Analysis of human risk assessment due to arsenic exposure through cow milk using multivariate and Monte Carlo simulation technique: a case study
摘要
This study assessed arsenic (As) contamination in cow milk across eight locations (A–H) in Sahibganj, Jharkhand, India, where the average As concentration in livestock drinking water was 0.44 mg/L. Milk As levels ranged from BDL (Location H, analyzed as zero) to 0.026 mg/kg (Location A), showing strong correlations with water (r = 0.983, p < 0.05) and fodder (r = 0.970, p < 0.05) concentrations. Principal component analysis (PCA) revealed that the first three principal components (PC) collectively explained 87.53% of total variance, effectively capturing the major trends in the data. PC1, explaining 65.97% of the variance, was primarily loaded with high positive contributions from arsenic in water (0.392), fodder (0.403), and soil (0.410). Deterministic and probabilistic risk assessment, performed using a Monte Carlo simulation, identified significant health threats. Deterministic non-carcinogenic risks (HQ) peaked at Location A for age groups 0–3 and 3–12 years (HQ > 1), while the other locations generally showed lower HQ values across all age groups. The probabilistic 95th percentile HQ also exceeded 1 for age groups 0–3 and 3–12 at Location A and for the 0–3 years age group at Locations B and C. Similarly, the deterministic carcinogenic risk (CR) reached 8.69E-04 for the 0–3 years age group at Location A, exceeding the USEPA limit (E-04–E-06), with the probabilistic 95th percentile at 1.15E-03; the lowest CR was 4.77E-06 at Location E for the 36–60 age group. These findings underscore a critical public health concern, particularly for vulnerable populations, necessitating urgent mitigation strategies to reduce As exposure through milk consumption in the region.