BACKGROUND: In the late twentieth century, emergence of high rates of treatment failure with antimonial compounds (SSG) for visceral leishmaniasis (VL) caused a public health crisis in Bihar, India. We hypothesize that exposure to arsenic through drinking contaminated groundwater may be associated with SSG treatment failure due to the development of antimony-resistant parasites. METHODS: A retrospective cohort design was employed, as antimony treatment is no longer in routine use. The study was performed on patients treated with SSG between 2006 and 2010. Outcomes of treatment were assessed through a field questionnaire and treatment failure used as a proxy for parasite resistance. Arsenic exposure was quantified through analysis of 5 water samples from within and surrounding the patient's home. A logistic regression model was used to evaluate the association between arsenic exposure and treatment failure. In a secondary analysis survival curves and Cox regression models were applied to assess the risk of mortality in VL patients exposed to arsenic. RESULTS: One hundred and ten VL patients treated with SSG were analysed. The failure rate with SSG was 59%. Patients with high mean local arsenic level had a non-statistically significant higher risk of treatment failure (OR = 1.78, 95% CI: 0.7-4.6, p = 0.23) than patients using wells with arsenic concentration <10 mug/L. Twenty one patients died in our cohort, 16 directly as a result of VL. Arsenic levels >/= 10 mug/L increased the risk of all-cause (HR 3.27; 95% CI: 1.4-8.1) and VL related (HR 2.65; 95% CI: 0.96-7.65) deaths. This was time dependent: 3 months post VL symptom development, elevated risks of all-cause mortality (HR 8.56; 95% CI: 2.5-29.1) and of VL related mortality (HR 9.27; 95% CI: 1.8-49.0) were detected. DISCUSSION/CONCLUSION: This study indicates a trend towards increased treatment failure in arsenic exposed patients. The limitations of the retrospective study design may have masked a strong association between arsenic exposure and selection for antimonial resistance in the field. The unanticipated strong correlation between arsenic exposure and VL mortality warrants further investigation.