Abstract
We conducted a spatial case-control study nested in a dengue incidence cohort to explore the role of the spatial and socioeconomic factors in the proportion of symptomatic (cases) and inapparent primary dengue virus infections (controls). Cohort participants were children and adolescents (2 to 16 years of age) at the beginning of the follow-up. Case definitions were, for symptomatic cases, fever plus a positive lab result for acute dengue (NS1, RT-PCR, ELISA IgM/IgG), and for inapparent infection a positive result for dengue IgG (ELISA) in subjects without symptoms and with a previously negative result at baseline. The covariates included sociodemographic factors, residential location, and socioeconomic context variables of the census tracts of residence of cases and controls. We used principal component analysis to reduce the contextual covariates, with the component values assigned to each one based on their residences. The data were modeled in a Bayesian context, considering the spatial dependence. The final sample consisted of 692 children, 274 cases and 418 controls, from the first year of follow-up (2014-2015). Being male, older age, higher educational level of the head of the family and having a larger number of rooms in the household were associated with a greater chance of presenting dengue symptomatic infection at the individual level. The contextual covariates were not associated with the outcome. Inapparent dengue infection has extensive epidemiological consequences. Relying solely on notifications of symptomatic dengue infections underestimates the number of cases, preserves a silent source of the disease, potentially spreading the virus to unaffected areas.