Carbon Tetrachloride Risk Assessment and Liver Toxicity

Chemical-Specific Risk Assessment Methodology

Carbon Tetrachloride (CASRN 56-23-5)

RoseyFields conducts Carbon Tetrachloride risk assessments using the following methodology. This document describes the toxicological basis, dose-response derivations, and exposure-assessment procedures applied by RoseyFields risk assessors when evaluating human health risk from Carbon Tetrachloride at a given site or in a given matter, and is intended to provide a transparent, defensible, and reproducible record of the scientific judgments underlying our numerical risk outputs.

1. Chemical Identity, Regulatory Background, and Exposure Sources

1.1 Chemical Identity

Carbon tetrachloride (CCl4; CASRN 56-23-5) is a chlorinated one-carbon aliphatic compound historically produced in large volumes as a solvent, a chemical intermediate in the manufacture of chlorofluorocarbons, a grain fumigant, a dry-cleaning agent, and a fire-extinguishing fluid. It is a dense, volatile, nonflammable liquid at room temperature, moderately water-soluble, and miscible with a wide range of organic solvents. Because of its volatility and historical use as a degreasing solvent and industrial intermediate, it is a common constituent of soil, soil-gas, and groundwater contamination at former manufacturing, dry-cleaning, and disposal sites, and has been detected as a trace atmospheric contaminant and drinking-water contaminant in a number of surveys.

1.2 Regulatory Background

Carbon tetrachloride has been the subject of quantitative noncancer and cancer dose-response assessment activity since the 1980s. Early risk values relied on linear extrapolation from hepatocellular tumor data in rats, mice, and hamsters (Della Porta et al., 1961; Edwards and Dalton, 1942; Edwards et al., 1942). The dose-response basis was substantially revised following publication of a comprehensive chronic bioassay of carbon tetrachloride vapor in rats and mice (Nagano et al., 2007b), which provided multiple-dose incidence data suitable for benchmark dose (BMD) modeling, and following broader application of physiologically based pharmacokinetic (PBPK) modeling to carbon tetrachloride (Paustenbach et al., 1988; Thrall et al., 2000). RoseyFields’ methodology below reflects this updated dose-response basis, and cites directly to the underlying peer-reviewed toxicological and epidemiological literature rather than to any single regulatory compilation.

1.3 Exposure Sources and Pathways

RoseyFields evaluates the following principal exposure pathways for carbon tetrachloride, selected based on physicochemical properties and documented use history:

• Ingestion of contaminated drinking water — carbon tetrachloride has been detected in public water supplies located near industrial or waste-disposal sources, and epidemiological work has examined birth-outcome associations with such contamination (Bove et al., 1995).
• Inhalation of vapor from contaminated groundwater or soil vapor intrusion into indoor air, and inhalation of ambient or occupational vapor — the compound’s high vapor pressure and history as a workplace solvent make inhalation a dominant pathway in many occupational and vapor-intrusion settings (Tomenson et al., 1995).
• Incidental soil ingestion and dermal contact with contaminated soil or water, evaluated as secondary pathways for surface and near-surface contamination scenarios.
• Dietary or fumigant residue exposure, historically relevant given past use as a grain fumigant, though generally a minor contemporary pathway.

Selection of applicable pathways for a given matter is site- and fact-specific and is documented separately in the site-specific conceptual site model; the dose-response values below apply irrespective of pathway, subject to the route-specific derivations described in Sections 3 through 7.

2. Point-of-Departure (POD) Selection

2.1 Noncancer Oral POD

The principal study RoseyFields relies upon for the oral noncancer point of departure is the 12-week gavage study of Bruckner et al. (1986), in which adult male Sprague-Dawley rats (15–16/group) received carbon tetrachloride in corn oil by gavage at 0, 1, 10, or 33 mg/kg, 5 days/week, for 12 weeks (time-weighted average doses of 0, 0.71, 7.1, and 23.6 mg/kg-day). Bruckner et al. (1986) measured serum alanine aminotransferase (ALT), ornithine carbamoyl transferase (OCT), and sorbitol dehydrogenase (SDH) as indicators of hepatocellular injury, and blood urea nitrogen (BUN) as an indicator of renal function, at 2-week intervals, together with terminal liver and kidney histopathology.

Elevated serum SDH activity was selected as the critical effect and the basis of the POD. Travlos et al. (1996), in a cross-chemical analysis of 13-week rodent toxicity studies, identified SDH as a specific and sensitive biomarker of hepatocellular injury, more closely correlated with histopathological change than ALT or OCT across a broad chemical set. In the Bruckner et al. (1986) data, only SDH showed a statistically and biologically significant increase at the 10 mg/kg dose group, coincident with mild centrilobular hepatocyte vacuolization, while the 33 mg/kg group showed overt toxicity (up to 50-fold elevation in SDH, significant reductions in body-weight gain, and extensive degenerative liver lesions including periportal fibrosis and bile duct hyperplasia). No effects were observed at 1 mg/kg-day, establishing a NOAEL of 1 mg/kg-day and a LOAEL of 10 mg/kg-day for this endpoint.

Rather than relying on the NOAEL/LOAEL directly, RoseyFields’ assessment applies benchmark dose (BMD) modeling to the continuous SDH response data, consistent with current risk-assessment practice for continuous endpoints. Of the dose-response models fit to the 10-week SDH data set, the power model provided the best fit (goodness-of-fit p ≥ 0.1; lowest Akaike Information Criterion), using a benchmark response (BMR) defined as a doubling of SDH activity relative to control (a level considered a minimal biologically significant increase). This model yielded a BMD2X of 7.32 mg/kg-day and a 95% lower confidence bound, BMDL2X, of 5.46 mg/kg-day. Because animals were dosed only 5 of 7 days per week, this value was time-adjusted by a factor of 5/7 to derive a POD (BMDL2X-ADJ) of 3.9 mg/kg-day, representing an average daily dose. Alternative data sets considered — 10- and 12-week ALT and OCT data, rat liver histopathology from Bruckner et al. (1986), and mouse liver histopathology from Condie et al. (1986) — were either less amenable to modeling or produced higher (less protective) BMDL values, and were not selected.

2.2 Noncancer Inhalation POD

For the inhalation noncancer assessment, RoseyFields relies on the 2-year whole-body inhalation bioassay of Nagano et al. (2007b), in which groups of 50 F344 rats/sex/group were exposed to 0, 5, 25, or 125 ppm carbon tetrachloride vapor, 6 hours/day, 5 days/week, for 104 weeks, with a parallel bioassay in Crj:BDF1 mice using the same protocol. Fatty change (steatosis) in the liver of rats was selected as the critical endpoint because it was the most sensitive histopathological indicator of hepatocellular injury present at 25 ppm but absent (comparable to control) at 5 ppm, and was more sensitive than the serum-chemistry changes (elevated AST, ALT, BUN) observed at the same exposure levels. An accompanying increase in the severity of proteinuria was noted at the lowest tested concentration (5 ppm); RoseyFields, consistent with the interpretation in the underlying pathology record, does not treat this as the basis for the POD because F344 rats have a high spontaneous background incidence of age-related nephropathy that confounds interpretation of this endpoint at low exposure.

Dose-response modeling of the incidence of fatty liver change was conducted on an internal-dose basis rather than directly on external (ppm) exposure, using the PBPK-derived dose metric described in Section 3. In male rats, the log-logistic dichotomous model provided the best fit (χ² p ≥ 0.1; lowest AIC); in female rats, the multistage model provided the best fit after the highest dose group was dropped due to lack of fit across the full dose range. A benchmark response of a 10% increase in incidence (BMR = 10%) was applied. The resulting internal-dose BMCL10 values were converted, via a human PBPK model, to human equivalent concentrations (HECs); the male-rat-derived HEC of 14.3 mg/m3 was selected as the inhalation POD because it represented the more sensitive (lower) of the two sex-specific estimates.

2.3 Cancer POD

For carcinogenic risk, RoseyFields relies on tumor incidence data from the same 2-year JBRC bioassay (Nagano et al., 2007b). Two tumor data sets were used: (1) hepatocellular adenoma/carcinoma in female BDF1 mice (incidence of 4/50, 9/49, 44/50, and 48/49 at 0, 5, 25, and 125 ppm, respectively; statistically significant dose-related trend by Peto’s test, p ≤ 0.01), used as the basis for the oral slope factor via route-to-route extrapolation because no oral bioassay was suitable for dose-response modeling (see Section 2.4); and (2) pheochromocytoma in male BDF1 mice (incidence of 0/50, 0/50, 16/50, and 32/50 at the same exposure levels; statistically significant dose-related trend, p ≤ 0.01), used as the basis for the inhalation unit risk. For each tumor type, the multistage (liver) or log-probit (pheochromocytoma) dose-response model was fit to PBPK-derived internal-dose estimates, and a benchmark dose associated with 10% extra risk (BMD10) and its 95% lower bound (BMDL10, or LED10/LEC10 in dose or concentration terms after HEC conversion) were derived as the POD for low-dose linear extrapolation.

2.4 Why the Oral Cancer Bioassays Were Not Used Directly

Historical oral bioassays of carbon tetrachloride (Della Porta et al., 1961, in hamsters; Edwards and Dalton, 1942 and Edwards et al., 1942, in mice; National Cancer Institute rat and mouse gavage bioassays) were reviewed but found unsuitable for direct dose-response modeling. The hamster and early mouse studies used a single dose group, dosed animals for substantially less than a lifetime, and did not report body-weight data needed to reconstruct dose with confidence, despite producing high tumor incidences (up to 100% in Della Porta et al., 1961). The NCI mouse bioassay produced tumor incidence at or near 100% in both dose groups, precluding characterization of a dose-response slope, while the NCI rat bioassay showed high early mortality at the higher dose, and a paradoxical inverse relationship between dose and tumor incidence in females, consistent with early deaths obscuring tumor formation at the high dose. Because both liver tumors and pheochromocytomas have been demonstrated by inhalation as well as oral exposure, RoseyFields’ methodology follows the approach of extrapolating the inhalation tumor dose-response relationship to the oral route via PBPK modeling, rather than relying on these inadequate direct oral data sets.

3. Pharmacokinetic Modeling and Dose Conversion

Carbon tetrachloride’s toxicity is metabolism-dependent: hepatic cytochrome P450 2E1 (CYP2E1) converts the parent compound to a trichloromethyl free radical and, subsequently, a trichloromethyl peroxy radical, which initiates lipid peroxidation and hepatocellular injury; this bioactivation step is the rate-limiting event linking external exposure to internal toxic burden. Because metabolism — rather than external concentration — governs toxicity, RoseyFields’ methodology relies on physiologically based pharmacokinetic (PBPK) modeling, rather than simple allometric scaling, for interspecies and route-to-route dose conversion.

The rat PBPK model applied is derived from Paustenbach et al. (1988), as refined by Thrall et al. (2000) using comparative gas-uptake data across rats, mice, and hamsters, and by Benson and Springer (1999). Two internal dose metrics are used depending on target organ: (a) time-averaged rate of hepatic carbon tetrachloride metabolism per kilogram liver (MRAMKL, µmol/hour/kg liver), used for liver effects (both noncancer fatty change and hepatocellular tumors), consistent with the CYP2E1-mediated mode of action; and (b) time-averaged arterial blood concentration (MCA, µmol/L), used for the adrenal gland/pheochromocytoma endpoint, reflecting a distinct dose metric appropriate to a non-hepatic target tissue.

External exposure concentrations from the rat and mouse bioassays (Nagano et al., 2007a, b) are first converted to internal dose metrics using the rodent PBPK model; benchmark dose modeling is performed on the internal-dose/response relationship; and the resulting internal-dose BMDL or BMCL values are then converted to human equivalent doses or concentrations (HEDs/HECs) using an analogous human PBPK model, holding the internal dose metric constant across species. This internal-dose equivalence approach is the mechanism by which RoseyFields’ methodology accomplishes route-to-route extrapolation (inhalation bioassay data informing an oral slope factor): a continuous-infusion approximation is applied in which carbon tetrachloride absorbed via the gastrointestinal tract is assumed to reach the liver at a comparable internal dose to inhaled vapor, with gastrointestinal absorption assumed to be complete. RoseyFields recognizes that to the extent oral bioavailability is less than 100%, this approach will tend to overestimate the resulting oral slope factor, which is a conservative (health-protective) direction for a cancer risk value.

For the two mouse liver-tumor and pheochromocytoma data sets, two independent published PBPK parameterizations (the “Fisher” and “Thrall” model variants, both built on the same general modeling framework of Paustenbach et al., 1988 and Thrall et al., 2000) were each used to generate an HED or HEC estimate; RoseyFields’ methodology follows the convention of averaging the two resulting values (e.g., HED of 1.40 and 1.69 mg/kg-day, averaged to 1.54 mg/kg-day, for the female mouse liver tumor data) to reduce dependence on any single model parameterization.

4. Uncertainty Factor Derivation

RoseyFields applies the standard multiplicative uncertainty factor (UF) framework to noncancer PODs, selecting each component factor based on the chemical-specific data available for carbon tetrachloride rather than defaulting uniformly to 10-fold values.

4.1 Oral RfD Uncertainty Factors (Total UF = 1,000)

Composite Oral UF = 10 (UFH) × 10 (UFA) × 3 (UFS) × 3 (UFD) = 1,000 (with UFS and UFD each rounded from 100.5 = 3.16 to the conventional value of 3).

4.2 Inhalation RfC Uncertainty Factors (Total UF = 100)

Composite Inhalation UF = 10 (UFH) × 3 (UFA) × 3 (UFD) = 100.

5. Derivation of the Oral Reference Dose and Reference Concentration

5.1 Oral RfD

RfD = POD ÷ Composite UF = 3.9 mg/kg-day ÷ 1,000 = 0.004 mg/kg-day.

This value is intended to represent a daily oral dose, including to sensitive subpopulations, that is not expected to be associated with an appreciable risk of deleterious noncancer effects (principally, hepatocellular injury) over a lifetime of exposure.

5.2 Inhalation RfC

RfC = POD ÷ Composite UF = 14.3 mg/m3 ÷ 100 = 0.1 mg/m3.

This value is intended to represent a continuous inhalation exposure concentration, including to sensitive subpopulations, that is not expected to be associated with an appreciable risk of deleterious effects (principally, hepatic fatty change) over a lifetime of exposure.

6. Carcinogenicity Classification and Cancer Potency Derivation

6.1 Weight-of-Evidence Classification

RoseyFields classifies carbon tetrachloride as likely to be carcinogenic to humans by all routes of exposure. This classification rests on (a) consistent induction of hepatocellular tumors across three rodent species (rat, mouse, hamster) and multiple routes of administration — oral gavage (Edwards, 1941; Edwards and Dalton, 1942; Edwards et al., 1942), inhalation (Nagano et al., 2007b), and parenteral injection (Reuber and Glover, 1967a, b, 1970; Kiplinger and Kensler, 1963) — together with induction of adrenal pheochromocytomas in mice by both oral and inhalation exposure; and (b) inadequate (confounded, poorly exposure-characterized) human epidemiological evidence, which has examined associations with non-Hodgkin lymphoma (Blair et al., 1998; Spirtas et al., 1991), lymphohematopoietic cancers (Checkoway et al., 1984; Wilcosky et al., 1984), and several other cancer sites, without identifying a clear association with liver cancer, the principal site of concern from animal data. Because tumors occur at sites remote from the portal of entry and the chemical is well absorbed by all exposure routes, RoseyFields’ methodology, consistent with current cancer risk assessment practice, applies the carcinogenicity finding across routes rather than limiting it to the route of the underlying bioassay.

The hypothesized mode of action for liver tumor formation — CYP2E1-mediated bioactivation, free-radical-driven hepatocellular cytotoxicity, and sustained compensatory regeneration/hyperplasia — is well supported at higher, hepatotoxic exposures. However, RoseyFields’ methodology does not apply a nonlinear (threshold) extrapolation approach for liver tumors, because the mode of action at low, non-cytotoxic exposures remains uncertain: notably, a statistically significant increase in liver adenoma incidence was observed in female mice at the lowest tested concentration (5 ppm), a concentration that did not produce hepatotoxicity in the same study (Nagano et al., 2007b), which is inconsistent with a purely cytotoxicity-driven, threshold mode of action at that dose. Genotoxicity data (reviewed comprehensively across bacterial, yeast, and mammalian test systems) are largely negative for direct mutagenicity, but positive findings that do occur are concentrated in assay systems sensitive to oxidative damage, consistent with an indirect, oxidative-stress-mediated genotoxic contribution that cannot be ruled out at low doses given sparse data in that range. For pheochromocytomas, no mode-of-action data are available at all, further supporting a linear low-dose extrapolation default for that tumor type.

6.2 Oral Slope Factor

Using the female mouse hepatocellular adenoma/carcinoma dose-response data (Nagano et al., 2007b), PBPK-based route-to-route extrapolation (Section 3) yielded an average human equivalent dose at the 10% extra risk level (average of the Fisher- and Thrall-model estimates) of 1.54 mg/kg-day (LED10); the corresponding central estimate (ED10) was 2.27 mg/kg-day.

Oral Slope Factor = 0.10 ÷ 1.54 mg/kg-day = 7 × 10⁻² per mg/kg-day (upper-bound estimate, from the LED10).

(For reference, the central-estimate slope from the ED10 is 0.10 ÷ 2.27 = 4 × 10⁻² per mg/kg-day; RoseyFields’ methodology applies the upper-bound value, 7 × 10⁻² per mg/kg-day, consistent with standard practice of using the LED10-based slope as the risk-protective estimate.) This slope factor should not be applied to exposures above the POD (1.54 mg/kg-day), because the fitted multistage model, rather than linear extrapolation, is a more appropriate characterization of risk above that dose.

Corresponding drinking water unit risk (assuming 2 L/day consumption by a 70-kg adult) is 2 × 10⁻⁶ per µg/L, yielding risk-specific concentrations of 50 µg/L (10⁻⁴ risk), 5 µg/L (10⁻⁵ risk), and 0.5 µg/L (10⁻⁶ risk).

6.3 Inhalation Unit Risk

Using the male mouse pheochromocytoma dose-response data (Nagano et al., 2007b) — the data set yielding the highest (most sensitive) potency estimate among the tumor data sets evaluated — PBPK-based interspecies extrapolation yielded an average human equivalent concentration at the 10% extra risk level of 1.78 × 10⁴ µg/m3 (LEC10); the central estimate (EC10) was 3.13 × 10⁴ µg/m3.

Inhalation Unit Risk = 0.10 ÷ 1.78 × 10⁴ µg/m3 = 6 × 10⁻⁶ per µg/m3.

This value should not be applied above the POD of approximately 1.8 × 10⁴ µg/m3. Risk-specific air concentrations are 17 µg/m3 (10⁻⁴ risk), 1.7 µg/m3 (10⁻⁵ risk), and 0.17 µg/m3 (10⁻⁶ risk).

RoseyFields notes, for case-specific application, that the oral slope factor and inhalation unit risk are derived from different tumor endpoints (liver tumors vs. pheochromocytoma) and different internal dose metrics, and are therefore not combined or summed; risk assessors should apply the route-appropriate value to each exposure pathway rather than attempting to reconcile the two into a single “potency” figure.

7. Application in RoseyFields Risk Assessment Practice

7.1 Hazard Quotient and Hazard Index

For noncancer endpoints, RoseyFields calculates a chronic daily intake (CDI) or exposure concentration for the pathway and population of interest and expresses risk as a Hazard Quotient (HQ):

HQoral = CDI (mg/kg-day) ÷ RfD (0.004 mg/kg-day)

HQinhalation = EC (mg/m3) ÷ RfC (0.1 mg/m3)

Where a receptor is exposed to carbon tetrachloride via multiple pathways, or to carbon tetrachloride together with other chemicals acting on the liver by a common mode of action, RoseyFields sums route- and chemical-specific HQs into a Hazard Index (HI) for that target organ; an HI or HQ exceeding 1.0 indicates that the exposure estimate exceeds the level associated with negligible risk of the critical (hepatic) effect and warrants further evaluation, though RoseyFields’ methodology treats this as a screening-level, not a bright-line causation, determination.

7.2 Excess Lifetime Cancer Risk

For carcinogenic risk, RoseyFields calculates an excess lifetime cancer risk (ELCR) as the product of the route-specific lifetime average daily dose (LADD) or exposure concentration and the corresponding slope factor or unit risk:

ELCRoral = LADD (mg/kg-day) × Oral Slope Factor (7 × 10⁻² per mg/kg-day)

ELCRinhalation = LEC (µg/m3) × Inhalation Unit Risk (6 × 10⁻⁶ per µg/m3)

Consistent with EPA risk-management benchmarks widely used in regulatory and litigation contexts, RoseyFields characterizes ELCR values in the 10⁻⁶ to 10⁻⁴ range as falling within the range typically considered for risk-management decision-making, while noting that this range reflects policy convention rather than a toxicological bright line, and that both the slope factor and unit risk used here are upper-bound (not central-tendency) potency estimates.

7.3 Exposure and Dose Reconstruction

Where litigation or regulatory work requires reconstruction of historical exposure — for example, occupational exposure reconstruction using industrial-hygiene monitoring data, or reconstruction of a historical drinking-water or vapor-intrusion exposure — RoseyFields’ methodology uses the PBPK models described in Section 3 to translate reconstructed external exposure concentrations into internal dose metrics comparable to those underlying the POD, which is particularly important given that carbon tetrachloride’s toxicity is governed by metabolic activation rather than parent-compound concentration alone. This allows RoseyFields to evaluate whether a case-specific exposure scenario (which may involve intermittent, variable, or short-duration exposure quite different from the continuous chronic exposure protocols used in the underlying bioassays) is toxicologically comparable to the exposure conditions underlying the RfD, RfC, slope factor, or unit risk, and to make an appropriately caveated, quantitative statement about excess risk or hazard for that specific exposure history.

7.4 Support for Litigation and Regulatory Work

RoseyFields’ risk assessors apply the methodology above to support: (a) preparation of quantitative risk assessment reports for regulatory submission (e.g., in the context of RCRA corrective action, CERCLA remedial investigation, or state voluntary cleanup programs); (b) expert opinion and testimony regarding general and specific causation in toxic tort matters, where the HQ/HI and ELCR outputs described above are one component (together with epidemiological and clinical evidence specific to the plaintiff or cohort) of a full causation analysis; (c) evaluation of remediation cleanup goals and risk-based screening levels back-calculated from the RfD, RfC, slope factor, and unit risk at a specified target HQ or risk level; and (d) review and critique of opposing experts’ dose-response or exposure-reconstruction methodology, using the chemical-specific data package summarized in this document as the point of comparison.

8. Key Uncertainties and Sensitive Populations

• Mode-of-action uncertainty at low dose: the cytotoxicity-driven mode of action for liver tumors is best supported at higher, hepatotoxic exposures; data at low, non-cytotoxic exposures are sparse, and the female mouse liver-adenoma finding at a non-hepatotoxic concentration (Nagano et al., 2007b) introduces uncertainty into any nonlinear (threshold) risk characterization.
• Human relevance of the pheochromocytoma endpoint: pheochromocytomas were observed in only one species (mouse); while morphological and immunohistochemical parallels to human pheochromocytoma support human relevance (Tischler et al., 1996; Eisenhofer et al., 2004), site concordance between animal and human tumors cannot be assumed as a general matter (National Research Council, 1994, as discussed in the underlying assessments), and human pheochromocytoma etiology is substantially influenced by hereditary factors not modeled in the rodent bioassay.
• Interindividual variability in CYP2E1 activity: because toxicity is metabolism-dependent, individuals or populations with elevated CYP2E1 activity (e.g., due to chronic ethanol consumption, certain pharmaceutical co-exposures, or genetic polymorphism) may be more susceptible to hepatotoxic and carcinogenic effects than assumed by the default intrahuman uncertainty factor; conversely, young children have lower CYP2E1 activity than adults and do not reach adult metabolic capacity until sometime between 1 and 10 years of age, which the available evidence suggests would not increase — and may decrease — susceptibility to this particular chemical’s free-radical-mediated liver injury mechanism relative to adults, in contrast to chemicals whose toxicity depends on the parent compound rather than a reactive metabolite.
• Reproductive and developmental toxicity database gaps: no adequate multigeneration reproductive toxicity study exists by any route; available developmental data (Narotsky et al., 1995, 1997a, b; Wilson, 1954; Hamlin et al., 1993) indicate that developmental effects (principally full-litter resorption) occur only at maternally toxic doses well above the hepatotoxic POD, but RoseyFields flags this as a residual database uncertainty, reflected in the applied database uncertainty factor, particularly relevant to case-specific evaluation of pregnant populations.
• PBPK model uncertainty: both interspecies extrapolation and route-to-route (inhalation-to-oral) extrapolation rely on PBPK models that, while an improvement over default allometric scaling, carry their own structural and parameter uncertainty; RoseyFields’ use of two independent model parameterizations (“Fisher” and “Thrall” variants) and averaging of their outputs is intended to partially address, but does not eliminate, this source of uncertainty.
• Occupational and other real-world co-exposure: human epidemiological cohorts historically exposed to carbon tetrachloride were typically co-exposed to other solvents (e.g., trichloroethylene, tetrachloroethylene) or to confounders such as smoking, limiting the ability of human data to independently confirm the potency estimates derived from animal bioassays (Blair et al., 1998; Checkoway et al., 1984); case-specific risk assessment should account for co-exposure to other hepatotoxicants or carcinogens present at a given site.
• Sensitive populations warranting particular attention in case-specific work include: individuals with pre-existing liver disease or heavy alcohol use (altered CYP2E1 induction and reduced hepatic reserve capacity); pregnant individuals (developmental toxicity database limitations, noted above); and individuals with concurrent exposure to other CYP2E1 substrates or inducers/inhibitors, which may alter internal dose at a given external exposure level relative to the bioassay animals.

9. Bibliography

(Author-year citations above correspond to the following peer-reviewed sources, listed alphabetically.)

Benson, JM; Springer, DL. (1999) Improved risk estimates for carbon tetrachloride. Report DE-FC04-96AL76406, Project No. 54940.

Blair, A; Hartge, P; Stewart, PA; et al. (1998) Mortality and cancer incidence of aircraft maintenance workers exposed to trichloroethylene and other organic solvents and chemicals: extended follow-up. Occup Environ Med 55:161–171.

Bove, FJ; Fulcomer, MC; Klotz, JB; et al. (1995) Public drinking water contamination and birth outcomes. Am J Epidemiol 141:850–862.

Bruckner, JV; MacKenzie, WF; Muralidhara, S; et al. (1986) Oral toxicity of carbon tetrachloride: acute, subacute and subchronic studies in rats. Fundam Appl Toxicol 6:16–34.

Checkoway, H; Wilcosky, T; Wolf, P; et al. (1984) An evaluation of the associations of leukemia and rubber industry solvent exposures. Am J Ind Med 5:239–249.

Condie, LW; Laurie, RD; Mills, T; et al. (1986) Effect of gavage vehicle on hepatotoxicity of carbon tetrachloride in CD-1 mice: corn oil versus Tween-60 aqueous emulsion. Fundam Appl Toxicol 7:199–206.

Della Porta, GD; Terracini, B; Shubik, P. (1961) Induction with carbon tetrachloride of liver cell carcinomas in hamsters. J Natl Cancer Inst 26:855–863.

Edwards, JE. (1941) Hepatomas in mice induced with carbon tetrachloride. J Natl Cancer Inst 2:197–199.

Edwards, JE; Dalton, AJ. (1942) Induction of cirrhosis of the liver and of hepatomas in mice with carbon tetrachloride. J Natl Cancer Inst 3:19–41.

Edwards, J; Heston, WE; Dalton, AJ. (1942) Induction of the carbon tetrachloride hepatoma in strain L mice. J Natl Cancer Inst 3:297–301.

Eisenhofer, G; Bornstein, SR; Brouwers, FM; et al. (2004) Malignant pheochromocytoma: current status and initiatives for future progress. Endocr Relat Cancer 11:423–436.

Hamlin, GP; Kholkute, SD; Dukelow, WR. (1993) Toxicology of maternally ingested carbon tetrachloride (CCl4) on embryonal and fetal development and in vitro fertilization in mice. Zool Sci 10:111–116.

Kiplinger, GF; Kensler, CJ. (1963) Failure of phenoxybenzamine to prevent formation of hepatomas after chronic carbon tetrachloride administration. J Natl Cancer Inst 30:837–843.

Nagano, K; Umeda, Y; Saito, M; et al. (2007a) Thirteen-week inhalation toxicity of carbon tetrachloride in rats and mice. J Occup Health 49:249–259.

Nagano, K; Sasaki, T; Umeda, Y; et al. (2007b) Inhalation carcinogenicity and chronic toxicity of carbon tetrachloride in rats and mice. Inhal Toxicol 19:1089–1103.

Narotsky, MG; Kavlock, RJ. (1995) A multidisciplinary approach to toxicological screening: II. Developmental toxicity. J Toxicol Environ Health 45:145–171.

Narotsky, MG; Brownie, CF; Kavlock, RJ; et al. (1997a) Critical period of carbon tetrachloride-induced pregnancy loss in Fischer 344 rats, with insights into the detection of resorption sites by ammonium sulfide staining. Teratology 56:252–261.

Narotsky, MG; Pegram, RA; Kavlock, RJ. (1997b) Effect of dosing vehicle on the developmental toxicity of bromodichloromethane and carbon tetrachloride in rats. Fundam Appl Toxicol 40:30–36.

Paustenbach, DJ; Clewell, HJ, III; Gargas, ML; et al. (1988) A physiologically based pharmacokinetic model for inhaled carbon tetrachloride. Toxicol Appl Pharmacol 96:191–211.

Reuber, MD; Glover, EL. (1967a) Cholangiofibrosis in the liver of buffalo strain rats injected with carbon tetrachloride. Br J Exp Pathol 48:319–322.

Reuber, MD; Glover, EL. (1967b) Hyperplastic and early neoplastic lesions of the liver in buffalo strain rats of various ages given subcutaneous carbon tetrachloride. J Natl Cancer Inst 38:891–899.

Reuber, MD; Glover, EL. (1970) Cirrhosis and carcinoma of the liver in male rats given subcutaneous carbon tetrachloride. J Natl Cancer Inst 44:419–427.

Spirtas, R; Stewart, PA; Lee, JS; et al. (1991) Retrospective cohort mortality study of workers at an aircraft maintenance facility. I. Epidemiological results. Br J Ind Med 48:515–530.

Thrall, KD; Vucelick, ME; Gies, RA; et al. (2000) Comparative metabolism of carbon tetrachloride in rats, mice, and hamsters using gas uptake and PBPK modeling. J Toxicol Environ Health A 60:531–548.

Tischler, AS; Sheldon, W; Gray, R. (1996) Immunohistochemical and morphological characterization of spontaneously occurring pheochromocytomas in the aging mouse. Vet Pathol 33:512–520.

Tomenson, JA; Baron, CE; O’Sullivan, JJ; et al. (1995) Hepatic function in workers occupationally exposed to carbon tetrachloride. Occup Environ Med 52:508–514.

Travlos, GS; Morris, RW; Elwell, MR; et al. (1996) Frequency and relationships of clinical chemistry and liver and kidney histopathology findings in 13-week toxicity studies in rats. Toxicology 107:17–29.

Wilcosky, TC; Checkoway, H; Marshall, EG; et al. (1984) Cancer mortality and solvent exposures in the rubber industry. Am Ind Hyg Assoc J 45:809–811.

Wilson, JG. (1954) Influence on the offspring of altered physiologic states during pregnancy in the rat. Ann NY Acad Sci 57:517–525.

This methodology document is prepared by RoseyFields for internal and client reference and reflects the peer-reviewed toxicological and epidemiological literature on carbon tetrachloride as summarized above. Application to any specific site, exposure scenario, or matter requires additional site-specific exposure assessment and is documented separately.