagenicity tests are negative.
In humans, there are no epidemiological studies that evaluate populations exposed only to chloroform, although there are several studies that examine populations exposed to chloroform in chlorinated drinking water. However, because chloroform is not the only potential carcinogen present in chlorinated water, the epidemiological data are considered inadequate to use as the basis for a quantitative risk assessment. Thus, a causal relationship between cancer and chloroform exposure cannot be determined based on epidemiological studies alone, although these studies can be used to provide general support for findings in animal studies.
A case-controlled study indicated a significant association between colon cancer and exposure to chlorinated drinking water contaminated with organic material (Young, Kanarek, and Tsiatis 1981/Ex. 1-118). Significant positive associations were also found for chloroform levels in drinking water and the incidence of mortality due to cancer of the bladder, rectum, and large intestine (Hogan, Chi, Hoel, and Mitchell 1979/Ex. 1-159). Similar results also have been found by others (Cantor, Hoover, Mason, and McCabe 1978/Ex. 1-50; and Gottlieb, Carr, and Morris 1981/Ex. 1-72). However, although these studies suggest an association between exposure to chloroform and an increased risk of cancer, a definite causal relationship between the development of colon and bladder cancer and exposure to chloroform cannot be determined solely from these studies.
In animals, several long-term studies provide strong evidence for the carcinogenic activity of chloroform. Chloroform has been shown to produce statistically significant increases in renal epithelial tumors in male rats and hepatocellular carcinomas in several strains of mice. The carcinogenic activity of chloroform in these studies is specific to the kidney and liver.
The carcinogenic activity of chloroform was investigated in rats exposed to chloroform by gavage for 78 weeks (NCI 1976a/Ex. 1-119). Male rats were administered doses of 90 or 180 mg/kg/day, and female rats were administered doses of 100 or 200 mg/kg/day. A statistically significant dose-related increase in renal epithelial tumors was observed in treated male rats compared with untreated, matched controls; these tumors were described as carcinomas and adenomas. No increase in the incidence of tumors was observed in chloroform-treated female rats.
In this same study, the carcinogenicity of chloroform was evaluated in mice exposed chronically to chloroform by gavage (NCI 1976a/Ex. 1-119). Male mice were exposed to doses of 138 or 277 mg/kg/day and females to 238 or 477 mg/kg/day for 78 weeks. There were significant dose-related increases in the incidence of hepatocellular carcinomas in chloroform-treated male and female mice. The increase of tumors in male mice for low and high doses was 36 percent and 98 percent, respectively. For female mice, the increases were 80 percent for the low dose and 95 percent for the high dose of chloroform.
The carcinogenic potential of chloroform in mice was further investigated in two additional studies (Roe, Palmer, and Worden 1979/Ex. 1-108; Jorgenson, Meierhenry, Rushbrook et al. 1985/Ex. 1-117). Doses of 17, 60, or 100 mg/kg/day were administered to four different strains of male and female mice [C57BL, CBA, CF/1, and ICI) by gavage for 80 weeks (Roe, Palmer, and Worden 1979/Ex. 1-108). The incidence of kidney tumors, described as hypernephromas, was significantly elevated in the ICI strain. Moderate to severe renal changes were observed in the male mice of the other strains, but no significant increase in renal tumors was reported. Tumors were not observed in female mice.
The carcinogenicity of chloroform administered in drinking water was investigated in male rats and female mice (Jorgenson, Meierhenry, Rushbrook et al. 1985/Ex. 1-117). Animals were treated with drinking water containing chloroform concentrations of 200, 400, 900, or 1800 mg/L for 104 weeks. There was a marked increase in the number of kidney tumors (described as tubular cell adenomas and adenocarcinomas) in rats. However, the incidence of tumors in female mice was not significantly increased.
Risk estimate for chloroform. The Jorgenson et al. (1985/Ex. 1-117) rat study, which demonstrated a statistically significant increase in the incidence of renal tumors in male rats, was the data set used for the quantitative risk estimation. (In the NPRM, OSHA inadvertently identified the NCI (1976a/Ex. 1-119) study as forming the basis for risk assessment; see Ex. 110.) Although there are no data concerning the carcinogenicity of chloroform following inhalation exposure, the risk from inhaled chloroform is considered to be equivalent to the risk from ingested chloroform. The linearized multistage, one hit, and Weibull models were used. The maximum likelihood estimates of excess cancers over an occupational lifetime for a population of 1,000 and the 95-percent upper-bound estimates are summarized in Table C15-6. The Weibull model is similar to the logit and probit models. However, by using only one data set, the logit, probit, and multihit models failed to converge.
The results of the data analysis presented here are similar to the results of other models described by the EPA (1984f/Ex. 1-216) for chloroform. These three models clearly demonstrate, based on the MLE estimates, that a significant cancer risk exists at the former PEL of 50 ppm. The risks estimated to exist at the former PEL are of the same order of magnitude as the risks determined by OSHA to be associated with other carcinogens that OSHA has regulated (e.g., benzene, ethylene oxide). Some commenters (Exs. 3-685, 3-741, 3-958, 8-89, and L3-1262) stated that OSHA's risk assessment approach for chloroform overstated the risk by not accounting for certain aspects of the mechanism by which chloroform induces cancer. Dow Chemical (Ex. 3-741, p. 45), Hoffmann-LaRoche (Ex. L3-1262), and the American Paper Institute (Ex. 3-685) presented evidence that the mouse liver tumors resulting from chloroform exposure arise secondarily to organ toxicity, which is a threshold phenomenon. As such, they argue that the use of a linearized, no-threshold model will overstate cancer risk. Theodore J. Berger, Assistant Vice President and Director of Corporate Environmental and Safety Affairs at Hoffmann-LaRoche, pointed out that the ACGIH TLV of 10 ppm for chloroform was based on this consideration, and that the 10-ppm level was one-fifth the level at which organ injury has been observed (in rats).
On the issue of the carcinogenic mechanism of chloroform, rulemaking participants (Exs. 3-685, 3-741, and L3-1262) and the EPA (1984f/Ex. 1-216) cite evidence that suggest that increased cell death brought about by the formation of reactive metabolites may be one mechanism by which chloroform has caused cancer in animals, particularly in the liver. The EPA (1984f/Ex. 1-216) also cites evidence that chloroform metabolites may deplete glutathione, which results in less effective cellular detoxification. In addition, a genotoxic mechanism cannot be entirely ruled out, although the data are equivocal; chloroform has produced positive results in the micronucleus test and host-mediated mutagenicity assay in the mouse; mutations in yeast; abnormal sperm morphology in mice; and sister chromatid the evidence presented by EPA (1984f/Ex. 1-216) and the comments on EPA's document by Dr. Bull, OSHA concludes that its use of the Jorgenson et al. (1985/Ex. 1-117) rat kidney data and multistage model is a reasonable approach for estimating the risk of cancer associated with exposure to chloroform. Furthermore, OSHA concludes that, even if one were to accept both that chloroform increases cancer risk via a cell-death mechanism, and that a threshold dose for this effect exists, the 10-ppm TLV recommended by the ACGIH provides an inadequate margin (fivefold) of protection against this life-threatening disease.
Dow Chemical Company (Ex. 3-741) and the American Paper Institute (Ex. 3-685) also commented that, because humans metabolize chloroform to a lesser degree than do rodents, quantitative risk assessments should consider such differences. Dow submitted a discussion (Ex. 3-958) of the preliminary results of an assessment based on the use of a physiologically based pharmacokinetic model (PB-PK) similar to that developed by Andersen et al. (1987) for methylene chloride. In this assessment, the researchers reported that the estimated cancer risk for chloroform was one to two orders of magnitude lower than the risks estimated using the multistage model. However, since this work is currently underway, details of the assessment are not available.
Dow Chemical also applied EPA's (1984f/Ex. 1-216) approach to the rat data from the Jorgenson et al. (1985/Ex. 1-117) study (Ex. 3-741, pp. 45-47). This approach uses metabolic data to express the active dose in units of average mg metabolite produced per day per liter of tissue; this method contrasts with OSHA's approach of using applied dose for the risk assessment. Dow's MLE estimate of lifetime occupational cancer risk associated with exposure to 2 ppm is 0.17 deaths per 1,000 workers (upper-confidence limit of 0.46/1,000), based on the amount of chloroform metabolized per unit volume of kidney tissue. The estimate based on chloroform metabolism in the liver is 0.27/1,000 (upper-confidence limit of 0.74/1,000). OSHA does not believe that these estimates, which account for interspecies differences in chloroform metabolism, are substantially different from OSHA's estimates, which are based on the use of applied dose; Dow's MLE estimate based on metabolism in the kidney is not quite half of OSHA's MLE estimate, and Dow's MLE estimate based on liver metabolism is the same as OSHA's. These findings give OSHA greater confidence in the estimates of chloroform-related cancer risk presented in Table C15-6 above.
The AFL-CIO (Ex. 194) supported OSHA's proposed PEL for chloroform. However, the New Jersey Department of Public Health (Ex. 144) urged OSHA to set a limit for chloroform based on EPA's IRIS data. The use of such an approach for setting exposure limits is discussed in Section VI.A of the preamble.
Based on the evidence presented above, OSHA concludes that a significant risk of cancer, which OSHA considers a material impairment of health and functional capacity, exists at the former PEL of 50 ppm, with estimated risks ranging from 22 to 34 excess deaths per 1,000 workers. The Supreme Court indicates that a reasonable person "might well consider a risk of 1.0 per 1,000 significant, and take steps to decrease or eliminate that risk" (I.U.D. v. A.P.I., 448 U.S. 655) (see the discussion in Section VI.A of this preamble). Based on OSHA's risk assessment, significant risk of cancer remains at the ACGIH TLV of 10 ppm (1.6 deaths per 1,000 workers). OSHA also finds that revising the PEL to 2 ppm will substantially reduce this risk by from 96 to 99 percent. Therefore, OSHA is establishing a 2-ppm limit as the PEL for chloroform. However, in the final rule, OSHA is establishing this limit as an 8-hour TWA limit, rather than a 60-minute limit as proposed. OSHA has determined that a TWA limit is more appropriate for chloroform since low-level exposure to chloroform presents a chronic, rather than acute, health hazard. OSHA also believes that establishing a TWA limit will simplify the development of compliance and exposure-monitoring strategies for employers, since an 8-hour TWA limit is more conventional than a 60-minute limit. Therefore, in the final rule, OSHA is establishing a 2-ppm 8-hour TWA PEL for chloroform.