Epidemiological basis for clinical diagnosis of childhood malaria in endemic zone in West Africa.

It is difficult to distinguish childhood malaria from other common febrile disorders by parasite count alone, because of the wide variation in tolerance of parasitaemia among individuals. We postulated that the proportion of febrile episodes among young children that can be attributed to parasitaemia varies according to simple clinical criteria. We studied 1114 children aged 2-9 years, who attended a dispensary in the Republic of Niger, with a case-control approach; each of 557 febrile children was matched with a non-febrile control by sex, age, ethnic group, and day of presentation. Febrile episodes were classified according to three clinical criteria: the presence of a likely non-malarial cause; the duration (less than 3 or more than 3 days before presentation); and the intensity (below 39 degrees C or 39 degrees C and above). There was no evidence for an association between febrile episodes and parasite count during the dry, low-transmission, season. During the rainy, high-transmission, season, by contrast, there was a highly significant relation (p less than 0.0001) between the likelihood of fever and the parasite count; each clinical criterion strengthened the association. There was no association between parasitaemia and low intensity fevers, with an obvious cause, that started 3 or more days before presentation, even in the rainy season; however, the relative risk of a fever that met all three criteria developing in those with vs those without parasitaemia was 27.5. The proportion of febrile cases attributable to detectable parasitaemia (population attributable risk) ranged from 0 to 0.92. Our results suggest that simple clinical criteria may be valuable in the selection of febrile patients for antimalarial treatment. In this geographic area, high fever of short duration and with no other obvious cause that occurs during the rainy season is most likely to be malaria.