Background: To evaluate the effect of insecticide spraying for vector control and elimination of infected dogs on the incidence of human infection with L. infantum, a randomized community intervention trial was carried out in the city of Teresina, Brazil.
Methods/principal findings: Within each of ten localities in the city, four blocks were selected and randomized to 4 interventions: 1) spraying houses and animal pens with insecticide; 2) eliminating infected dogs; 3) combination of spraying and eliminating dogs, and 4) nothing. The main outcome is the incidence of infection assessed by the conversion of the Montenegro skin test (MST) after 18 months of follow-up in residents aged ≥ 1 year with no previous history of visceral leishmaniasis (VL). Reactions were measured at 48-72 h, induration of ≥ 5 mm considered positive. Interventions were executed after the baseline interview and repeated 6 and 12 months later. The effects of each type of intervention scheme on the incidence of infection were assessed by calculating relative risks and 95% confidence intervals using Poisson population-averaged regression models with robust variance. Among the 1105 participants, 408 (37%) were MST positive at baseline. Of the 697 negatives, only 423 (61%) were reexamined at the end of the follow-up; 151 (36%) of them converted to a positive MST. Only dog culling had some statistically significant effect on reducing the incidence of infection, with estimates of effectiveness varying between 27% and 52%, depending on the type of analysis performed.
Conclusions/significance: In light of the continuous spread of VL in Brazil despite the large scale deployment of insecticide spraying and dog culling, the relatively low to moderate effectiveness of dog culling and the non-significant effect of insecticide spraying on the incidence of human infection, we conclude that there is an urgent need for revision of the Brazilian VL control program.