Geostatistical Modeling and Prediction of Rift Valley Fever Seroprevalence among Livestock in Uganda

Abstract

ABSTRACT. Uganda reported cases of Rift Valley fever virus (RVFV) for the first time in almost 50 years in 2016, following an outbreak of Rift Valley fever (RVF) that caused four human infections, two of which resulted in death. Subsequent outbreak investigation serosurveys found high seroprevalence of IgG antibodies without evidence of acute infection or IgM antibodies, suggesting the possibility of undetected RVFV circulation prior to the outbreak. After the 2016 outbreak investigation, a serosurvey was conducted in 2017 among domesticated livestock herds across Uganda. Sampling data were incorporated into a geostatistical model to estimate RVF seroprevalence among cattle, sheep, and goats. Variables resulting in the best fit to RVF seroprevalence sampling data included annual variability in monthly precipitation and enhanced vegetation index, topographic wetness index, log human population density percent increase, and livestock species. Individual species RVF seroprevalence prediction maps were created for cattle, sheep, and goats, and a composite livestock prediction was created based on the estimated density of each species across the country. Seroprevalence was greater in cattle compared with sheep and goats. Predicted seroprevalence was greatest in the central and northwestern quadrant of the country, surrounding Lake Victoria, and along the Southern Cattle Corridor. We identified areas that experienced conditions conducive to potential increased RVFV circulation in 2021 in central Uganda. An improved understanding of the determinants of RVFV circulation and locations with high probability of elevated RVF seroprevalence can guide prioritization of disease surveillance and risk mitigation efforts.