Clinical Pharmacist and Master’s student in Clinical Pharmacy with research interests in pharmacovigilance, behavioral interventions in mental health, and AI applications in clinical decision support. Experience includes digital health research with Bloomsbury Health (London) and pharmacovigilance practice in patient support programs. Published work covers drug awareness among healthcare providers, postpartum depression management, and patient safety reporting. What was reviewed?
This scientific opinion presents an updated inorganic arsenic risk assessment that synthesizes toxicokinetics, biomarkers, genotoxicity, epidemiology, dose–response modeling, and risk characterization for inorganic arsenic (iAs) in food and water within the EU context. It integrates EFSA’s 2021 exposure assessment with a new hazard appraisal, identifies a human-derived reference point, and applies a margin-of-exposure (MOE) framework rather than setting a health-based guidance value because iAs is a genotoxic carcinogen. According to the flowchart on page 14, the review systematically screens studies, converts epidemiologic risk estimates to quantal data, performs benchmark dose modeling, and selects a reference point for risk characterization. The diagram on page 16 depicts interlinked iAs metabolic pathways, highlighting methylation steps that influence toxicity.
Who was reviewed?
The review evaluates human epidemiological evidence emphasizing low to moderate exposure populations, typically with drinking-water arsenic <~150 µg/L or equivalent biomarker-based intakes, and converts water or urinary iAs measures into total dietary exposures (µg/kg bw/day). It focuses on adult cohorts and case–control studies from the United States, Europe, and Asia, with specific consideration of pregnancy and in-utero exposure for developmental endpoints. While children often exhibit higher dietary exposures per body weight, risk characterization is anchored to adult outcome data, noting that adults in these cohorts likely had higher early-life exposures as well.
Most important findings
| Critical points for HTMC | Details relevant to certification and regulation |
|---|---|
| Human reference point (RP) | Benchmark dose lower confidence limit at 5% response (BMDL05) = 0.06 µg iAs/kg bw/day, derived from a high-quality US skin cancer case–control study using urinary total inorganic arsenic; a second European study supports a skin-cancer–based RP. This RP was judged to also cover lung and bladder cancers, skin lesions, ischemic heart disease, chronic kidney disease, respiratory disease, certain reproductive outcomes, infant mortality, and neurodevelopmental effects. |
| Risk characterization approach | Because iAs is a genotoxic carcinogen with additional epigenetic effects, EFSA applies an MOE approach rather than setting a tolerable intake. There is no EFSA precedent for an MOE of low concern when using human cancer data; therefore, no generic “safe” MOE is defined. |
| MOE vs. EU exposure | Adult dietary exposure mean: 0.03–0.15 µg/kg bw/day; P95: 0.07–0.33 µg/kg bw/day. Against the RP (0.06), MOEs ≈2–0.4 (means) and 0.9–0.2 (P95), indicating health concern; an MOE of 1 corresponds to ~5% relative increase in skin cancer risk above background. Children’s MOEs are smaller due to higher exposure per kg. |
| Causally supported endpoints | Sufficient, causal associations at low–moderate exposure for skin, bladder, and lung cancers; skin lesions; ischemic heart disease; carotid atherosclerosis; chronic kidney disease; respiratory disease; certain developmental and perinatal outcomes; neurodevelopmental effects; inconsistent or insufficient evidence for several other cancers and metabolic endpoints. |
| Exposure contributors | Main contributors across ages: rice and rice-based products, other grains, and drinking water. For infants/young children, infant cereals and biscuits materially contribute. |
| Biomarkers & measurement | Preferred biomarker for intake is urinary total inorganic arsenic (u-tiAs) (sum of iAs, MMA, DMA), with seafood restrictions pre-sampling; hair and nails inform longer-term exposure but are imperfect. Analytical speciation typically uses HPLC-ICP-MS; method details and interferences are discussed. |
| EU maximum levels (MLs) | Regulation (EU) 2023/915 sets MLs for iAs in rice categories (e.g., 0.15 mg/kg for non-parboiled milled rice; 0.10 mg/kg for rice for infant foods), rice-based drinks, infant formula, baby food, and juices; drinking water limit remains 10 µg/L. |
| Mechanisms & susceptibility | Conditional uncertainty analysis indicates that mean exposures may exceed modeled BMDs with probabilities ranging from unlikely (~0.17) to likely (~0.86), depending on study; EFSA calls for guidance on BMD modeling with epidemiology and on quantitative risk assessment for genotoxic carcinogens. |
| Uncertainty highlights | Conditional uncertainty analysis indicates that mean exposures may exceed modeled BMDs with probabilities ranging from unlikely (~0.17) to likely (~0.86), depending on the study; EFSA calls for guidance on BMD modeling with epidemiology and on quantitative risk assessment for genotoxic carcinogens. |
Key implications
For regulators, the inorganic arsenic risk assessment supports tightening or maintaining strict MLs for rice categories and enforcing water at 10 µg/L while using MOE to flag concern. Certification programs should require validated speciation methods and batch-level iAs verification near the RP-aligned MOE. Industry can apply rice sourcing, polishing, and rinsing/cooking controls to lower iAs, with infant foods prioritized. Research gaps include epigenetic causality and early-life susceptibility. Practically, standardize u-tiAs controls, seafood washout before biomonitoring, and transparent MOE reporting.
Citation
EFSA Panel on Contaminants in the Food Chain (CONTAM). (2024). Update of the risk assessment of inorganic arsenic in food. EFSA Journal, 22, e8488.
