Arsenic in drinking water causes a widespread concern in Bangladesh, where a major proportion of tube wells is contaminated. Arsenic ingestion causes skin lesions, which is considered as definite exposure. A prevalence comparison study of respiratory effects among subjects with and without arsenic exposure through drinking water was conducted in Bangladesh. Exposed participants were recruited through health awareness campaign programs. Unexposed participants were randomly selected, where tubewells were not contaminated with arsenic. A total of 169 individuals participated (44 exposed individuals exhibiting skin lesions; 125 unexposed individuals). The arsenic concentrations ranged from 136 to 1000 micro g l(-1). The information regarding respiratory system signs and symptoms were also collected and the analyses were confined to nonsmokers. The crude prevalence ratio for chronic bronchitis and chronic cough amounted to 2.1 (95% CI 0.7-6.1). The prevalence ratios for chronic bronchitis increased with increasing exposure, i.e., 1.0, 1.6, 2.7 and 2.6 using unexposed as the reference. The prevalence ratios for chronic cough were 1.0, 1.6, 2.7 and 2.6 for the exposure categories, using the same unexposed as the reference. The dose-response trend was the same (P < 0.1) for both conditions. These results add to evidence that long-term ingestion of arsenic exposure can cause respiratory effects.