Cattle-associated risk factors for human tuberculosis in rural livestock keeping communities, Uganda

Tuberculosis (TB) is a leading infectious cause of human death worldwide. TB can also infect cattle, resulting in productivity losses, trade barriers, and zoonotic transmission via milk, meat, or direct contact. While the majority of TB cases are non-zoonotic, an unknown proportion are acquired from cattle; in Africa, this proportion is estimated to be 0.4% to 10%. We conducted a cross-sectional study in rural communities in southeastern and northwestern Uganda between 2014 and 2016 to evaluate the association between tuberculosis skin test (TST) positivity in humans and cattle-associated risk factors. Human and cattle skin testing was performed in communities followed by a survey of household practices. TST data are available on 493 humans, 250 men and 243 women; 184 individuals in total—111 men and 73 women—tested positive. Separate log binomial models were fit to estimate relative risks (RR) for herd TST positivity stratified on gender and for raw milk consumption, using generalized estimating equations. Having at least one TST positive bovid in the household’s herd was significantly associated with decreased risk of TB among men (PR 0.61, 95% CI 0.47, 0.79) but was not significantly associated with TB among women (PR 1.26, 95% CI 0.80, 1.97). This was contrary our a priori hypothesis of higher effect of exposure among men—the primary caretakers of cattle—than women. This apparent protective effect may be the result of residual confounding by socioeconomic status: wealthier individuals may be less likely to be TB positive, but more likely to have TST positive herds by virtue of larger herd sizes, ability to purchase new and possibly infected stock, and propensity to keep more TB-susceptible European breeds. For raw milk consumption, effect estimates were close to one and not statistically significant; adjustment for confounders or pathways mediated by exposure to non-zoonotic TB did not change the size or magnitude of effect estimates. The importance of cattle-associated risk factors for human TB burden may be setting-specific, as suggested by the lack of consensus reached by prior research. In settings where bovine TB prevalence is low, such as Uganda, cattle-associated zoonotic transmission may be rare.