EFSA CONTAM Panel 2024 — Updated risk assessment of inorganic arsenic in food
This 191-page EFSA Scientific Opinion, adopted November 2023, is the first comprehensive update to EFSA’s 2009 iAs risk assessment and integrates the revised 2021 exposure assessment. The CONTAM Panel reviewed new epidemiological evidence and BMD modelling, concluded that iAs is a genotoxic carcinogen, and selected a new Reference Point (RP): BMDL₀₅ of 0.06 µg iAs/kg bw per day from a US skin cancer (squamous cell carcinoma) case-control study, based on a 5% relative increase in background incidence (BMR05). Because iAs is both a threshold and non-threshold carcinogen (genotoxic), the Panel applied a Margin of Exposure (MOE) approach rather than deriving a health-based guidance value. MOEs for adults range from 2 to 0.4 for mean consumers and 0.9 to 0.2 at the 95th percentile — all below an MOE of 1 (the exposure level associated with a 5% increase in background skin cancer incidence), raising a health concern.
Key numbers
- Reference Point (RP): BMDL₀₅ = 0.06 µg iAs/kg bw per day (skin cancer, squamous cell carcinoma; US case-control study; BMR of 5% relative increase in background incidence after adjustment for confounders)
- 20 epidemiological studies met validity criteria for BMD modelling; 6 cohort studies had follow-up of 3–12 years
- Low BMDLs (≤0.17 µg iAs/kg bw/day) calculated for: skin cancer, bladder and lung cancer, respiratory disease, skin lesions, chronic kidney disease, ischemic heart disease
- The RP of 0.06 µg iAs/kg bw/day is within the range of mean adult dietary exposure estimates from EFSA 2021: 0.03–0.15 µg iAs/kg bw/day; and below all 95th percentile adult estimates: 0.07–0.33 µg iAs/kg bw/day
- MOEs for adults (mean exposure): range 2 to 0.4; at 95th percentile: 0.9 to 0.2 (all raise health concern)
- An MOE of 1 would correspond to an exposure level associated with a 5% increase in background skin cancer incidence — and the majority of adult 95th percentile estimates exceed this
- Probability that mean adult exposure exceeds associated BMDs: unlikely (≈0.17) to likely (≈0.86) based on conditional uncertainty analysis considering both skin cancer studies
- EU maximum levels (from Commission Regulation EU 2023/915, Table 2 of this Opinion): non-parboiled milled (white) rice 0.15 mg/kg; parboiled and husked rice 0.25 mg/kg; rice flour 0.25 mg/kg; rice waffles/wafers/crackers/cakes/flakes/puffed breakfast rice 0.30 mg/kg; rice for infant/young children food production 0.10 mg/kg
- Note: 2023/915 values are lower than 2015/1006 values (0.20, 0.25, 0.30, 0.10) for white, parboiled, waffles categories — white rice limit tightened from 0.20 to 0.15 mg/kg
- Previous EFSA 2009 BMDL₀₁: 0.3–8 µg/kg bw/day for 1% extra risk (skin, bladder, lung cancer, skin lesions); JECFA 2011 BMDL₀.₅: 3.0 µg/kg bw/day for lung cancer
- Inapplicability of PTWI: JECFA 2011 withdrew PTWI of 15 µg/kg bw as no longer appropriate given evidence of carcinogenicity at lower doses
- iAs is a genotoxic carcinogen — both thresholded and non-thresholded mechanisms; methylation of iAs to trivalent methylated species (MMA(III), DMA(III)) is an activation process producing more reactive and toxic forms
- Children: MOEs are smaller for children (higher dietary exposure) but the Panel concludes children are covered by the adult risk characterisation because epidemiological evidence is from adults who would also have had higher childhood dietary exposure to iAs
Methods (brief)
Systematic literature review (20 epidemiological studies selected from a broader set meeting validity criteria: statistically significant trend test and/or statistically significant association with iAs as continuous variable; quantile data required for BMD modelling). Exposure data from EFSA 2021 scientific report. BMD modelling using adjusted incidences and number of cases based on adjusted risk ratios from cohort or case-control studies. BMR = 5% relative increase in background incidence after adjustment for confounders. Uncertainty analysis: conditional uncertainty analysis using probability that mean exposure exceeds BMD.
Limitations
Risk characterisation based on relatively large epidemiological studies in which susceptible individuals of higher genetic risk (e.g., differing arsenic biotransformation capacity, DNA repair capacity) may not be adequately represented. Most mechanistic evidence for genotoxicity is from in vitro studies; in vivo evidence is consistent but less extensive. Animal carcinogenicity data cannot serve as primary basis for risk assessment due to major toxicokinetic differences between laboratory animals and humans (particularly methylation capacity). The EU skin cancer epidemiology in the second supporting study (basal cell carcinoma, Hungary/Romania/Slovakia) supports but does not independently derive the RP.
Implications
- Certification: This is the current controlling EFSA opinion for HMT&C’s iAs threshold scientific basis. The shift from BMDL₀₁-based approach (2009) to MOE-based approach using BMDL₀₅ of 0.06 µg/kg bw/day materially changes the risk characterisation framework. The 2023/915 update tightening white rice from 0.20 to 0.15 mg/kg is directly relevant to HMT&C limit-setting. Flag: the new white rice EU ML (0.15 mg/kg) is lower than the 2021 exposure assessment’s reference MLs — this is a regulatory tightening that post-dates the 2021 exposure report.
- Courses: Definitive 2024 authority on iAs risk characterisation in food. The MOE approach, the skin cancer RP, and the genotoxic carcinogen classification are all core teaching content for regulatory affairs and QA teams.
- App: Confirms rice and rice-based products as primary dietary iAs contributors in Europe. The RP of 0.06 µg/kg bw/day and typical adult exposures of 0.03–0.15 µg/kg bw/day mean essentially all adult rice consumers are exposed at or above the RP — a critical point for consumer-facing risk communication.
- Microbiome: Not primary topic.