Risk mapping and determinants of the potential presence of Anopheles stephensi in Benin: a geostatistical and logistic regression approach

Anopheles stephensi is an emerging malaria vector, historically confined to the Indian subcontinent but recently confirmed in several African countries. Its distinctive capacity to exploit anthropogenic water storage containers—cisterns, water tanks, wells, and discarded tires—enables adaptation to urban and peri-urban environments independent of seasonal rainfall, thereby representing a growing and novel threat to vector control strategies. In Benin, favorable environmental and demographic conditions suggest a potential risk of ecological establishment, although this has not yet been confirmed by comprehensive entomological surveys. This study aimed to identify the determinants associated with the ecological suitability for Anopheles stephensi and to map the spatial distribution of predicted potential occurrence risk across all 77 municipalities of Benin. An ecological study encompassing all 77 municipalities of Benin was conducted using agricultural, livestock, socioeconomic, demographic, and environmental data collected between March and May 2025. The dependent variable (GITES_BINARY) did not reflect observed vector presence, but represented an ecological proxy of habitat suitability derived from systematic field surveys of artificial water storage structures, based on pre-established An. stephensi habitat suitability criteria (inter-rater reliability: Cohen's κ = 0.82). Multivariable logistic regression followed by probabilistic spatial modeling with k-fold cross-validation (k = 5) was used to estimate the likelihood of potential ecological suitability. Model performance was assessed using AUC (with 95% CI), sensitivity, specificity (at the optimal Youden threshold), and Brier calibration score. Human population density (p = 0.001) and swine population size (p < 0.001) were positively associated with ecological suitability for Anopheles stephensi , whereas cattle, sheep, and goat populations showed significant negative associations (p < 0.001). Model performance: AUC = 0.81 (95% CI 0.72–0.90); sensitivity = 0.76; specificity = 0.78 (Youden threshold); Brier score = 0.17. Risk maps represent probabilities of ecological suitability rather than confirmation of established vector presence. This study identifies urban and peri-urban municipalities of Benin as zones of highest ecological suitability for Anopheles stephensi , primarily driven by human population density and swine husbandry. The absence of associations with rainfall and temperature is consistent with the distinctive ecology of An. stephensi , which is independent of seasonal rainfall and insensitive to high temperatures. These findings provide prioritization tools for entomological surveillance and preventive interventions, with particular attention to the distinct characteristics of urban, peri-urban, and rural biotopes.