<b>OBJECTIVEb>Analyses of bladder cancer mortality in the Black Foot Disease (BFD) endemic area of southwest Taiwan conducted by Morales et al. showed a discontinuity in risk at 400 microg/L arsenic in the drinking water in a stratified analysis and no discontinuity in a continuous analysis. As the continuous analysis presentation had been used by both the NRC and the EPA to assess the carcinogenic risk from arsenic ingestion, an explanation of the discontinuity was sought.

<b>METHODSb>Review of 40 years of published health studies of the BFD-endemic area of SW Taiwan showed that earlier publications had limited their cancer associations with arsenic levels in artesian well waters and that the reports of Morales et al., NRC, and EPA failed to do so. Underlying data for the Morales et al. study were obtained from the appendix to the NRC report. Bladder cancer mortality rates were calculated from case counts and person-years of observation for each study village. Villages were categorized by water source according to the descriptions from the underlying study. Graphic and regression analyses were conducted of the bladder cancer mortality rates using exposure as a continuous variable and simultaneously stratifying by water source.

<b>RESULTSb>The median village well arsenic levels ranged from 350 to 934 microg/L for villages solely dependent on artesian well water and from 10 to 717 microg/L for villages not solely dependent on artesian well water. Bladder cancer mortality rates were found to be dependent upon the arsenic level only for those villages that were solely dependent on artesian well water for their water source. Bladder cancer mortality rates were found to be independent of arsenic level for villages with non-artesian well water sources.

<b>CONCLUSIONSb>The data indicate that arsenic exposure levels do not explain the bladder cancer mortality risk in SW Taiwan among villages not dependent upon artesian well water. The association for villages dependent upon artesian well water may be explained either by arsenic acting as a high-dose carcinogen or in artesian well water as a co-carcinogen with some other aspect of artesian well water (possibly humic acid). Arsenic exposure level alone appears to be an insufficient exposure measure to describe the risk of bladder cancer mortality in the BFD-endemic area. Risk analyses that fail to take water source into account are likely to misrepresent the risk characterization, particularly at low arsenic levels.

Adolescent ; Adult ; Aged ; Arsenic Poisoning ; complications ; epidemiology ; Child ; Child, Preschool ; Death Certificates ; Environmental Exposure ; Female ; Humans ; Infant ; Infant, Newborn ; Male ; Middle Aged ; Mortality ; trends ; Retrospective Studies ; Risk Assessment ; Taiwan ; epidemiology ; Urinary Bladder Neoplasms ; etiology ; mortality ; Water Supply