During an epidemic of visceral leishmaniasis (VL) in eastern Sudan, Médecins Sans Frontières distributed 357,000 insecticide-treated bednets (ITN) to 155 affected villages between May 1999 and March 2001. To estimate the protective effect of the ITN, we evaluated coverage and use of ITN, and analysed VL incidence by village from March 1996 to June 2002. We provided ITN to 94% of the individuals >5 years old. Two years later, 44% (95% CI 39-48%) of nets were reasonably intact. Because ITN were mainly used as protection against nuisance mosquitoes, bednet use during the VL transmission season ranged from <10% during the hot dry months to 55% during the beginning of the rainy season. ITN were put up from 9 to 11 p.m., leaving children unprotected during a significant period of sandfly-biting hours after sunset. Regression analysis of incidence data from 114 villages demonstrated a significant reduction of VL by village and month following ITN provision. The greatest effect was 17-20 months post-intervention, with VL cases reduced by 59% (95% CI: 25-78%). An estimated 1060 VL cases were prevented between June 1999 and January 2001, a mean protective effect of 27%. Although results need to be interpreted with caution, this analysis indicates a potentially strong reduction in VL incidence following a community distribution of ITN. The effectiveness of ITN depends on behavioural factors, which differ between communities.