OBJECTIVE: Nepal reports a visceral leishmaniasis (VL) incidence of 5 per 10 000 per year on the basis of notification by health facilities, but little community-based epidemiological information exists. We report data on prevalence rates of Leishmania donovani infection in ten communities in East Nepal. METHODS: Ten clusters with highest VL incidence rates were purposefully selected in Nepal. All households were mapped and socio-demographic data and data on past VL incidence were collected. An exhaustive serological survey was performed of individuals aged >2 years, by collecting finger prick blood on filter paper in November-December 2006. The samples were tested by direct agglutination, and a titre >or=1:1600 was taken as marker of infection. A generalized estimating equation (GEE) model was used to assess risk factors for Direct Agglutination Test (DAT) positivity taking into account the clustering at household and village level. RESULTS: The sero-survey (n = 5397) showed an infection prevalence rate of 9%, (range 5-15% per cluster) with higher prevalence in men (9.9%) than in women (8.3%) (P = 0.049). Male gender, increasing age and poverty were significant risk factors in the final GEE model. CONCLUSION: Leishmania infection rate in high-transmission areas in Nepal is associated with gender, age and socio-economic status.