Incidence of Leishmania donovani infection and Visceral Leishmaniasis (VL) was assessed in a prospective study in Indian and Nepalese high-endemic villages. DAT-seroconversion was used as marker of incident infection in 3 yearly surveys. The study population was followed up to month 30 to identify incident clinical cases. In a cohort of 9034 DAT-negative individuals with neither active signs nor history of VL at baseline, 42 VL cases and 375 asymptomatic seroconversions were recorded in the first year, giving an infection:disease ratio of 8.9 to 1. In the 18 months' follow-up, 7 extra cases of VL were observed in the seroconverters group (N=375), against 14 VL cases among the individuals who had not seroconverted in the first year (N=8570) (RR=11.5(4.5<RR<28.3)). Incident asymptomatic L. donovani infection in VL high-endemic foci in India and Nepal is nine times more frequent than incident VL disease. About 1 in 50 of these new but latent infections led to VL within the next 18 months.