Abstract
Country-specific interventions targeting high-risk groups are necessary for a global reduction in Tuberculosis (TB)/HIV burden. We analyzed the data of 13,802 TB cases diagnosed in Harare, Zimbabwe, during 2013-2017. Pearson's chi-square tests and multivariate logistic regression models were used to identify patient characteristics significantly associated with TB/HIV coinfection. Of the 13,802 TB cases analyzed, 9,725 (70.5%) were HIV positive. A significantly higher odds of having TB/HIV coinfection diagnosis was found among females, patients aged 25-64 years, previously treated cases, and acid-fast bacillus sputum smear-negative cases. Compared with nondisseminated pulmonary TB, miliary TB (adjusted odds ratio [aOR]: 1.469, 95% CI: 1.071, 2.015) and TB meningitis (aOR: 1.715, 95% CI: 1.074, 2.736) both had a significantly higher odds for TB/HIV coinfection, whereas pleural TB (aOR 0.420, 95% CI: 0.354, 0.497) and all other extrapulmonary TB (EPTB) (aOR: 0.606, 95% CI: 0.516 0.712) were significantly less likely to have TB/HIV coinfection. The risk for TB/HIV coinfection varied significantly by patients' sociodemographic and clinical characteristics in Harare. Our finding that different forms of EPTB have different relationships with HIV coinfection has extended the knowledge base about clinical markers for TB/HIV coinfection which can lead to a greater public health impact on eliminating TB/HIV infection.