In a study in three neighbouring villages of southern Madagascar, where Schistosoma mansoni is hyperendemic, ultrasound examination using the Niamey protocol showed marked differences in the burden of disease from one village to another. Hepatosplenic schistosomiasis was more frequent in the village with the highest geometric mean egg counts and the earliest onset of infections, demonstrating that the morbidity induced by S. mansoni may vary greatly within a given area. True representativeness of study populations, a keystone of epidemiological studies, is mandatory to obtain a clear picture of a wide area. Ultrasound examinations in a small number of villages, or even a single one, may be a questionable approach. Using logistic regression analysis, the explanatory variables found to be significantly associated with a risk of severe hepatosplenic disease in our study were sex, age, village of residence and S. mansoni egg counts. On the other hand, a concurrent infection with an intestinal helminth seems to reduce the risk of severe hepatosplenic disease. Further studies should assess the role and possible impact of intestinal helminths on S. mansoni associated-morbidity.