Spatial distribution models of seroreactive sheep to Leptospira spp. in Veracruz, Mexico

Leptospirosis is an infectious zoonotic disease of special importance in tropical regions of the world and is closely related to climatic conditions. In Mexico, at least eight Leptospira serogroups are known to affect sheep, but little is known about their distribution. The aim was to analyse the spatial distribution of seroreactive sheep to eight serogroups of Leptospira spp. through ecological niche modelling from the state of Veracruz. We carried out a cross-sectional, multistage, and stratified epidemiological study, sampling 405 sheep in different regions of the state (north, center, and south). The sera were analysed using the microscopic agglutination test (MAT) to identify seropositivity to eight Leptospira serogroups (Icterohaemorrhagiae, Pyrogenes, Grippotyphosa, Canicola, Pomona, Hardjo, Wolffi, and Tarassovi). Management variables in the sampled herds were evaluated through a survey among the producers, which was analysed using the chi-squared test for cross-tabulation. Geospatial modelling was conducted using MAXENT and 19 climatic variables, and validation was carried out using the area under the curve (AUC). No positive animals were found for Pomona in any area of Veracruz, and there was only one case of seroreactivity to Grippotyphosa. The total seroprevalence found was 53.83% (95% confidence interval [CI] 48.84-58.75). The main serogroup found was Sejroe (55.31%, 95% CI 50.32-60.20%), followed by Canicola (8.64%, 95% CI 6.17-11.92%), Icterohaemorrhagiae (4.69%, 95% CI 2.93-7.36%), Tarassovi (3.95%, 95% CI 2.35-6.47%), Pyrogenes (2.47%, 95% CI 1.26-4.64%), Australis (0.99%, 95% CI 0.32-2.69%), and Grippotyphosa (0.25%, 95% CI 0.01-1.59%). The predictive model for Australis was not significant. Acceptable predictive models (AUC > 0.7-0.8) were found for Canicola, Icterohaemorrhagiae, Pyrogenes, and Tarassovi, while for Sejroe, it was excellent (AUC > 0.85); consequently, the climatic variables that most contributed to the model were those related to precipitation. The potential distribution of Pyrogenes, Icterohaemorrhagiae, and Canicola was located to a greater extent in the three regions; Pyrogenes and Tarassovi were distributed mostly in the north and central regions, and Sejroe was mostly located in the center and south of the state. Ecological niche modelling could support epidemiological control and surveillance programs for affected sheep herds in the state of Veracruz.