RESUMEN
BACKGROUND: This study was carried out to determine the incidence and predictors of anti-tuberculosis drug induced hepatotoxicity among TB/HIV co-infected patients at Jimma University Hospital, Ethiopia. METHODS/PRINCIPAL FINDINGS: A nested case-control study was conducted by reviewing charts of all TB/HIV co-infected patients who commenced anti-TB treatment from January 2008 to December 2011 at Jimma University Hospital. Patients who had developed hepatotoxicity after at least 5 days of standard doses of anti-TB drug therapy were labeled as "cases" and those without hepatotoxicity were "controls". Each case with anti-TB drug induced hepatotoxicity was compared with 3 controls selected randomly from the cohort. From a cohort of 296 TB/HIV co-infected patients 8 were excluded from the study as the causality between anti-TB drugs and hepatotoxicity was not confirmed, 33 had developed hepatotoxicity. On bivariate logistic regression analysis, body mass index (BMI) <18.5 Kg/m(2) [Pâ=â0.01; OR (95%CI): 3.6 (1.4-9.5)], disseminated pulmonary TB [Pâ=â0.00; OR (95%CI): 5.6 (2.2-14.6)], CD4 count ≤50 [Pâ=â0.016; OR (95%CI): 3.6(1.27-10.23)] and WHO stage 4 [Pâ=â0.004, OR (95%CI): 3.8 (1.68-8.77)] were significantly associated with anti-TB drug induced hepatotoxicity. Predictor variables with p-value <0.05 by bivariate analysis were analyzed using multivariable logistic regression analysis and identified disseminated pulmonary TB [Pâ=â0.001; AOR (95%CI)â=â5.6 (2.1-15.0)] and BMI <18.5 [Pâ=â0.014; AOR (95%CI)=â3.6 (1.3-10.1)] as independent predictors of anti-TB drug induced hepatotoxicity. CONCLUSIONS: The incidence of anti-TB drug induced hepatotoxicity was 11.5%. The results suggest that in the presence of disseminated pulmonary TB and/or BMI <18.5 Kg/m(2), TB/HIV co-infected patients should be closely followed for the occurrence of hepatotoxicity during the intensive phase of TB treatment to prevent morbidity and mortality.