We designed a cohort in order to assess the long-term effects of triple-drug antiretroviral combinations in 608 patients infected with human immunodeficiency virus type 1 (HIV-1). We recruited patients who had been previously treated with nucleoside analogues as well as treatment-naive patients who were starting triple-drug antiretroviral combinations consisting of nucleoside analogues, either alone or in combination with a protease inhibitor. After a median follow-up time of 22 months, the incidence rates of acquired immune deficiency syndrome-defining events and death were, respectively, 6.9 (95% confidence interval [CI], 5.3-8.8) and 2.9 (95% CI, 1.9-4.2) per 100 person-years. Advanced clinical stage of disease (P=.004), a low CD4(+) cell count (P=.002), and a low quality-of-life score (P=.001) at baseline were independent predictors of clinical progression. The initial triple-drug combination was modified a total of 647 times in 321 patients. The only independent predictor of treatment modification was previous exposure to a nucleoside analogue in patients who did not receive a new nucleoside analogue at inclusion (P=.001). Plasma HIV RNA values below 500 copies/mL were obtained in 88% of the treatment-naive patients and in 57% of the previously treated patients (P<.001). Compared with previously treated patients who received > or = 1 new nucleoside analogue at enrollment, previously treated patients who did not receive a new nucleoside analogue at enrollment were twice as likely to have plasma HIV RNA values >500 copies/mL at the last visit (adjusted odds ratio [OR], 1.8; 95% confidence interval [CI], 1.2-2.8), and the antiretroviral-naive patients were significantly less likely to have plasma HIV RNA values >500 copies/mL at the last visit (adjusted OR, 0.2; 95% CI, 0.1-0.4).