Background:   Numerous public initiatives offering protease inhibitor (PI) -sparing HAART have recently commenced in sub-Saharan Africa, the first being in Botswana. The efficacy, tolerability, development of drug resistance, and adherence of various PI-sparing HAART regimens therefore requires intensive study in the region. We herein report preliminary data from an ongoing randomized clinical trial among HAART-treated adults in Botswana.     

Methods:  Ongoing, open-label, randomized study that enrolled 650 ART-naïve adults between December 2002 and December 2004. A 3x2x2 factorial design comparing efficacy and tolerability among factors:  A = zidovudine (ZDV)/lamivudine (3TC) vs ZDV/didanosine (ddI) vs stavudine (d4T)/3TC; B = efavirenz (EFV) vs nevirapine (NVP); and C = community-based directly observed therapy vs standard of care. There were 2 balanced CD4 strata:  <201 vs 201 to 350 with viral load >55,000. Following recent Data and Safety Monitoring Board (DSMB) recommendations, ZDV/ddI-containing HAART arms were closed due to inferiority in primary endpoint (virological failure). We report pooled data from patients receiving ZDV/ddI- vs ZDV/3TC- or d4T/3TC-containing HAART regimens.   

Results:  Baseline characteristics were:  451 female (69.4%) and 199 males; median age 35.9 years with median follow-up of 89.7 (65.1 to 119.7) weeks; lost to follow-up rate of 3.4%; median baseline CD4⁺ 199 (IQR 134 to 239); and plasma viral load 193,000 (IQR 65,000 to 468,000). Rates of virologic failure with genotypic resistance were 11% ZDV/ddI vs 2% ZDV/3TC or d4T/3TC; p = 0.0002. Across all arms, median CD4⁺ increase was 140 (IQR 73 to 229) and 201 (IQR 114 to 322) with 92.3% and 88.8% having undetectable viral loads at 1 and 2 years, respectively. Kaplan-Meier survival estimates at 2 years were 88.6% (CD4⁺ <50), 94.4% (CD4⁺ 51 to 200), and 96.6% (CD4⁺ 201 to 350). Treatment-modifying toxicities occurred in 124 patients, the most common being neutropenia (3.4%), hepatotoxicity (2.3%), Stevens Johnson syndrome (1.8%), anemia (1.5%), and lactic acidosis (1.2%). There was no difference in death rates or time to first treatment modifying toxicity by dual nucleoside reverse transcriptase inhibitor (NRTI) combination.

Conclusions:  ZDV/ddI-containing HAART is inferior when used as an initial PI-sparing regimen. Response to PI-sparing HAART regimens containing ZDV/3TC and d4T/3TC, the 2 most commonly offered dual NRTI combinations in the region, is excellent. Based on these findings, longer term outcomes are needed to determine whether there are significant differences in virologic suppression, resistance patterns, and toxicity profiles.