RESUMO
BACKGROUND: Highly active antiretroviral therapy (HAART) is strongly effective in reducing morbidity and mortality in HIV-1-positive individuals. Its main drawback is the potential toxicity. Data on the frequency and determinants of severe hepatotoxicity in a clinical setting are still sparse. METHODS: This is a prospective study of HIV-1-positive individuals with known HBsAg and HCV-Ab serology. The study end point was progression to alanine aminotransferase (ALT) levels > or =200 IU/L after HAART initiation. Cumulative probability of progression to this end point was estimated by the Kaplan-Meier method. Crude and adjusted hazard ratios (HR) were estimated by proportional hazards regression model. RESULTS: One thousand two hundred fifty-five patients were included. HBsAg was found in 91 (7.2%), HCV-Ab in 578 (46.5%) patients; almost all injection drug users (451 of 482; 93.6%) were HCV-Ab positive. Sixty-one individuals progressed to the end point with a probability of 7.9% (95% confidence interval [CI], 5.6-10.0) of progression at 24 months from starting. Independent factors predicting progression to the end point were baseline ALT levels (HR, 5.29; 95% CI, 3.24-8.65; every 10 IU/L higher), HCV-Ab positivity (HR, 4.01; 95% CI, 1.48-10.85) or both HBsAg and HCV Ab positivity (HR, 3.85, 95% CI, 1.01-14.61), and previous non-HAART therapy (HR, 1.84, 95% CI, 1.04-3.42). Patients receiving stavudine-containing regimens had a lower risk than those receiving zidovudine-containing regimens (HR, 0.30, 95% CI, 0.12-0.71). CONCLUSIONS: There was a low risk of ALT > or =200 IU/L in our cohort. Hepatitis C coinfection and elevated ALT levels at HAART initiation are important predictors of progression to ALT > or =200 IU/L; stavudine-containing regimens were associated with a lower risk compared with zidovudine-containing regimens.