Dai et al. 2024 — Dietary arsenic exposure and health risk assessment in Chongqing, China (2018-2023)
This surveillance and risk-assessment study reports total arsenic (tAs) concentrations across 4,900 food samples collected in Chongqing municipality from 2018 to 2023, and combines these with dietary consumption data from 969 Chongqing residents to estimate population-level dietary arsenic exposure by age, sex, and place of residence. The study applies a two-dimensional Monte Carlo simulation following FAO/WHO probabilistic assessment methodology and uses the margin-of-exposure (MOE) approach to characterize non-carcinogenic risk (skin damage) and carcinogenic risk (lung cancer, urinary system cancers). Rice and rice products emerge as the dominant arsenic source, accounting for 69.97 to 74.37 percent of total dietary arsenic exposure, and all age groups show MOE values below 100 for carcinogenic endpoints, indicating potential elevated cancer risk with the highest concern in young children aged 3 to 6 years.
Key numbers
All concentrations are total arsenic (tAs) in mg/kg wet weight by ICP-MS (GB 5009.268-2016) or hydride generation atomic fluorescence (GB 5009.11-2014). Values are reported as median bound (MB, using 1/2 LOD for non-detects) and upper bound (UB, using LOD for non-detects).
Overall food: detection rate 36.40%; mean tAs 0.0207 (MB) to 0.0234 (UB) mg/kg; P95 0.0673 (MB) to 0.0674 (UB) mg/kg; n=4,900.
Per-category summary (mean, MB / UB mg/kg; P95 MB/UB mg/kg; n):
- Rice and rice products: mean 0.0981 / 0.0985; P95 0.1833 / 0.1833; detection rate 89.87%; n=464
- Special dietary foods: mean 0.0603 / 0.0614; P95 0.1500 / 0.1500; detection rate 75.78%; n=450
- Other grains: mean 0.0507 / 0.0521; P95 0.2100 / 0.2100; detection rate 71.43%; n=91
- Grains and grain products (combined): mean 0.0624 / 0.0637; P95 0.1725 / 0.1725; detection rate 64.79%; n=869
- Nuts and seeds: mean 0.0261 / 0.0283; P95 0.1176 / 0.1176; detection rate 52.17%; n=46
- Beans and bean products: mean 0.0113 / 0.0139; P95 0.0593 / 0.0593; detection rate 47.78%; n=91
- Aquatic animals and their products: mean 0.0223 / 0.0251; P95 0.0947 / 0.0947; detection rate 34.72%; n=72
- Leafy vegetables: mean 0.0159 / 0.0177; P95 0.0650 / 0.0650; detection rate 50.13%; n=393
- Eggs and egg products: mean 0.0164 / 0.0186; P95 0.0677 / 0.0677; detection rate 24.91%; n=530
- Wheat and wheat products: mean 0.0131 / 0.0156; P95 0.0711 / 0.0711; detection rate 33.12%; n=314
- Vegetables overall: mean 0.0119 / 0.0139; P95 0.0420 / 0.0420; detection rate 35.36%; n=1,383
- Meat and meat products: mean 0.0079 / 0.0108; P95 0.0152 / 0.0152; detection rate 18.22%; n=807
- Fruits and fruit products: mean 0.0080 / 0.0109; P95 0.0295 / 0.0295; detection rate 18.81%; n=489
- Milk and dairy products: mean 0.0042 / 0.0073; P95 0.0080 / 0.0100; detection rate 18.26%; n=115 — lowest across all categories
Dietary inorganic arsenic (iAs) exposure estimates are modeled, not directly measured: the paper converts tAs to iAs using EFSA 2014 / WHO 2011 food-category-specific iAs:tAs ratios (e.g., rice/grains 84%, meat 100%, aquatic products 2%, dairy 26%). The exposure values below are iAs equivalents, not directly speciated measurements.
Estimated daily iAs intake (two-dimensional Monte Carlo, MB/UB µg/kg/day):
- Children aged 3-6: mean 1.046 / 1.116; P95 2.789 / 2.875 — highest of all age groups
- Adult males (mean): 0.412 / 0.467; P95 1.115 / 1.153
- Adult females (mean): 0.410 / 0.441; P95 0.999 / 1.040
- Rural residents (mean): 0.525 / 0.560; P95 1.283 / 1.320
- Urban residents (mean): 0.388 / 0.422; P95 0.956 / 0.981
- Residents aged 60+: lowest mean exposure across age groups
Risk characterization (MOE, JECFA 2011 BMDL endpoints):
- Non-carcinogenic (skin damage, BMDL = 5.4 µg/kg/day): MOE > 1 in all populations — relatively low risk
- Carcinogenic lung cancer (BMDL = 3 µg/kg/day): MOE < 100 in all populations — potential risk present
- Carcinogenic urinary cancer (BMDL = 5.2 µg/kg/day): MOE < 100 in all populations
Rice and rice products contribute 69.97 to 74.37 percent of total dietary arsenic exposure.
Methods (brief)
Analytical method: GB 5009.11-2014 (hydride generation atomic fluorescence spectrometry, HGAFS) and GB 5009.268-2016 (ICP-MS); LOD range 0.0006 to 0.0400 mg/kg by food category. All reported values are total arsenic (tAs); inorganic arsenic is not directly speciated. Conversion to iAs uses published EFSA/WHO iAs:tAs ratios per food category (see Table 1 of the paper). Non-detect handling follows WHO/EFSA substitution method: 1/2 LOD for MB, LOD for UB. Exposure modeling: Latin hypercube sampling, 100,000 iterations; dietary data from three consecutive 24-hour recalls in 969 Chongqing residents (2018 CNHS). Risk characterization via MOE using JECFA 2011 BMDL values.
Important limitation: the paper does not directly measure iAs. The iAs exposure estimates depend on the validity of the EFSA/WHO conversion ratios, which vary by processing method and geography and may not reflect Chongqing-specific speciation patterns. For regulatory or certification purposes, the tAs concentration data in Table 2 are the directly measured values; the iAs estimates should be treated as modeled approximations.
Implications
Certification: Rice and rice products are the dominant dietary arsenic vector in this Chinese urban/rural population, consistent with the broader literature. The tAs P95 of 183 ppb in rice (wet weight) aligns with the range reported in other Chinese occurrence surveys. The modeled iAs:tAs conversion used here (84% for grains per WHO 2011) is a simplification; certification work requiring iAs should use direct HPLC-ICP-MS speciation.
Courses: Illustrates the probabilistic dietary exposure assessment methodology (Monte Carlo) and the distinction between total arsenic measurement and modeled inorganic arsenic — a key pedagogical point for QA professionals.
App: tAs concentrations in Table 2 can inform contamination profile ranges for rice, grains, vegetables, eggs, meat, dairy, aquatic products, and fruits within a Chinese market context; note geographic and market specificity.
Microbiome: Not applicable.