Spatial association between sporotrichosis in cats and in human during a Brazilian epidemics

Abstract

Sporotrichosis is a subcutaneous granulomatous disease caused by the fungus Sporothrix spp. In Brazil, S. brasiliensis is reported in regions of outbreaks and epidemics in the zoonotic form of the disease where cats play an important role in the transmission of the disease to humans. Therefore, it is important to assess how the presence of infected cats impacts the risk for sporotrichosis in humans. The objective of this study was to analyze the spatial association of sporotrichosis in cats and in humans from Belo Horizonte, a Brazilian city where an epidemics of sporotrichosis occurs since the first human case register in 2015, through an inhomogeneous Poisson process model. Feline and human cases of sporotrichosis recorded between January 2016 and June 2019 were georeferenced by address and spatial point patterns were generated. Feline case intensity and human demographic density were calculated using a kernel smoothed estimate. The distance to the nearest feline case was also compute. Model parameters were estimated by Maximum Likelihood Estimate. The model validation was performed by the evaluation of partial residual, leverage and influence measure. There were 343 cases of cats and 135 human cases of sporotrichosis. The average incidence of human sporotrichosis in the period was 1.343 per 100 thousand inhabitants, which is relatively low in relation to the population, but higher than that observed in other regions in zoonotic outbreak of the disease. The southern region of the municipality has a higher intensity of feline cases. According to the fitted model, the risk for human sporotrichosis is greater when at distances very close to a feline case, with a virtually stable effect for distances greater than 1 km. Regarding the intensity of feline cases there is a gradual increase in risk as the intensity of cases increases. From the leverage analysis it was observed that the model was particularly sensitive to the occurrence of human cases in the south and east regions, places with extreme values ​​of covariates. Poisson point process model seems to be a reasonable approach in spatial epidemiology when multiple sources of infection are involved, and there is a low incidence of the disease as long as it is reasonable to assume independence between cases. Interventions for disease prevention and control in humans are suggested to encompass disease control in cats and the search for feline cases, focused on diagnosis and control, close to reported human cases.