Background: We analysed the spatial and spatiotemporal patterns of visceral leishmaniasis (VL) mortality at the municipality level in an endemic state in the southern Amazon region of Brazil. Individual-level factors associated with death due to VL were also investigated.
Methods: All VL cases and deaths reported between 2007 and 2018 were included. The global and local bivariate Moran's index assessed the space-time autocorrelation of smoothed triennial VL mortality. Kulldorff's scan statistics investigated spatial and spatiotemporal clusters. A multivariable logistic regression explored sociodemographic, diagnostic and clinical variables associated with death due to VL.
Results: We observed an overall VL mortality and lethality of 0.14 cases/100 000 inhabitants and 11.2%, respectively. A total of 14% of the municipalities registered at least one VL-related death. In the southeastern mesoregion of the state, we detected high-risk spatial (relative risk [RR] 14.14; p<0.001) and spatiotemporal (RR 15.91; p<0.001) clusters for VL mortality. Bivariate Moran's analysis suggested a high space-time autocorrelation of VL mortality. Death by VL was associated with age ≥48 y (odds ratio [OR] 7.2 [95% confidence interval {CI} 3.4 to 15.3]), displacement for notification (OR 3.3 [95% CI 1.5 to 7.2]) and occurrence of oedema (OR 2.8 [95% CI 1.3 to 6.1]) and bleeding (OR 5.8 [95% CI 2.6 to 12.8]).
Conclusions: VL mortality has a heterogeneous spatiotemporal distribution. The death-related factors suggest late diagnosis as an underlying cause of mortality.
Keywords: case fatality rate; geographic information system; kala azar; risk factors; spatiotemporal analysis.
© The Author(s) 2022. Published by Oxford University Press on behalf of Royal Society of Tropical Medicine and Hygiene.