An Epidemiologic Study of Arsenic-Related Skin Disorders and Skin Cancer and the Consumption of Arsenic-Contaminated Well Waters in Huhhot, Inner Mongolia, China

• Published in: Human and ecological risk assessment 2007-07, Vol.13 (4), p.713-746
• Main Authors: Lamm, Steven H., Luo, Zhen-Dong, Bo, Fu-Bao, Zhang, Ge-You, Zhang, Ye-Min, Wilson, Richard, Byrd, Daniel M., Lai, Shenghan, Li, Feng-Xiao, Polkanov, Michael, Tong, Ying, Loo, Lian, Tucker, Stephen B., the Inner Mongolia Cooperative Arsenic Project (IMCAP)
• Format: Article
• Language: English
• Subjects: Arsenic; arsenic-related skin effects; Drinking water; Epidemiology; Inner Mongolia; Mortality; Skin cancer; skin cancer risk; threshold (hockey-stick) model; Wells
• ISSN: 1080-7039; 1549-7860
• DOI: 10.1080/10807030701456528

Well-use histories were obtained and dermatological examinations were conducted for 3,179 of the 3,228 (98.5%) residents of 3 villages in Inner Mongolia with well water arsenic levels as high as 2,000 ppb (ug/L). Eight persons were found to have skin cancer, 172 had hyperkeratoses, 121 had dyspigmentation, 94 had both hyperkeratoses and dyspigmentation, and, strikingly, none had Blackfoot disease. All 8 subjects with skin cancer also had both hyperkeratoses and dyspigmentation. Arsenic levels were measured for 184 wells and individual well-use histories were obtained. Arsenic exposure histories were summarized as both highest arsenic concentration (highest exposure level for at least 1-year duration) and cumulative arsenic exposure (ppb-years). Sixty-nine percent of the participants had highest arsenic concentrations below 100 ppb; 71% had cumulative arsenic exposures below 2,000 ppb-years. Exposure-response analyses included frequency-weighted, simple linear regression, and most-likely estimate (hockey-stick) models. Skin cancer cases were only found for those with a highest arsenic concentration greater than 150 ppb, and those with exposure less than 150 ppb had a statistically significant deficit. A frequency-weighted model showed a threshold at 150 ppb, and a hockey-stick model showed a threshold at 122 ppb. Considerations of duration, age, latency, and misclassification did not appear to markedly affect the analysis. The non-malignant skin findings showed thresholds of 40-50 ppb in the hockey-stick models. Application of these analytic models to the data from other epidemiological studies of arsenic ingestion and malignant and non-malignant skin disorders can be used to examine patterns of arsenic carcinogenicity.