A 3-year longitudinal survey was carried out from 1998 to 2000 in a village in eastern Sudan where a visceral leishmaniasis (VL) outbreak occurred. Leishmania-specific antibodies were analysed by enzyme-linked immunosorbent assay and immunoblotting. Immunoblot analysis detected antibodies to Leishmania in 80% of the healthy subjects and half of them harboured high immunoglobulin (Ig) G antibody levels, similar to those of VL patients. These antibodies belonged to the IgG1 and IgG3 subclasses but neither their respective levels nor the immunoblot recognition patterns were predictive of VL. During this epidemic, a large proportion of subjects had a high antileishmanial antibody response, indicating that they were infected by Leishmania though most of them remained healthy during the whole study period. These results obtained in the context of an outbreak contrast with those obtained from studies performed in endemic areas characterized by lower parasite transmission levels. Furthermore, the clinical and serological follow-up of our study subjects showed that VL occurred mainly in subjects who had been serologically positive for 5-24 months rather than resulting from primo infection by the parasite.