Background:  An abacavir (ABC)-containing nucleoside reverse transcriptase inhibitor (NRTI) regimen has potential advantages over standard first-line regimens in Africa (nevirapine [NVP] + 2 NRTI), as it avoids drug interactions with TB therapy, spares drug classes for second-line therapy and has a low pill burden. However the tolerability and toxicity of ABC in Africa are unknown.

Methods:  A 24-week randomized double-blind trial comparing the safety of NVP or ABC (NORA) was conducted in 2 Ugandan centers within the DART trial. 599 symptomatic ARV-naive adults with CD4 <200 cells/mm³ were allocated to zidovudine/lamivudine (Combivir) plus either 300 mg ABC and NVP placebo (n = 300) or 200 mg NVP and ABC placebo (n = 299) twice daily.

Results:  Of the total, 429 patients were women (72%); at baseline, 111 (19%) had WHO stage 4 disease, median age was 37 years (range 19 to 67), and CD4 99 cells/mm³ (1-199). Most (569 or 95%) patients completed 24 weeks, however, 24 (4%) died and 6 (1%) were lost to follow-up. As defined in ICH-GCP, 34 independently adjudicated serious adverse events occurred on blinded drug (plus 30 days) in 33 patients:   20 patients (6, or 2.0%, on ABC and 14, or 4.7%, on NVP) were classified as having serious adverse reactions (definitely/probably/uncertainly related to blinded ABC/NVP) (primary endpoint:  HR = 0.41; 95%CI 0.16 ro 1.08; log rank p = 0.06). Serious adverse events considered unlikely to be related to blinded drug were predominantly anemia (n = 7). Serious adverse reactions reported in 19 of 20 patients were consistent with potential hypersensitivity reactions:  similar proportions on ABC (4 of 6) and NVP (7 of 13) had respiratory, constitutional (3 ABC, 7 NVP), or gastrointestinal symptoms (2 ABC, 5 NVP), rash (6 ABC, 10 NVP),fever (6 ABC, 9 NVP), or oral/mucosal involvement (2 ABC, 5 NVP); but hepatic involvement was seen only in 3 patients on NVP. The remaining serious adverse reaction was asymptomatic hepatitis (NVP). In total, 14 (4.7%) ABC and 30 (10.0%) NVP patients discontinued blinded ABC/NVP (p = 0.01):  toxicity (6, or 2.0%, vs 15, or 5.0%, respectively, p = 0.05), all rash or possible hypersensitivity reactions (6 ABC, 13 NVP), or hepatotoxicity (4 NVP), and including the 6 ABC and 14 NVP patients with serious adverse reactions), to start TB therapy (6, or 2.0%, vs 13, or 4.3%), or for other reasons (n = 4:  pregnancy [1 NVP], personal [2 ABC, 1 NVP]).

Conclusions:  Compared with NVP, ABC had a lower discontinuation rate and a trend toward a lower rate of serious adverse reactions in African patients initiating ART with low CD4 counts. With the possible exception of hepatotoxicity, there was considerable overlap in the clinical manifestations of reactions to ABC and NVP in this population.