Background: Leishmaniasis is a vector-borne disease spread by female phlebotomine sandflies. The most severe form of the disease is visceral leishmaniasis (VL), which causes swelling of internal organs. Over 97% of human VL cases in the Americas are within Brazil, where the average mortality rate is 7% in treated cases, and can reach 90% in untreated cases. Domestic canines are the main reservoir for human cases in Brazil because they live in close proximity and can remain asymptomatic for long periods of time. Due to this zoonotic transmission, sole treatment of human cases will not contain the spread of the disease. Current methods of control have been unsuccessful, and thus a better understanding of the transmission to dogs and the effect of their
environment is required.
Methods & Results: This study analysed the data of 777 domestic canines in
Camaçari, Bahia state, Northeastern Brazil, collected in 2011-12, to investigate the
association between the presence of vegetation and the risk of canine visceral leishmaniasis. The level of vegetation was indexed using Landsat 7 satellite imagery through two methods: Normalised Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI). Generalised linear modelling determined that whether the dog lived nearthe coast and whether they lived with other infected dogs were significant to the odds of disease. An empirical variogram detected residual spatial correlation when the data was modelled with a generalised linear mixed model. This was addressed through a generalised linear geostatisitcal model (GLGM). The Laplace method was used to estimate the parameters, which were used as starting values in the Monte Carlo method.
The GLGM determined that an increase in vegetation of 20% would lead to an increase in the odds of canine visceral leishmaniasis of 18.9%. The model fit was verified by an empirical variogram.