The effects of landscape changes on the Brazilian spotted fever incidence in the Atlantic Forest over the last two decades

Abstract

Zoonoses have become an increasing area of concern since the last decades. Landscape changes increase the spatial overlap between reservoir hosts and human habitats, thereby raising the risk of effective spillover. It is crucial to monitor the emergence patterns of zoonoses to reduce their incidence, which can be done in light of environmental epidemiology. The Brazilian spotted fever (BSF) is a vector-born disease transmitted by ticks of the Amblyomma genus and whose etiological agents are Rickettsia bacteria. Despite low incidence of the disease, it has a high annual lethality rate. Moreover, the BSF sylvatic cycle involves amplifier hosts that can raise up pathogen prevalence in the environment. The Atlantic Forest is a tropical biome that has undergone several landscape changes, especially deforestation for pasture and agriculture expansion, and is known to contain most of the BSF incidence. Hence, this thesis aims at characterizing the effects of landscape changes on the BSF incidence, and in particular, at assessing which landscape features promote high incidence levels. Several hypotheses based on the host ecology were tested using different variables of landscape structure and configuration. Epidemiologic data of BSF by municipality were downloaded. Data on land use and land cover were retrieved from the MapBiomas initiative. Additional data on municipalities, hydrography, elevation, and climate were gathered. For each municipality, landscape structure and additional data were processed using geographic information systems. Landscape configuration metrics were calculated using the FRAGSTATS software. Subsequently, the relationships between BSF incidence and landscape features were assessed using generalized linear mixed models in RStudio. First, general explanatory model using environmental principal component analysis was carried out. Second, competing hypotheses regarding the role of specific environmental features were tested using multiple Akaike information criterion model selection. The results shows that BSF cases have increased over the last two decades and most of them were located in the Atlantic Forest. There was a significant correlation between BSF incidence and forest cover, especially riparian forests cover immersed in pasture and agriculture landscape matrices. The results are interpreted by the ecological requirements of BSF tick vectors and their hosts in the context of landscape changes in the Atlantic Forest over the period studied. Finally, applications for an enhanced preventive framework of the disease are discussed.