TCE Toxicology and Cancer Risk

Primary source document informing this summary: U.S. EPA, Risk Evaluation for Trichloroethylene, CASRN 79-01-6, EPA Document #740R18008 (November 2020).

1. Methodology Overview

RoseyFields conducts trichloroethylene (TCE) risk assessments using the following methodology:

Dr. Rosenfeld characterizes hazard by synthesizing the peer-reviewed animal toxicology and human epidemiology literature on TCE; we identify points of departure (PODs) for the most sensitive, well-supported non-cancer endpoints across organ systems and apply standard uncertainty-factor methodology to derive health-protective reference values; we evaluate carcinogenic potential using the weight of the epidemiological and mechanistic evidence and apply a linear low-dose slope factor; and we combine these hazard values with site- or scenario-specific exposure estimates to calculate hazard quotients, hazard indices, and excess lifetime cancer risk for the populations of concern. The summary below documents the toxicokinetic, toxicological, and epidemiological evidence base for TCE and explains how RoseyFields applies the resulting reference values in practice.

2. Chemical Identity and Background

Trichloroethylene (TCE, CASRN 79-01-6, molecular formula C2HCl3, molecular weight 131.39 g/mol) is a colorless, volatile liquid with a sweet odor resembling chloroform (O’Neil et al., 2006). It has a boiling point of 87.2°C, vapor pressure of approximately 73.7 mmHg at 25°C, water solubility of 1,280 mg/L at 25°C, and a log octanol-water partition coefficient (log Kow) of 2.42 (Lide, 2007). These properties make TCE volatile and moderately lipophilic, driving both its behavior in the environment and its disposition in biological tissues.

TCE has historically been used as a metal degreasing solvent, an extraction solvent, an intermediate in the manufacture of fluorinated refrigerants, and a component of some consumer products (e.g., aerosol degreasers, spot removers, and adhesives). Occupational exposure occurs primarily via inhalation and dermal contact during vapor degreasing, cold cleaning, and related metalworking operations; environmental exposure occurs through contaminated groundwater and drinking water, indoor air vapor intrusion, and ambient air near industrial sources.

3. Toxicokinetics (ADME)

Absorption

TCE is lipophilic and readily crosses biological membranes, partitioning into blood largely through binding with lipid components (Cichocki et al., 2016). It is efficiently absorbed by all three exposure routes: essentially complete following inhalation, rapid from the gastrointestinal tract following ingestion, and rapid across intact skin following dermal contact with either liquid or vapor. In human volunteer studies, peak TCE levels in expired air occurred within 15 minutes of liquid contact and 30 minutes of vapor contact, and dermal exposure disrupts the stratum corneum in a manner that can enhance its own absorption over time.

Distribution

Once absorbed, TCE distributes widely throughout the body and preferentially partitions into lipid-rich tissues, including brain, muscle, heart, kidney, lung, liver, and adipose tissue (Cichocki et al., 2016). TCE crosses the placenta and has been detected in maternal and fetal blood, and it is excreted into breast milk; one study reported detectable TCE in all eight sampled breast-milk specimens from mothers in urban areas (Pellizzari et al., 1982), while a separate study detected TCE in 7 of 20 samples at concentrations up to 6 ng/mL (Beamer et al., 2012).

Metabolism

TCE is extensively metabolized via two competing hepatic pathways: (1) oxidation through the cytochrome P450 (CYP) mixed-function oxidase system, principally CYP2E1, yielding chloral, trichloroethanol, and trichloroacetic acid, among other metabolites; and (2) glutathione (GSH) conjugation, yielding S-(1,2-dichlorovinyl)glutathione (DCVG) and its downstream cysteine conjugate DCVC. Oxidative metabolism is the dominant pathway, and production of conjugative metabolites increases in CYP2E1-null mice, indicating redundancy among CYP isoforms in the oxidative pathway (Luo et al., 2018). GSH conjugation, although a minor pathway quantitatively, is toxicologically important because its downstream metabolites are bioactivated by renal beta-lyase to reactive species implicated in kidney toxicity and carcinogenicity (Green et al., 1997b). Rodents generally metabolize TCE more rapidly than humans through both pathways (Lash et al., 2014), and there is evidence that rats may generate reactive GSH-pathway metabolites at a greater rate than humans based on beta-lyase activity (Lash et al., 1990), a species difference that carries directly into the interspecies uncertainty factors applied in dose-response modeling (Section 7).

Elimination

Unmetabolized TCE and CO2 are eliminated via exhalation, while polar metabolites (trichloroacetic acid, trichloroethanol, and GSH-pathway metabolites) are excreted primarily in urine and feces, with minor contributions from sweat and saliva. Pulmonary elimination is slow, with terminal half-lives as long as 44 hours reflecting substantial partitioning into adipose tissue. Urinary elimination kinetics differ markedly by species: TCE and its metabolites have a reported elimination half-life of just over 50 hours in humans compared with approximately 16 hours in rats (Ikeda and Imamura, 1973).

Because internal dose (rather than external exposure concentration) drives target-organ toxicity, and because rodents and humans metabolize TCE at different rates through different pathways, physiologically based pharmacokinetic (PBPK) modeling is used to convert applied doses in animal studies into human-equivalent internal dose metrics — for example, the amount of TCE oxidized in liver per unit body weight, or the amount of TCE metabolized via the GSH pathway — before dose-response modeling and uncertainty-factor application. This PBPK-based approach underlies essentially all of the point-of-departure derivations discussed below.

4. Toxicological Evidence Base (Animal Studies)

Animal toxicology on TCE spans multiple exposure routes (inhalation, oral gavage, and drinking water), durations (single-dose to chronic), and species (principally rats and mice), and has consistently identified the liver, kidney, immune system, nervous system, and developing conceptus as the most sensitive targets. Representative studies underlying the point-of-departure derivations are summarized below by organ system.

Liver

Kjellstrand et al. (1983) exposed male NMRI mice by inhalation to TCE concentrations ranging from 37 to 3,600 ppm for 1–24 hours/day over 30–120 days and observed a dose-responsive increase in relative liver weight at all exposure levels, with vacuolization and inflammatory cell infiltration at higher concentrations; benchmark dose modeling of these data yielded a BMDL10 of 21.6 ppm for increased liver-to-body-weight ratio. Buben and O’Flaherty (1985) administered TCE by gavage to male Swiss-Cox mice (100–3,200 mg/kg-bw/day, 5 days/week for 6 weeks) and identified a BMDL10 of 82 mg/kg-bw/day for the same endpoint, accompanied by cytotoxicity, histopathologic change, and reduced hepatic glucose-6-phosphatase activity. Woolhiser et al. (2006) exposed female Sprague-Dawley rats by inhalation (100–1,000 ppm, 6 hours/day, 5 days/week for 4 weeks) and derived a BMDL10 of 25 ppm for increased liver-to-body-weight ratio.

Kidney

Maltoni et al. (1986) administered TCE by gavage to male rats 4–5 days/week for 52 weeks and reported pathologic changes in the renal tubule, yielding a BMDL10 of 34 mg/kg-bw/day when internal dose was modeled using the glutathione-pathway metabolite dose metric associated with renal bioactivation. Kidney toxicity and carcinogenicity are mechanistically linked to the DCVG/DCVC bioactivation pathway described in Section 3 (Green et al., 1997b).

Nervous System

Arito et al. (1994) exposed male rats by inhalation to TCE at concentrations as low as 12 ppm (8 hours/day, 5 days/week for 6 weeks) and reported significant decreases in wakefulness, yielding a LOAEL-based POD. In humans, Ruijten et al. (1991) evaluated workers with a mean of 16 years of occupational TCE exposure and identified trigeminal nerve effects — increased latency of the masseter reflex — at a LOAEL of 14 ppm, providing one of the few human-derived quantitative PODs for TCE. Mechanistic work using cellular and rodent models has also implicated mitochondrial complex I inhibition by TCE as a plausible pathway for dopaminergic neurodegeneration (Gash et al., 2008; Liu et al., 2018), relevant to the epidemiological signal for Parkinson’s disease discussed in Section 5.

Immune System

Sanders et al. (1982) exposed female mice to TCE in drinking water for 16 or 24 weeks and reported immunosuppression at a LOAEL of 18 mg/kg-bw/day. Selgrade and Gilmour (2010) exposed female mice to a single 3-hour inhalation dose of TCE followed by respiratory bacterial challenge and observed dose-related mortality due to immunosuppression, yielding a BMDL01 of 13.9 ppm (applied dose) — the most robust POD for acute non-cancer effects (see Section 7). Keil et al. (2009) administered TCE in drinking water to female B6C3F1 mice, a strain not genetically prone to autoimmunity, for 27–30 weeks and observed increased anti-dsDNA and anti-ssDNA autoantibodies at a LOAEL of 0.35 mg/kg-bw/day — the study selected by EPA as the best overall chronic non-cancer POD.

Reproductive and Developmental Effects

Narotsky et al. (1995) administered TCE by gavage to pregnant rats during gestation and identified a BMDL01 of 32.2 mg/kg-bw/day for increased full-litter resorptions (gestational days 6–15 dosing) and a LOAEL of 475 mg/kg-bw/day for delayed parturition. Johnson et al. (2003) dosed pregnant rats with TCE-contaminated drinking water throughout gestation and reported congenital heart defects at a BMDL01 of 0.0207 mg/kg-bw/day, the lowest POD identified in the toxicological database, though this study carries medium confidence due to statistical and methodological limitations discussed in Section 9. Fredriksson et al. (1993) exposed neonatal rat pups by gavage on postnatal days 10–16 and reported decreased rearing activity, a marker of developmental neurotoxicity, at a LOAEL of 50 mg/kg-bw/day. In humans, Chia et al. (1996) evaluated semen parameters in male workers occupationally exposed to TCE over a mean of 5.1 years and identified decreased normal sperm morphology and hyperzoospermia at a human-derived BMDL10 of 1.4 ppm.

Carcinogenicity Bioassays

Rodent cancer bioassays, most notably Maltoni et al. (1986), demonstrate multisite carcinogenicity of TCE, including kidney tumors in rats and liver and lung tumors in mice, with inhalation unit risk estimates from the most sensitive of these bioassays ranging from roughly 1 × 10⁻² to 2 × 10⁻¹ per ppm. These animal findings are considered supportive of, but not the primary basis for, the human cancer risk values discussed in Section 8, which instead rely on high-quality human epidemiological data.

5. Epidemiological Evidence Base (Human Studies)

The human evidence base for TCE spans occupational cohort and case-control studies, community drinking-water exposure studies, and biomonitoring investigations, and is central to both the cancer and non-cancer hazard characterization for this chemical.

Cancer Epidemiology

Charbotel et al. (2006) conducted a case-control study of renal cell carcinoma among workers in the Arve Valley region of France with detailed, task-based quantitative estimates of cumulative TCE exposure derived from decades of workplace measurement; this High-quality study forms the principal quantitative basis for TCE’s cancer slope factor and inhalation unit risk (Section 8). Raaschou-Nielsen et al. (2003) followed a large Danish cohort of workers at companies using TCE and reported elevated standardized incidence ratios for kidney cancer, liver cancer, and non-Hodgkin lymphoma (NHL), providing corroborating multisite evidence used to adjust the kidney-cancer-based risk values upward for the two additional cancer sites. Additional occupational cohort studies — including Anttila et al. (1995) in Finnish workers and Axelson et al. (1994) in an updated Swedish cohort — have examined cancer incidence and mortality among TCE-exposed workers, contributing to the broader weight-of-evidence meta-analyses of kidney cancer, liver cancer, and NHL risk.

Immunological and Hematological Studies

Bassig et al. (2016) compared hematological alterations and markers of B-cell activation among workers exposed to benzene, formaldehyde, and TCE, contributing biomarker-level human evidence relevant to TCE’s association with lymphohematopoietic effects, including NHL.

Community Drinking-Water Exposure Studies

Communities with TCE-contaminated drinking water have been the subject of several major epidemiological investigations. At Marine Corps Base Camp Lejeune, North Carolina, Ruckart et al. (2013) conducted a case-control study of birth defects and childhood cancers, and Ruckart et al. (2014) conducted a cross-sectional evaluation of preterm birth, small-for-gestational-age status, and birth weight, in each case examining associations with volatile organic compound exposure (including TCE) in contaminated base drinking water. Bove et al. (2014a) and Bove et al. (2014b) separately evaluated mortality among Marine and Navy personnel and civilian employees exposed to the same contaminated water supply. Earlier work by Bove (1996) and Bove et al. (1995) examined public drinking water contamination in New Jersey in relation to birthweight, prematurity, fetal death, and birth defects, with statistically equivocal findings for cardiac defects specifically.

Neurological Studies

Ruijten et al. (1991), described above, provides the principal human quantitative POD for TCE neurotoxicity. Separately, Goldman et al. (2012) conducted a twin study examining solvent exposure and Parkinson’s disease risk and reported an association between TCE exposure and Parkinson’s disease, complementing the mechanistic mitochondrial toxicity findings of Gash et al. (2008).

6. Health Outcomes Associated with TCE Exposure, by Body System

The following overview organizes the illnesses and health outcomes associated with TCE exposure by body system and distinguishes findings with strong, quantitatively modeled dose-response support from those with suggestive or less consistent evidence.

Kidney

Strong, quantitatively modeled: Renal cell carcinoma is the cancer type with the strongest and most consistent human evidence, anchoring the quantitative cancer slope factor and inhalation unit risk (Charbotel et al., 2006). Non-cancer renal tubular pathology in rodents is also quantitatively modeled (Maltoni et al., 1986).

Liver

Strong, quantitatively modeled: Non-cancer hepatotoxicity — increased liver weight, cytotoxicity, and hypertrophy — is robustly demonstrated across multiple rodent studies and species (Kjellstrand et al., 1983; Buben and O’Flaherty, 1985; Woolhiser et al., 2006).

Suggestive/less consistent: Liver cancer is elevated in some human cohorts (Raaschou-Nielsen et al., 2003) and in mice but not rats in bioassay data; it is included in the combined cancer risk value as a secondary, supportive cancer site rather than the primary quantitative anchor.

Immune System

Strong, quantitatively modeled: Acute immunosuppression with increased susceptibility to respiratory infection (Selgrade and Gilmour, 2010) and chronic autoimmunity, evidenced by increased anti-dsDNA/anti-ssDNA autoantibodies (Keil et al., 2009), are the two endpoints EPA identified as the best-supported non-cancer PODs overall and are the basis of the reference values applied by RoseyFields (Section 7).

Suggestive/less consistent: Associations between TCE exposure and systemic autoimmune disease in humans, including scleroderma, have been reported but are less consistent across the epidemiological literature than the rodent autoimmunity findings. NHL is discussed under Cancer below.

Nervous System

Strong, quantitatively modeled: Human trigeminal nerve effects among long-tenured workers (Ruijten et al., 1991) and rodent effects on sleep/wakefulness (Arito et al., 1994) provide quantitative non-cancer PODs.

Suggestive/less consistent: An association with Parkinson’s disease has been reported in a human twin study (Goldman et al., 2012) and is supported mechanistically by evidence of mitochondrial complex I inhibition in animal and cellular models (Gash et al., 2008), but this association has not been quantitatively modeled for reference-value derivation.

Reproductive System

Strong, quantitatively modeled: Decreased sperm quality in occupationally exposed men (Chia et al., 1996) and delayed parturition and full-litter resorptions in rodents (Narotsky et al., 1995) are quantitatively characterized.

Developmental Effects

Strong, quantitatively modeled (with caveats): Full-litter resorption is a high-confidence rodent endpoint (Narotsky et al., 1995). Congenital heart defects (Johnson et al., 2003) and developmental neurotoxicity/decreased rearing activity (Fredriksson et al., 1993) are quantitatively modeled but carry medium confidence because of study-specific statistical and methodological limitations (Section 9).

Suggestive/less consistent: Community studies near contaminated drinking water sources have examined birth defects, preterm birth, and birth weight (Ruckart et al., 2013; Ruckart et al., 2014; Bove, 1996; Bove et al., 1995), with mixed and, for cardiac defects specifically, statistically equivocal results.

Cancer (Systemic)

Strong, quantitatively modeled: Kidney (renal cell) cancer, based on high-quality human case-control data (Charbotel et al., 2006), supported by rodent bioassay evidence of multisite carcinogenicity (Maltoni et al., 1986) and by a documented mutagenic mode of action for kidney tumors.

Moderate, incorporated via adjustment factor: Liver cancer and non-Hodgkin lymphoma are supported by cohort evidence (Raaschou-Nielsen et al., 2003) and meta-analysis, and are incorporated into the combined cancer risk values as adjustment factors applied to the kidney-cancer-based potency estimate, rather than independently modeled with the same level of confidence as kidney cancer.

7. Derivation of Non-Cancer Reference Values (Points of Departure and Uncertainty Factors)

EPA’s TCE risk evaluation derives its non-cancer reference values using a PBPK-based points-of-departure (POD) framework rather than a single classical oral reference dose (RfD) of the kind used in some IRIS toxicological reviews. Because TCE toxicity is understood to be metabolite-driven and the relevant metabolic pathways differ by target organ and by species, EPA modeled internal-dose PODs (idPODs) for each candidate endpoint using PBPK-derived dose metrics (e.g., total TCE metabolized per unit body weight, amount oxidized in liver, amount bioactivated via the glutathione pathway), then converted those internal-dose PODs to human-equivalent concentrations (HECs, for inhalation) and human-equivalent doses (HEDs, expressed as an oral-equivalent mg/kg-bw/day value) at both the 50th percentile (HEC50/HED50) and 99th percentile (HEC99/HED99) of the modeled human population, the latter intended to be protective of toxicokinetic variability across sensitive individuals. RoseyFields treats the HED99 values as functionally equivalent to a health-protective oral reference dose for the corresponding endpoint.

Uncertainty Factors Applied

Consistent with EPA’s RfD/RfC Guidance, standard composite uncertainty factors were applied to each POD:

• UFS (subchronic-to-chronic): applied when a study of less-than-chronic duration is used to represent a chronic endpoint.
• UFA (interspecies): typically 3 when PBPK modeling has already accounted for toxicokinetic differences between species, addressing only the residual toxicodynamic uncertainty.
• UFH (intraspecies/human variability): typically 3 when the HEC99/HED99 value is used (partially accounting for human variability), or 10 when a study was not amenable to PBPK modeling and no HEC99/HED99 adjustment was available.
• UFL (LOAEL-to-NOAEL): applied when a study’s POD is a LOAEL rather than a NOAEL or benchmark dose lower confidence limit (BMDL).

Composite uncertainty factors across the selected endpoints ranged from 10 (e.g., Kjellstrand et al., 1983, liver; Selgrade and Gilmour, 2010, acute immune) to 300 (e.g., Fredriksson et al., 1993, developmental neurotoxicity), reflecting differences in study duration, POD type, and the availability of PBPK-based human variability adjustment.

Selected Best-Overall Non-Cancer Endpoints

Of all endpoints evaluated, EPA identified acute immunosuppression and chronic autoimmunity as the best-overall non-cancer endpoints for risk characterization, based on data quality, dose-response robustness, biological severity, and reduced extrapolation uncertainty relative to competing endpoints (including the lower, but less well-supported, PODs for congenital heart defects and developmental neurotoxicity):

RoseyFields uses these HED99 values — approximately 1.34 mg/kg-bw (acute, single-exposure equivalent) and 0.048 mg/kg-bw/day (chronic) — as the working oral-equivalent reference values for TCE non-cancer risk characterization, applying the corresponding HEC99 values directly for inhalation exposure pathways. Organ-system-specific PODs (liver, kidney, nervous system, and reproductive/developmental endpoints, summarized in Section 4) remain available and are substituted when a site-specific exposure scenario or sensitive subpopulation makes a different endpoint more relevant than the default immune-system values.

8. Carcinogenicity Classification and Cancer Slope Factor

TCE is classified as carcinogenic to humans, a conclusion supported by high-quality human epidemiological data across three cancer sites (kidney, liver, and non-Hodgkin lymphoma), corroborating rodent bioassay evidence of multisite carcinogenicity, and an established mutagenic mode of action for kidney tumorigenesis that supports linear, non-threshold low-dose extrapolation.

Inhalation Unit Risk (IUR)

The inhalation unit risk for TCE is 2.2 × 10⁻² per ppm (approximately 4 × 10⁻⁶ per µg/m³), representing the plausible upper-bound lifetime extra risk of cancer per unit air concentration from chronic inhalation. This value is anchored to human kidney cancer risk from Charbotel et al. (2006) — calculated directly from that study as 5.49 × 10⁻³ per ppm — and then adjusted upward by a factor of four to account for the additional contribution of liver cancer and NHL. That four-fold adjustment factor was derived from two independent calculations: a ratio of 3.28 based on summary relative-risk estimates from meta-analysis of kidney, liver, and NHL studies, and a ratio of 4.36 based on site-specific standardized incidence ratios reported in Raaschou-Nielsen et al. (2003); the geometric and arithmetic mean of these two ratios (3.8) was rounded up to 4.

Oral Slope Factor (OSF)

The oral slope factor is 4.64 × 10⁻² per mg/kg-day (rounded to 5 × 10⁻² per mg/kg-day), representing the plausible upper-bound lifetime extra cancer risk per unit of chronic oral dose. The OSF was derived by PBPK-model-based route-to-route extrapolation from the same Charbotel et al. (2006) kidney cancer risk estimate, using site-specific internal dose metrics for each of the three cancer types and summing the resulting site-specific oral slope factors to obtain a combined adjustment factor of 5.0 relative to kidney cancer alone (somewhat larger than the 4-fold inhalation adjustment because of additional variability introduced by using distinct dose metrics for different target tissues).

Because the human kidney cancer data provide High-quality, directly estimable risk values, this approach avoids much of the interspecies extrapolation uncertainty inherent in rodent-bioassay-based cancer potency estimates, though it is instead subject to the epidemiological and exposure-reconstruction uncertainties inherent in the Charbotel et al. (2006) case-control design. Consistent with 2016 NIOSH guidance, an excess lifetime cancer risk benchmark of 1 × 10⁻⁴ is used as a reference point for occupational risk determinations, though this is treated as a risk-management reference point rather than a scientific bright line.

9. Uncertainties and Sensitive Populations

Key Uncertainties

• PBPK dose-metric selection: for kidney effects specifically, HEC50/HED50-to-HEC99/HED99 ratios span 7- to 10-fold (versus a typical 2- to 3-fold spread for other endpoints), reflecting greater uncertainty in modeling the glutathione-conjugation pathway and reduced confidence in mouse-derived data for this metric.
• Developmental endpoints with lower confidence: the congenital heart defect POD (Johnson et al., 2003) and the developmental neurotoxicity POD (Fredriksson et al., 1993) carry medium (rather than high) data-quality ratings due to statistical and methodological limitations, even though they represent the lowest numerical PODs identified.
• Route-to-route and species extrapolation: the oral slope factor is derived by extrapolating a human inhalation-based cancer risk estimate through a PBPK model rather than from directly observed oral human cancer data.
• Cancer type combination: summing extra risks across three cancer sites using ratio-based adjustment factors introduces additional imprecision, though EPA’s overall assessment concludes this is unlikely to change the combined IUR/OSF by more than roughly two-fold.
• Acute-to-chronic extrapolation of cancer risk: methods for extrapolating theoretical lifetime cancer risk to single short-term exposures carry substantial uncertainty, so cancer risk is evaluated only for lifetime-averaged (chronic) exposure scenarios, not acute exposures.

Sensitive and Susceptible Subpopulations

• Pregnant women and the developing fetus, given the quantitative and mechanistic evidence for full-litter resorption, delayed parturition, and (with medium confidence) congenital heart defects.
• Infants, via TCE detected in breast milk following maternal exposure.
• Children and adolescents, given developmental neurotoxicity findings in juvenile animal models.
• Individuals with pre-existing autoimmune predisposition, given the chronic autoimmunity POD used as a primary non-cancer reference value.
• Long-tenure workers with high cumulative occupational exposure, reflecting the exposure durations underlying both the epidemiological cancer data and the neurological findings of Ruijten et al. (1991).
• Individuals with genetic polymorphisms affecting CYP2E1 or glutathione-pathway enzyme activity, which could increase production of reactive metabolites relative to the general population.

10. Application of Reference Values in RoseyFields Risk Assessments

Non-Cancer Risk: Hazard Quotient and Hazard Index

For each relevant exposure pathway and averaging time, RoseyFields calculates a hazard quotient (HQ) as the ratio of the estimated exposure concentration or dose to the appropriate POD-derived reference value (HEC99/HED99, or route-specific reference concentration/dose):

HQ = Exposure Estimate (ADC or ADD) ÷ Reference Value (HEC99 or HED99)

An HQ at or below 1 indicates the exposure is not expected to result in the associated adverse effect with an adequate margin of safety; an HQ above 1 signals potential concern warranting further evaluation. Where multiple non-cancer endpoints or exposure pathways are relevant to the same population, RoseyFields sums the individual HQs into a hazard index (HI), generally grouped by common target organ or effect (e.g., a liver-toxicity HI, an immune-system HI) rather than combined indiscriminately across dissimilar endpoints, consistent with EPA’s guidance for cumulative risk assessment.

Cancer Risk: Excess Lifetime Cancer Risk

For scenarios involving chronic or lifetime exposure, RoseyFields calculates excess lifetime cancer risk (ELCR) using the lifetime average daily concentration or dose (LADC/LADD) and the applicable slope factor:

ELCR = LADC × IUR (0.022 per ppm), or LADD × OSF (0.0464 per mg/kg-day)

Resulting ELCR estimates are compared against applicable regulatory benchmarks (commonly in the 1 × 10⁻⁶ to 1 × 10⁻⁴ range depending on the regulatory and population context) to characterize whether estimated risk falls within an acceptable range for the exposure scenario under evaluation.

Dose Reconstruction

Where direct exposure monitoring data are unavailable, RoseyFields reconstructs exposure using validated modeling approaches appropriate to the exposure setting — for example, near-field/far-field inhalation modeling for occupational scenarios, dermal absorption modeling accounting for evaporative loss, and fate-and-transport or vapor-intrusion modeling for groundwater- or soil-vapor-impacted sites — combined with population-specific behavioral parameters (exposure duration, frequency, body weight, and, for cancer risk, total years of exposure over a lifetime averaging period). These reconstructed exposure estimates are then compared against the appropriate acute, chronic non-cancer, or lifetime cancer reference value using the HQ/HI and ELCR frameworks described above.

11. Conclusion

TCE presents a well-characterized, multi-organ hazard profile supported by a substantial toxicological and epidemiological evidence base. Non-cancer effects on the liver, kidney, nervous system, immune system, and reproductive/developmental systems are each supported by quantitatively modeled animal or human dose-response data, with acute immunosuppression and chronic autoimmunity serving as the best-overall non-cancer endpoints for reference-value derivation. TCE is carcinogenic to humans, with a quantitative cancer potency anchored in high-quality human kidney cancer data and adjusted to reflect corroborating evidence for liver cancer and non-Hodgkin lymphoma. RoseyFields applies the resulting reference values — non-cancer HEC99/HED99 values and cancer inhalation unit risk/oral slope factor estimates — within a standard hazard quotient, hazard index, and excess lifetime cancer risk framework, informed by scenario-specific dose reconstruction, to characterize risk for the populations and exposure pathways relevant to each site-specific assessment.

Bibliography