Assessment of health risk from exposure to arsenic in selected arsenic affected areas of Bangladesh

Abstract

Estimation of population exposed to different levels of As in drinking water has been made for 21 As-affected upazilas covered under the BAWMSP survey. In the highly affected 7 upazilas (80-100% contaminated wells), about 64% of population is exposed to As concentration above the Bangladesh drinking water standard of 50 ppb. In the 7 upazilas with 60-80% contamination, population exposed to unsafe level of As is about 55%, while in the 7 upazilas with 40-60% contamination, it is about 37%. Overall, about 54.7% of the population in the 21 upazilas is exposed to unsafe level of As in their drinking water. Population exposed to As concentration above the WHO guideline value of 10 ppb could not be estimated from the survey data, since the lowest reported As concentration range in the BAMWSP database is 0-50 ppb. In the As affected areas, percentage of population exposed to unsafe level of As (i.e. > 50 ppb) is somewhat less than the percentage of contaminated wells, possibly suggesting that some people are using As-safe water from sources other than their contaminated wells. For example, in the 7 upazilas with 80-100% contaminated wells, the average percentage of contaminated wells is about 87.9%, while percentage of population exposed to contaminated water (i.e. As> 50 ppb) is about 63.5%. Model predictions of the number of arsenicosis patients using two sets of model parameters did not match the actual patient data. Parameters of Yu et al. (2003) overpredicted the total arsenicosis patients by a factor of over 70; parameters of Ahmed (2003) under-predicted patient number among population exposed to relatively low level of arsenic (below 100 ppb) and significantly over-predicted patient number among population exposed to relatively high concentrations of As. The model parameters of Yu et al. (2003) were derived from a study in West Bengal; which suggests that if health effects in Bangladesh eventually become similar to those experienced in West Bengal, a huge number of people would be come affected with arsenicosis. Model parameters estimated in this study also failed to match the actual data. This is not surprising because the survey data show that prevalence of arsenicosis does not correlate well with As concentration, while the health risk model assumes prevalence to increase with increasing As in drinking water. Survey data show that the actual prevalence ratio among population exposed to relatively low As concentration (e.g., up to 100 ppb) is quite high; while the prevalence ratio appears to show a decreasing trend as As concentration exceeds 500 ppb, especially among female population. It appears that in the As affected areas, drinking As safe water (As < 50 ppb) does not guarantee safety from adverse health effects. For example, among a total of 2832 arsenicosis patients identified in the 21 upazilas, about 43% are drinking water with As concentration below 50 ppb. Other parameters, e.g., arsenic exposure through food chain, food habit and nutrition, genetic makeup probably have significant influence on the prevalence of arsenic. However, the data gathered during the BAMWSP survey do not allow analysis of such parameters. Many could however question the reliability of As measurements made during the BAMWSP survey using field kits. Along with efforts to better understand the health effects of arsenic, efforts should also be made to develop better models for predicting long-term health effect of arsenic in Bangladesh and in other countries.