ATSDR 2012 — Toxicological Profile for Cadmium

Summary

This is the September 2012 finalized Toxicological Profile for Cadmium prepared by the US Agency for Toxic Substances and Disease Registry under CERCLA Section 104(i) authority. It supersedes the 2008 draft-for-public-comment version and represents ATSDR’s comprehensive synthesis of the cadmium toxicology literature through approximately 2012. The profile establishes four minimal risk levels for cadmium: an acute-duration inhalation MRL of 0.03 µg Cd/m³, a chronic-duration inhalation MRL of 0.01 µg Cd/m³, an intermediate-duration oral MRL of 0.5 µg Cd/kg/day, and a chronic-duration oral MRL of 0.1 µg/kg/day. The chronic oral MRL is derived from an urinary cadmium benchmark dose lower confidence limit (UCDL10) of 0.5 µg/g creatinine for a 10 percent increase in the prevalence of abnormal low-molecular-weight proteinuria, translated via a toxicokinetic model to a dietary intake of 0.33 µg/kg/day in females at age 55 (peak renal cortex concentration), divided by an uncertainty factor of 3 for human variability. ATSDR’s chronic oral MRL is the tightest of the internationally-circulating cadmium reference values, below both the EFSA TWI and the JECFA PTMI when all are expressed on a daily basis, and below typical US dietary cadmium intake as estimated by the agency itself.

Key numbers

Minimal Risk Levels derived by this profile:

RouteDurationMRLEquivalent
InhalationAcute (≤14 days)3 × 10⁻⁵ mg Cd/m³0.03 µg Cd/m³
InhalationIntermediate (15–364 days)Not derivedWould be lower than chronic and therefore not protective
InhalationChronic (≥1 year)0.01 µg Cd/m³1 × 10⁻⁵ mg Cd/m³
OralAcute (≤14 days)Not derivedUncertainty at the most sensitive endpoint
OralIntermediate (15–364 days)0.5 µg Cd/kg/day5 × 10⁻⁴ mg/kg/day
OralChronic (≥1 year)0.1 µg/kg/day1 × 10⁻⁴ mg/kg/day

Chronic oral MRL derivation parameters:

ParameterValue
Critical endpointRenal tubular dysfunction, biomarkered by urinary low-molecular-weight proteins (β2-microglobulin, pHC, retinol binding protein)
Effect metric10 percent increase in prevalence of abnormal biomarker levels at UCD exposure
UCD10 (dose giving 10 percent increase)1.34 µg Cd/g creatinine (females)
UCDL10 (95% lower confidence limit, point of departure)0.5 µg Cd/g creatinine (females)
Toxicokinetic translationPharmacokinetic model simulating peak renal cortex cadmium at age 55 under constant intake
Dietary intake corresponding to UCDL10, females0.33 µg Cd/kg/day
Dietary intake corresponding to UCDL10, males0.70 µg Cd/kg/day (not used as point of departure; female value is more conservative)
Uncertainty factor3 for human variability (diabetics excluded from several contributing studies)
Final MRL0.33 ÷ 3 ≈ 0.11, rounded to 0.1 µg/kg/day

Intermediate oral MRL derivation:

ParameterValue
Critical studyBrzóska and Moniuszko-Jakoniuk 2005d
Critical endpointDecreased bone mineral density, lumbar spine, in young female rats exposed to cadmium chloride in drinking water for 6 to 12 months
Point of departureBMDLsd1 of 0.05 mg Cd/kg/day (9-month lumbar spine data)
Uncertainty factor100 (10 animal-to-human, 10 human variability)
Final MRL0.05 ÷ 100 = 0.5 µg Cd/kg/day

Acute inhalation MRL derivation:

ParameterValue
Critical studyNTP 1995, F344 rats exposed to cadmium oxide, 6.2 hours/day, 5 days/week for 2 weeks
Critical endpointAlveolar histiocytic infiltrate, focal inflammation, minimal fibrosis in alveolar septa
LOAEL0.088 mg Cd/m³
Duration-adjusted LOAEL0.016 mg Cd/m³
Regional deposited dose ratio (pulmonary)0.617
LOAELHEC0.01 mg Cd/m³
Uncertainty factor300 (10 for LOAEL-to-NOAEL, 3 for animal-to-human with dosimetric adjustments, 10 for human variability)
Final MRL3 × 10⁻⁵ mg Cd/m³ (0.03 µg Cd/m³)

Context for the chronic oral MRL relative to typical US exposure (p. 43, ATSDR’s own framing):

ParameterValue
Chronic oral MRL0.1 µg/kg/day
Age-weighted US dietary cadmium intake (from Choudhury et al. 2001)0.3 µg/kg/day
Geometric mean urinary cadmium, US adults ≥20 years, CDC 20110.247 µg/g creatinine
UCDL10 point of departure0.5 µg/g creatinine

ATSDR acknowledges that its chronic oral MRL is below typical US dietary cadmium intake, and that the UCDL10 point of departure is approximately twofold above the CDC 2011 US adult geometric mean urinary cadmium concentration.

Methods (brief)

ATSDR’s Minimal Risk Level methodology under CERCLA follows standardized ATSDR guidance: identification of sensitive endpoints by route and duration, selection of the most appropriate point of departure from the peer-reviewed literature, application of uncertainty factors per the ATSDR Minimal Risk Level framework, and peer review by a nongovernmental panel. The 2012 profile reflects peer review by a panel and made available for public review prior to finalization.

The chronic oral MRL derivation is notable for using a meta-analysis of environmental exposure studies rather than a single pivotal study. ATSDR synthesized dose-response relationships between urinary cadmium and the prevalence of abnormal renal biomarker levels across multiple large-scale environmental exposure studies, treated the UCD as the exposure metric, and computed UCD10 and UCDL10 values at the population level. The translation from urinary cadmium to dietary cadmium intake used a pharmacokinetic model simulating lifetime constant intake and peak renal cortex concentration at age 55. The uncertainty factor of 3 rather than 10 for human variability reflects ATSDR’s judgment that the environmental-exposure study base already represents the general population including many sensitive subgroups; the factor of 3 was retained specifically to address the exclusion of diabetics from several of the contributing studies, because diabetics are hypothesized to be especially sensitive to cadmium renal toxicity.

Limitations the profile itself acknowledges: the cut-off values used to define “abnormal” urinary biomarker levels vary widely across environmental exposure studies (for beta-2-microglobulin, from 283 to 1,129 µg/g creatinine), which introduces noise into the dose-response synthesis. Several biomarker-increase studies do not provide the quantitative dose-response data needed for MRL derivation. The pharmacokinetic model used for urinary-to-dietary translation is itself a model with its own uncertainties, particularly around inter-individual variability in absorption and half-life.

Implications

  • Certification: ATSDR’s chronic oral MRL of 0.1 µg Cd/kg/day is the tightest major reference value for cadmium and is below typical US dietary cadmium intake. This is important for HMT&C’s defensibility story: an HMT&C threshold calibrated to ATSDR is more conservative than one calibrated to EFSA TWI or JECFA PTMI, and the calibration-to-ATSDR rationale is “US public health agency reference value, most conservative of the options, explicitly acknowledged by the agency to be below typical US intake, most defensible for a certification program asserting population-level health protection.” A calibration to EFSA or JECFA instead would require the rationale-for-deviation statement per the CLAUDE.md firewall rule.
  • Courses: the ATSDR methodology (meta-analysis of environmental exposure studies, UCDL10 as point of departure, pharmacokinetic translation to dietary intake) is a different procedural anchor than either EFSA CONTAM or EPA IRIS; the three-way contrast is a teachable pattern.
  • App: the ATSDR chronic oral MRL gives the app an additional per-ingredient exposure threshold to evaluate against, potentially tighter than the EFSA TWI. The profile’s own observation that its MRL is below typical US dietary intake (~3x lower) is also material to consumer-facing risk communication, which should not describe typical US dietary cadmium as “safe” but rather as exceeding the public-health agency’s derived risk threshold.
  • Microbiome: the profile touches on gut microbiome effects briefly but does not focus on them; dedicated ingest needed for microbiome-specific synthesis.

Provenance notes

License us-government-work. Document is a federal agency publication under 17 U.S.C. § 105 and freely redistributable. The profile supersedes the September 2008 draft-for-public-comment version; any prior references to the 2008 draft for cadmium should be treated as superseded. SHA-256 provenance recorded. The canonical access URL is ATSDR’s tp5.pdf, which is the numbered profile file for cadmium in ATSDR’s series.

Wiki pages updated on ingest