Hexavalent Chromium (Cr-VI) Risk Assessment and Lung Cancer

RoseyFields conducts hexavalent chromium risk assessments using the following methodology.

Chemical Identity, Regulatory Background, and Exposure Sources

Hexavalent chromium, or Cr(VI) (CASRN 18540-29-9), refers to a group of compounds containing chromium in the +6 oxidation state, distinct from the far less toxic trivalent form, Cr(III), which is an essential trace nutrient. Cr(VI) compounds—chromic acid, sodium and potassium dichromate, chromate pigments, and related salts—are strong oxidizers, are generally water-soluble, and are consequently highly mobile in soil and groundwater. Cr(VI) does not persist indefinitely in the environment; it is reduced to Cr(III) by organic matter, ferrous iron, and other reducing agents, but this reduction capacity can be exceeded or overwhelmed in contaminated groundwater, which is why Cr(VI) plumes have proven so persistent at sites such as the Hinkley, California groundwater contamination associated with a natural-gas compressor station’s cooling-tower operations.

RoseyFields’ assessors most frequently encounter Cr(VI) exposure in three settings: occupational exposure via inhalation of chromic acid mist or chromate dust in electroplating, chrome plating, stainless steel and alloy welding, chromate pigment and dye manufacturing, and leather tanning; occupational or environmental dermal and inhalation exposure at wood-preservative (chromated copper arsenate) treatment facilities; and oral exposure via contaminated drinking water, where Cr(VI) has leached from industrial cooling towers, chrome plating waste, or naturally oxidized geologic chromium into groundwater supplies. Combustion of coal and other fossil fuels is an additional, more diffuse source of airborne Cr(VI).

Cr(VI)’s regulatory profile reflects both its occupational and its environmental significance. OSHA’s 2006 Cr(VI) standard (29 C.F.R. §§ 1910.1026, 1915.1026, 1926.1126) set an 8-hour time-weighted-average permissible exposure limit of 5 µg/m³, a roughly tenfold reduction from the prior limit. IARC classifies Cr(VI) compounds as Group 1 (carcinogenic to humans), based on well-established occupational lung cancer evidence. At the federal level, the Safe Drinking Water Act maximum contaminant level for total chromium (both Cr(III) and Cr(VI)) is 100 µg/L, but no chromium-speciated federal MCL exists. California has taken a different path: after its original 2014 hexavalent chromium-specific MCL of 10 µg/L was invalidated by a state court in 2017 on cost-benefit grounds, the State Water Resources Control Board readopted a Cr(VI) MCL of 10 µg/L effective October 1, 2024, with a compliance extension to October 1, 2028 for small systems. Cr(VI) is also a RCRA hazardous waste constituent, a CERCLA hazardous substance, and a listed Proposition 65 carcinogen in California. RoseyFields’ assessors track this patchwork of federal and state standards closely, since the applicable regulatory benchmark in a given matter often depends on jurisdiction as much as on the underlying toxicity value discussed below.

Point-of-Departure Selection

EPA’s most recent (August 2024) toxicological review of Cr(VI) departs from a single-critical-study model in favor of organ/system-specific reference values, each with its own point of departure, reflecting the breadth of the modern toxicological database. RoseyFields’ assessors work from this same organ-specific architecture rather than collapsing it into a single number, because litigation and site-specific matters frequently turn on a particular target organ.

For the oral pathway, the overall chronic RfD is driven by gastrointestinal tract toxicity: diffuse epithelial hyperplasia of the duodenum in female B6C3F1 mice from the two-year drinking-water bioassay conducted by the National Toxicology Program (NTP, 2008). A LOAEL analysis—rather than benchmark dose (BMD) modeling—was used for this endpoint because the hyperplasia data in female mice were not well suited to dose-response modeling; the resulting POD is a human-equivalent dose (HED) of 0.0911 mg/kg-day. (The male mouse hyperplasia data were amenable to BMD modeling, yielding a POD roughly five-fold lower before uncertainty factors, but the female LOAEL-based value was carried forward as the organ-specific RfD basis.) Three additional organ-specific oral reference values were derived in parallel: a hepatic osRfD based on chronic inflammation in female F344 rats (NTP, 2008); a developmental osRfD based on decreased F1 offspring postnatal growth in a continuous-breeding study in BALB/c mice (NTP, 1997); and a hematological osRfD based on decreased hemoglobin in male F344 rats after only 22 days of exposure (NTP, 2008). RoseyFields’ assessors apply the organ-specific value most relevant to the health outcome at issue in a given matter, while defaulting to the overall (GI-tract-based) RfD of 9 × 10⁻⁴ mg/kg-day when no more specific endpoint is implicated, since GI toxicity is both the most sensitive endpoint and one for which confidence is rated medium-high.

For the inhalation pathway, the critical effect—and the basis for the overall RfC—is upper respiratory tract toxicity: ulceration of the nasal septum in occupationally exposed chromate production workers, drawn primarily from the longitudinal clinical study of Gibb et al. (2000a), with corroborating LOAELs from Cohen et al. (1974) and Lindberg and Hedenstierna (1983). Nasal septal effects were selected as the critical effect because they are well established, occur at the lowest exposure levels of any Cr(VI)-related endpoint, and are therefore considered protective of the other (hepatic, developmental, hematological, immune, and male reproductive) effects that inhalation exposure can also produce. Notably, the RfC’s principal and supporting studies are entirely human occupational data, in contrast to the RfD’s reliance on rodent bioassays—a distinction that materially shapes the pharmacokinetic and uncertainty-factor treatment described in Sections 3 and 4.

Pharmacokinetic Modeling and Dose Conversion

Cr(VI)’s dosimetry is unusually route- and site-dependent, and RoseyFields’ assessors treat the choice of dose conversion method as a substantive scientific judgment rather than a mechanical step, because EPA itself used different methods for different tumor and toxicity sites within the same assessment.

For gastrointestinal endpoints—both the GI tract hyperplasia underlying the RfD and the small intestinal tumors underlying part of the oral cancer assessment—EPA developed a gastric physiologically based pharmacokinetic (PBPK) model that simulates the reduction of ingested Cr(VI) to the much less toxic Cr(III) by gastric juices, saliva, and first-pass hepatic metabolism before any Cr(VI) reaches the small intestine. This model is necessary because only the Cr(VI) that survives gastric reduction reaches the small intestine to produce the hyperplasia and tumors observed in the NTP bioassays, and the fraction surviving is nonlinear with dose and differs meaningfully between rodents and humans due to differences in stomach pH, gastric transit time, and body-weight-normalized gastric fluid volume. The gastric PBPK model was used to calculate human-equivalent doses for the GI tract RfD, the hepatic and developmental osRfDs, and the mouse small-intestinal-tumor oral slope factor.

By contrast, for effects occurring in the oral cavity or tongue—both the hepatic ALT elevations evaluated using simple NOAEL/LOAEL analysis in some supporting studies and, more importantly, the squamous cell tumors of the oral mucosa and tongue that anchor the final oral slope factor—no gastric reduction has yet occurred, because these tissues are exposed to ingested Cr(VI) before it reaches the stomach. For this portal-of-entry exposure, EPA instead applied simple body-weight-to-the-three-quarters-power (BW¾) allometric scaling (human-equivalent dose equals ingested dose multiplied by [BW{animal}/BW{human, 80 kg}]^0.25), consistent with EPA’s default cross-species scaling guidance (U.S. EPA, 2011), rather than the gastric PBPK model, because there is no adequately developed theory to construct an oral-cavity-specific dosimetric adjustment factor. RoseyFields’ assessors view this dual approach—a mechanistic PBPK model for post-gastric-reduction sites and simple allometric scaling for pre-gastric-reduction sites—as scientifically appropriate but also as a source of method-dependent uncertainty when comparing potency estimates across tumor sites, discussed further in Section 9.

For the inhalation RfC, no PBPK modeling was performed and none was needed, because the critical and supporting studies are direct human occupational measurements of nasal effects at measured air concentrations—there is no animal-to-human extrapolation step in this pathway at all. This mirrors the pattern RoseyFields’ assessors have observed in other chemical-specific assessments (e.g., mercury vapor) where human-based inhalation values require no interspecies pharmacokinetic conversion, in contrast to oral, rodent-based values.

Uncertainty Factor Derivation

Because Cr(VI) toxicity is characterized by organ-specific reference values with different critical studies, RoseyFields’ assessors apply a different composite uncertainty factor to each osRfD and to the RfC, rather than a single UF across the whole assessment.

For the GI tract and hepatic oral reference values, a composite UF of 100 is applied to each, composed of an interspecies factor (UFA) of 3—reflecting that the PBPK modeling already resolves the toxicokinetic component of animal-to-human extrapolation, leaving only residual toxicodynamic (pharmacodynamic) uncertainty—an intraspecies factor (UFH) of 3, addressing human variability and the possibility that the available data do not capture the most susceptible individuals, and a LOAEL-to-NOAEL factor (UFL) of 10, applied because BMD modeling was not feasible for either endpoint and a LOAEL was used directly as the POD. No subchronic-to-chronic factor (UFS) was needed for either value because both derive from the NTP (2008) two-year chronic bioassay.

For the developmental oral reference value, a composite UF of 10 is applied (UFA = 3, UFH = 3), with no LOAEL to-NOAEL factor because the POD was extrapolated from a NOAEL (0.700 mg/kg-day HED, from the continuous breeding study in BALB/c mice; NTP, 1997) rather than a LOAEL.

For the hematological oral reference value, a composite UF of 10 is likewise applied (UFA = 3, UFH = 3), with the POD derived via BMD modeling (a BMDL corresponding to a one-standard-deviation change in hemoglobin) rather than a LOAEL, so no UFL was needed. Notably, no subchronic-to-chronic factor was applied here either, but for a different reason than for the chronic-study-based values: EPA determined that Cr(VI)-induced hematological effects are most pronounced early in exposure and attenuate over time, so using longer-duration exposure data would actually understate the relevant short-term risk; the short-term, low-dose data from NTP (2008) were used directly as reflecting the appropriately protective exposure duration for this endpoint, rather than being adjusted upward for presumed chronic under-estimation.

For the inhalation RfC, a composite UF of 100 is applied, but with a materially different composition than the oral values: because the critical and supporting studies are entirely human, no interspecies factor is needed at all. Instead, the UF of 100 comprises an intraspecies factor (UFH) of 3, a LOAEL-to-NOAEL factor (UFL) of 10—applied because nasal septal ulceration showed a high incidence even at the lowest concentration tested across multiple studies, precluding NOAEL identification—and a subchronic-to-chronic factor (UFS) of 3, applied because the underlying occupational studies were not lifetime chronic exposures, tempered by the fact that nasal septal effects have a short onset time and are therefore reasonably captured even by less-than-lifetime occupational monitoring.

Derivation of the Oral RfD and Inhalation RfC

Applying the organ-specific PODs and their respective composite uncertainty factors yields the following organ specific reference values, with the GI-tract value serving as the overall RfD:

Overall chronic oral RfD = 9 × 10⁻⁴ mg/kg-day (GI tract hyperplasia; NTP, 2008; POD [LOAEL-HED] = 0.0911 mg/kg-day ÷ UF 100 = 9.11 × 10⁻⁴, rounded to 9 × 10⁻⁴).

Supporting organ-specific values, available for use where a matter implicates a more specific endpoint, include a hepatic osRfD of 7 × 10⁻⁴ mg/kg-day (chronic inflammation; NTP, 2008), a developmental osRfD of 0.07 mg/kg-day (decreased postnatal growth; NTP, 1997), and a hematological osRfD of 0.01 mg/kg-day (decreased hemoglobin; NTP, 2008).

Overall chronic inhalation RfC = 3 × 10⁻⁵ mg/m³ (upper respiratory tract—nasal septum ulceration; Gibb et al., 2000a; POD [LOAEL] = 3.4 × 10⁻³ mg/m³ ÷ UF 100 = 3.4 × 10⁻⁵, rounded, and cross-checked against the Cohen et al. (1974) and Lindberg and Hedenstierna (1983) candidate values of 9.5 × 10⁻⁶ and 6.6 × 10⁻⁶ mg/m³, respectively, all of which are protective of, or consistent with, the adopted value).

RoseyFields’ assessors note that this RfC is more than an order of magnitude below OSHA’s occupational PEL of 5 µg/m³ (0.005 mg/m³)—an expected and appropriate divergence, since the RfC is designed to be protective of continuous lifetime exposure of the general population, including children, rather than of a working-age adult during an 8-hour shift.

Carcinogenicity Classification and Quantitative Cancer Risk Values

EPA’s carcinogenicity conclusions for Cr(VI) differ by exposure route, and RoseyFields’ assessors apply each accordingly. By the oral route, Cr(VI) is classified as likely to be carcinogenic to humans, based on robust animal evidence and only slight direct human evidence, but strong supporting mechanistic evidence. By the inhalation route, Cr(VI) is classified more strongly, as carcinogenic to humans, reflecting long-standing, consistent epidemiologic evidence of lung cancer among chromate production, chrome plating, and pigment workers—a classification EPA first reached in 1998 and has maintained, and one shared by OSHA, NIOSH, NTP, IARC, and California EPA.

The oral cancer database rests on the same NTP (2008) two-year drinking-water bioassay that anchors the noncancer RfD, which produced “clear evidence of carcinogenic activity”: squamous cell tumors of the oral cavity and tongue in rats, and adenomas and carcinomas of the small intestine in mice. Because Cr(VI) carcinogenesis operates through a mutagenic mode of action that is sufficiently supported in animals and considered relevant to humans, EPA applied linear low-dose extrapolation from the point of departure rather than a threshold (RfD-style) approach, and applied age-dependent adjustment factors (ADAFs) to account for presumed increased early-life susceptibility to mutagenic carcinogens, in the absence of chemical-specific pediatric data.

Oral slope factor (adult exposure, no ADAF) = 0.16 (mg/kg-day)⁻¹, based on squamous cell carcinomas and papillomas of the oral mucosa and tongue in female rats (NTP, 2008; BMDL10%ER = 2.70 mg/kg-day ingested dose; human-equivalent dose = 0.645 mg/kg-day via BW¾ scaling; slope = 0.1/0.645 ≈ 0.155, rounded to 0.16). Comparable slope factors of similar magnitude (0.098–0.11 (mg/kg-day)⁻¹) were derived from the male rat oral cavity data and from small-intestinal tumors in male and female mice via the gastric PBPK model, providing reasonable cross validation across sex, species, and tumor site.

Total lifetime exposure oral slope factor (with ADAFs applied) = 0.27 (mg/kg-day)⁻¹ (rounded from 0.265, per the detailed ADAF derivation; RoseyFields’ assessors note that EPA’s own executive summary elsewhere rounds this same lifetime value to 0.26, an internal rounding inconsistency in the source document that assessors resolve by relying on the fully documented derivation table rather than the headline summary figure).

For the inhalation pathway, the inhalation unit risk is based on the long-studied Baltimore, Maryland chromate production worker cohort, most recently reanalyzed by Gibb et al. (2020) to characterize how the relative risk of lung cancer mortality varies with age at exposure, building on the original cohort description and exposure-response analysis in Gibb et al. (2000b).

Inhalation unit risk (adult exposure, no ADAF) = 1.1 × 10⁻² per µg Cr(VI)/m³.

Total lifetime exposure inhalation unit risk (with ADAFs applied) = 1.8 × 10⁻² per µg Cr(VI)/m³.

Application in Practice

RoseyFields’ risk assessors apply Cr(VI)’s organ-specific and route-specific toxicity values through the same core risk assessment architecture used across chemicals, adapted to Cr(VI)’s multi-organ, multi-route profile.

For noncancer hazard evaluation, a Hazard Quotient is calculated as HQ = CDI/RfD (oral) or HQ = EC/RfC (inhalation), using either the overall RfD/RfC or the organ-specific value most relevant to the health outcome at issue—for example, using the hematological osRfD of 0.01 mg/kg-day rather than the overall GI-tract RfD when a case centers on anemia or hemoglobin changes. Where multiple Cr(VI) exposure pathways or co-occurring chemicals affect the same target organ, individual HQs are summed into a target-organ-specific Hazard Index. Assessors are careful not to conflate organ-specific and overall reference values within a single HQ or HI calculation, since doing so would either understate risk (using the less sensitive overall value for a more sensitive endpoint) or double-count uncertainty (applying multiple organ-specific UFs to the same exposure without a coherent basis for summation).

For cancer risk evaluation, excess lifetime cancer risk is calculated as ELCR = CDI × OSF (oral) or ELCR = EC × IUR (inhalation), using route-specific age-dependent adjustment factors when early-life exposure is at issue—10-fold for exposure ages younger than 2 years, 3-fold for ages 2 to younger than 16, and 1-fold for ages 16 and older, applied to age-specific exposure estimates and summed across the exposure period of interest, exactly as with other mutagenic carcinogens in RoseyFields’ portfolio (e.g., methylene chloride). Assessors do not apply the oral slope factor or inhalation unit risk to exposures exceeding the underlying points of departure, and they document which tumor site and sex (e.g., female rat oral cavity tumors for the oral pathway) drives the applied slope factor, since alternative, similarly well-supported slope factors exist for other tumor sites and could be relevant to a sensitivity analysis.

Exposure and dose reconstruction for Cr(VI) matters typically requires speciation-specific sampling or modeling, since total chromium measurements (routine in many historical environmental datasets) do not distinguish the far more toxic Cr(VI) from background Cr(III); RoseyFields’ assessors routinely evaluate whether historical “total chromium” data can be reasonably apportioned to Cr(VI), based on process knowledge (e.g., chrome plating baths are essentially pure Cr(VI)) or site-specific geochemistry, before applying the toxicity values derived above. In litigation and regulatory-compliance matters—drinking water contamination litigation, chrome-plating and welding occupational exposure claims, and RCRA/CERCLA cost-recovery actions—this full chain from critical study through organ-specific POD, uncertainty factors, and route-specific reference value or cancer potency, to case-specific HQ/HI or ELCR, is documented transparently to withstand Daubert or analogous scrutiny, with particular attention to the Cr(VI)-versus-total-chromium speciation issue, which is frequently contested.

Key Uncertainties and Sensitive Populations

EPA’s 2024 review identifies an unusually well-characterized set of susceptible populations and life stages for Cr(VI), and RoseyFields’ assessors apply each in case-specific work. Individuals with chronically elevated stomach pH— whether from hypochlorhydria, use of proton-pump inhibitors or other GERD medications, or simple population variability—are expected to detoxify ingested Cr(VI) less effectively via gastric reduction, increasing GI tract uptake and dose for a given ingested amount. This same mechanism makes neonates, infants, and toddlers younger than 30 months particularly susceptible, since young children have physiologically elevated gastric pH and correspondingly reduced gastric reduction capacity. Elderly individuals (65 and older) are identified as susceptible both because age related conditions can elevate stomach pH and because pre-existing GI, liver, respiratory, and hematological conditions common in this population can be independently exacerbated by Cr(VI) exposure. Individuals with genetic polymorphisms conferring reduced DNA repair capacity, and carriers of mutated cystic fibrosis transmembrane conductance regulator (CFTR) alleles—shown in animal models to enhance intestinal tumorigenesis when CFTR is suppressed—represent additional genetically defined susceptible subgroups relevant to Cr(VI)’s mutagenic carcinogenic pathway specifically.

RoseyFields’ assessors also flag several structural uncertainties in the assessment itself. Below an oral dose of approximately 1 mg/kg-day, EPA’s own pharmacokinetic analysis acknowledges substantial uncertainty as to whether Cr(VI) reaches systemic tissues beyond the GI tract and liver in consistent, dose-related amounts—meaning inconsistent results across studies at low doses are expected rather than anomalous, a consideration relevant when evaluating any single low-dose study in isolation. The oral cancer slope factor also rests on two different dose conversion methods depending on tumor site (gastric PBPK modeling for small intestinal tumors versus simple BW¾ scaling for oral cavity tumors), and while the resulting potency estimates are reassuringly similar in magnitude across sites, RoseyFields’ assessors treat this methodological divergence as a genuine source of cross-study uncertainty rather than eliding it. The developmental osRfD carries only low confidence because decreased offspring growth was observed only at doses where other toxicity may have been occurring in the dams, and the hepatic osRfD’s confidence is likewise tempered by the minimal severity of the chronic inflammation on which it is based. Finally, EPA’s own evidence review found the data inadequate to assess female reproductive toxicity at all, and inconsistent across high confidence rodent studies for immune and male reproductive effects—gaps RoseyFields’ assessors disclose explicitly whenever a case implicates those endpoints, rather than presenting the existing organ-specific values as a complete characterization of Cr(VI)’s toxicological profile.

Bibliography