Geostatistical modelling of the effect of vegetation on canine visceral leishmaniasis in Camaçari, Bahia State, Northeastern Brazil using remote-sensed data

Abstract Visceral leishmaniasis (VL) is a vector-borne disease that can cause fever, hepatosplenomegaly, weight loss and pancytopenia. In Brazil, VL occurs when a female phlebotomine sand fly ( Lutzomyia longipalpis ) takes a blood meal from an infected reservoir. Domestic canines are likely the main source of the infectious agent since they live in close proximity to humans and can remain asymptomatic for several years. Consequently, sole treatment of human cases likely cannot prevent transmission. Current methods of control have been unsuccessful, and thus, a better understanding of canine transmission and the effect of the environment is required. Phlebotomine sand flies feed on plant sap and nectar, and lay eggs around tree roots, hence, we hypothesise that dogs are more likely to contract VL near vegetation. In this retrospective study, we investigate the effect of vegetation on canine visceral leishmaniasis (CVL) in the community of Camaçari, Bahia, Northeastern Brazil using geostatistical modelling. The normalised vegetation index (NDVI) is used to quantify vegetation. We find that, on average, an increase in NDVI of 0.1 led to a 1.21-fold increase in the odds of CVL. The odds of CVL were highest along the coast where plant species known to attract Lutzomyia longipalpis have been identified.