First-line antiretroviral therapy durability in a 10-year cohort of naive adults started on treatment in Uganda

Introduction: The majority of studies from resource-limited settings only report short-term virological outcomes of patients on antiretroviral treatment (ART). We aim to describe the long-term durability of first-line ART and identify factors associated with long-term virological outcomes. Methods: At the Infectious Diseases Institute in Kampala, Uganda, 559 adult patients starting ART in 2004 were enrolled into a research cohort and monitored with viral load (VL) testing every six months for 10 years. We report the proportion and cumulative probability of 1) achieving virologic suppression (at least one VL <400 copies/ml); 2) experiencing virologic failure in patients who achieved suppression (two consecutive VLs >1000 copies/ml or one VL >5000, for those without a subsequent one); 3) treatment failure (not attaining virologic suppression or experiencing virologic failure). We used Cox regression methods to determine the characteristics associated with treatment failure. We included gender, baseline age, WHO stage, body mass index, CD4 count, propensity score for initial ART regimen, VL, time-dependent CD4 count and adherence. Results: Of the 559 patients enrolled, 472 (84.8%) had at least one VL (67 died, 13 were lost to follow-up, 4 transferred, 2 had no VL available); 73.6% started on d4T/3TC/nevirapine and 26.4% on AZT/3TC/efavirenz. Patients in the two groups had similar characteristics, except for the higher proportion of patients in WHO Stage 3/4 and higher VL in the efavirenz-based group. Four hundred thirty-nine (93%) patients achieved virologic suppression with a cumulative probability of 0.94 (confidence interval (CI): 0.92-0.96); 74/439 (16.9%) experienced virologic failure with a cumulative probability of 0.18 (CI: 0.15-0.22). In the multivariate analysis, initial d4T/3TC/ nevirapine regimen (hazard ratio (HR): 3.02; CI: 3.02 (1.66-5.44, p < 0.001)) and baseline VL >= 5 log10 copies/ml (HR: 2.29; CI: 1.29-4.04) were associated with treatment failures; patients of older age (HR: 0.87 per five-year increase; CI: 0.77-0.99), with adherence >95% (HR: 0.04; CI: 0.02-0.11) and with higher time-dependent CD4 count (HR: 0.94 per 50 cells/mu l increase; CI: 0.92-0.99, p < 0.001) were less likely to experience treatment failure. Conclusions: The long-term virological outcomes from this cohort are promising and comparable to those from research-rich settings. Our results provide further evidence that efavirenz is associated with better virological outcomes.