Background: Reasons for the variation in reported treatment outcomes from antiretroviral therapy (ART) programmes in developing countries are not clearly defined.
Methods: Among ART-naïve individuals in a workplace ART programme in South Africa we determined virological outcomes at 12 months, and risk factors for suboptimal virological outcome, defined as plasma HIV-1 viral load > or = 400 copies/ml.
Results: Among 1760 individuals starting ART before July 2004, 1172 were in follow-up at 12 months of whom 953 (81%) had a viral load measurement (median age 41 yrs, 96% male, median baseline CD4 count 156 x 10(6)/l). 71% (681/953) had viral load < 400 copies/ml at 12 months. In a multivariable analysis, independent predictors of suboptimal virological outcome at 12 months were <1 log decrease in viral load at six weeks (odds ratio [OR] 4.71, 95% confidence interval [CI] 2.56-8.68), viral load at baseline (OR 3.63 [95% CI 1.88-7.00] and OR 3.54 [95% CI 1.79-7.00] for 10,001-100,000 and >100,000 compared to < or = 10,000 copies/ml, respectively), adherence at six weeks (OR 3.50 [95% CI 1.92-6.35]), WHO stage (OR 2.08 [95% CI 1.28-3.34] and OR 2.03 [95% CI 1.14-3.62] for stage 3 and 4 compared to stage 1-2, respectively) and site of ART delivery. Site of delivery remained an independent risk factor even after adjustment for individual level factors. At 6 weeks, of 719 patients with self-reported adherence and viral load, 72 (10%) reported 100% adherence but had <1 log decrease in viral load; conversely, 60 (8%) reported <100% adherence but had > or = 1 log decrease in viral load.
Conclusion: Virological response at six weeks after ART start was the strongest predictor of suboptimal virological outcome at 12 months, and may identify individuals who need interventions such as additional adherence support. Self reported adherence was less strongly associated but identified different patients compared with viral load at 6 weeks. Site of delivery had an important influence on virological outcomes; factors at the health system level which influence outcome need further investigation to guide development of effective ART programmes.