To determine the time-lag effect of meteorological factors on the relative risk (RR) of dengue incidence in Coronel Fabriciano city, Brazil, we applied a distributed lag nonlinear model, a modeling framework that can simultaneously represent nonlinear exposure-response dependencies and delayed effects, to establish the association between dengue incidence and weather predictors. The weekly number of notified dengue cases during the period 2004-2010 was used for analysis. When considering the rainfall, the highest RR (1.2) was observed for lag 10. Observing the cumulative effect of the precipitation, the RR for 12th and 13th week was RR = 4. The highest risk, 1.25, was observed at 25 °C, denoting that the risk of dengue transmission increases with temperature. Climate-based models that take into account the time lag between rainfall, temperature, and dengue can be useful in dengue control programs to be applied in tropical countries.
Keywords: dengue fever; lag nonlinear model (DLNM); precipitation; temperature.