Numerous national antiretroviral (ARV) treatment initiatives offering protease inhibitor-sparing combination antiretroviral therapy (cART) have recently commenced in southern Africa, the first of which began in Botswana in January 2002. Evaluation of the efficacy and tolerability of various protease inhibitor-sparing cART regimens requires intensive study in the region, as does investigation of the development of drug resistance and the optimal means of sustaining adherence. The "Tshepo" Study is the first large-scale, randomized, clinical trial that addresses these important issues among HIV-1 subtype C-infected ARV treatment-naive adults in southern Africa. The Tshepo Study is a completed, open-labeled, randomized study that enrolled 650 ARV-naive adults between December 2002 and 2004. The study is a 3 x 2 x 2 factorial design comparing the efficacy and tolerability among factors: (1) 3 combinations of nucleoside reverse transcriptase inhibitors (NRTIs): zidovudine (ZDV) + lamivudine (3TC), ZDV + didanosine (ddI), and stavudine (d4T) + 3TC; (2) 2 different nonnucleoside reverse transcriptase inhibitors (NNRTIs): nevirapine and efavirenz; and (3) 2 different adherence strategies: the current national "standard of care" versus an "intensified adherence strategy" incorporating a "community-based directly observed therapy." Study patients were stratified into 2 balanced CD4 T-cell count groups: less than 201 versus 201-350 cells per cubic millimeter with viral load greater than 55,000 copies per milliliter. Following Data Safety Monitoring Board recommendations in April 2006, ZDV/ddI-containing arms were discontinued due to inferiority in primary end point, namely, virologic failure with resistance. We report both overall data and pooled data from patients receiving ZDV/ddI- versus ZDV/3TC- and d4T/3TC-containing cART through April 1, 2006. Four hundred fifty-one females (69.4%) and 199 males with a median age of 33.3 years were enrolled into the study. The median follow-up as of April 1, 2006, was 104 weeks, and loss to follow-up rate at 2 years was 4.1%. The median baseline CD4 T-cell count was 199 cells per cubic millimeter [interquartile ratio (IQR) 136-252], and the median plasma HIV-1 RNA level was 193,500 copies per milliliter (IQR 69-250, 472-500). The proportion of participants with virologic failure and genotypic resistance mutations was 11% in those receiving ZDV/ddI-based cART versus 2% in those receiving either ZDV/3TC- or d4T/3TC-based cART (P = 0.002). The median CD4 T-cell count increase at 1 year was 137 cells per cubic millimeter (IQR 74-223) and 199 cells per cubic millimeter (IQR 112-322) at 2 years with significantly lower gain in the ZDV/ddI arm. At 1 and 2 years, respectively, 92.0% and 88.8% of patients had an undetectable plasma HIV-1 RNA level (< or = 400 copies/mL). Kaplan-Meier survival estimates at 1 and 2 years were 96.6% and 95.4%. One hundred twenty patients (18.2%) had treatment-modifying toxicities, of which the most common were lipodystrophy, anemia, neutropenia, and Stevens-Johnson syndrome. There was a trend toward difference in time to treatment-modifying toxicity by pooled dual-NRTI combination and no difference in death rates. The preliminary study results show overall excellent efficacy and tolerability of NNRTI-based cART among HIV-1 subtype C-infected adults. ZDV/ddI-containing cART, however, is inferior to the dual NRTIs d4T/3TC or ZDV/3TC when used with an NNRTI for first-line cART.