A community-based study of subclinical flavivirus infections in children in an area of Tamil Nadu, India, where Japanese encephalitis is endemic.

A characteristic feature of the epidemiology of Japanese encephalitis (JE) is the occurrence of a large number of subclinical infections. The reporting of only overt cases underestimates the total level of virus transmission, a knowledge of which is essential for the evolution of control strategies. We carried out a 3-year prospective serological study between 1989 and 1991 in a primary health centre in Tamil Nadu where JE is endemic. Each year paired specimens, taken before and after the transmission season from a cohort of schoolchildren aged 5-9 years, were tested for haemagglutination inhibition (HI) antibody titres in order to study seroconversion. The seroconversion rates in the successive years were 37.5, 42.1 and 25 percentage points, and in a third of such seroconversions it was possible to establish a specific diagnosis. Seroconversion was attributable predominantly to JE virus and minimally to West Nile virus. Relatively high dengue virus activity occurred only in 1991. There were statistically significant differences in seroconversion rates between villages and this was related to variations in the ratio cattle:humans:pigs. Very high seroconversion rates occurred among children who were negative for HI antibodies before the transmission season. HI antibodies declined to undetectable levels 6-8 months later in half the children who had seroconverted. The average net annual increase of 16.2 percentage points in seropositivity was nevertheless much higher than values reported from other areas of endemicity. The overall incidence of JE cases was 15 per 10,000 children aged 5-9 years, and the estimated ratio overt:inapparent infection was 1:270.