BACKGROUND — Antiretroviral therapy (ART) significantly reduces tuberculosis (TB) incidence among persons with human immunodeficiency virus (HIV), but the safety and effectiveness of concomitant treatment for both diseases remain unclear. OBJECTIVE — To evaluate the impact of ART and anti-tuberculosis treatment on survival and risk of adverse events (AE) among co-infected individuals. METHODS — In a retrospective cohort study, clinical data were collected from 618 TB-HIV patients treated with rifampin, isoniazid and pyrazinamide ± ethambutol between 1 January 1995 and 31 December 2003. Patients were categorized into two groups: highly active ART (HAART) or no ART. Different HAART regimens were evaluated. Bivariate analysis, multivariate logistic regression and survival analysis using Cox proportional hazards regression were used. RESULTS — One-year mortality was lower for patients receiving HAART (adjusted hazard ratio [aHR] 0.17, 95%CI 0.09–0.31) compared to no ART. HAART increased the risk of AE (aHR 2.08, 95%CI 1.29–3.36). The odds of AE when receiving a ritonavir + saquinavir HAART regimen was eight-fold higher compared to no ART (OR 8.31, 95%CI 3.04–22.69), while efavirenz-based HAART was not associated with a significantly increased risk of AE (OR 1.42, 95%CI 0.76–2.65). CONCLUSION — HIV patients with TB have significantly better survival if they receive HAART during anti-tuberculosis treatment. Efavirenz-based HAART is associated with fewer AEs than protease inhibitor-based HAART.