Between September 2003 and April 2004, the supply of antimonial drugs to Amudat Hospital, in north-eastern Uganda, was interrupted and all cases of visceral leishmaniasis presenting at the hospital could only be treated with amphotericin B deoxycholate (AmB). This allowed the safety and effectiveness of the AmB to be evaluated, in comparison with an historical cohort of patients treated, at the same hospital, with meglumine antimoniate (Sb(V)). Demographic and clinical data were collected before and after treatment. Adverse effects were recorded passively in all the subjects, and actively, using a standardized questionnaire, in a sub-group of the patients given AmB. The in hospital case-fatality 'rates' were 4.8% [95% confidence interval (CI) = 2.4%-8.8%] among the 210 patients treated with AmB and 3.7% (CI = 1.4%-7.9%) among the 161 patients treated with Sb(V) (P>0.20). Adverse effects requiring treatment interruption were rare in both cohorts. Treatment failures (i.e. non-responses or relapses) were observed in 2.9% (CI = 1.2%-6.4%) of the patients treated with AmB and 1.2% (CI = 0.1%-4.4%) of the patients treated with Sb(V) (P>0.20). For the treatment of visceral leishmaniasis in Uganda, AmB therefore had a similar effectiveness and safety profile to that of meglumine antimoniate.