The effect of livestock density on Trypanosoma brucei gambiense and T. b. rhodesiense: a causal inference-based approach

Domestic and wild animals are important reservoirs of the rhodesiense form of human African trypanosomiasis (rHAT), however quantification of this effect offers utility for deploying non-medical control activities, and anticipating their success when wildlife are excluded. Further, the uncertain role of animal reservoirs—particularly pigs—threatens elimination of transmission (EOT) targets set for the gambiense form (gHAT). Using a new time series of high-resolution cattle and pig density maps, HAT surveillance data collated by the WHO Atlas of HAT, and methods drawn from causal inference and spatial epidemiology, we conducted a retrospective ecological cohort study in Uganda, Malawi, Democratic Republic of Congo (DRC) and South Sudan to estimate the effect of cattle and pig density on HAT risk. For rHAT, we found a positive effect for cattle (RR 1.61, 95% CI 0.90, 2.99) and pigs (RR 2.07, 95% CI 1.15, 2.75) in Uganda, and a negative effect for cattle (RR 0.88, 95% CI 0.71, 1.10) and pigs (RR 0.42, 95% CI 0.23, 0.67) in Malawi. For gHAT we found a negative effect for cattle in Uganda (RR 0.88, 95% CI 0.50, 1.77) and South Sudan (RR 0.63, 95% CI 0.54, 0.77) but a positive effect in DRC (1.17, 95% CI 1.04, 1.32). For pigs, we found a positive gHAT effect in both Uganda (RR 2.02, 95% CI 0.87, 3.94) and DRC (RR 1.23, 95% CI 1.10, 1.37), and a negative association in South Sudan (RR 0.66, 95% CI 0.50, 0.98). While ecological bias may drive the findings in South Sudan, estimated E-values and simulation studies suggest unmeasured confounding and underreporting are unlikely to explain our findings in Malawi, Uganda, and DRC. Our results indicate cattle and pigs are important reservoirs of rHAT in Uganda but not Malawi, and that pigs—and possibly cattle–may be gHAT reservoirs. Author summary Domestic animals, including cattle and pigs, are known to be important for the transmission of the rhodesiense form of human African trypanosomiasis (rHAT), however the relative importance of these reservoirs compared to wild animals is uncertain and likely focus-specific. For the gambiense form (gHAT) transmission is predominantly human-to-human, however pigs are thought to be a possible reservoir. In this study we used pre-existing data on livestock density and HAT risk to estimate the strength of the effect of cattle and pig density on rHAT risk in Uganda and Malawi, and gHAT risk in Uganda, Democratic Republic of Congo (DRC), and South Sudan. We found evidence that cattle and pigs increase the risk of rHAT in Uganda but not Malawi, that pigs increase the risk of gHAT in Uganda and DRC, and that cattle increase the risk of gHAT in DRC alone. These results indicate that control of both forms of HAT should include domestic animals in a One Health framework, however control of rHAT in Malawi is unlikely to be achieved if measures exclude wild animals.