Background: Most patients in resource-limited countries receive ART with minimal immunological/virologic monitoring. Data on long-term virologic outcomes in such populations are lacking.

Methods:  The Development of AntiRetroviral Therapy in Africa (DART) was a randomized trial comparing clinically driven monitoring (CDM) vs routine laboratory (CD4/hematology/biochemistry) and clinical monitoring (LCM) in HIV-infected adults initiating ART with CD4 <200 cells/mm³ in Africa. No real-time virologic monitoring was performed. Ugandan patients still on first-line ART had an HIV RNA viral load test by Roche Taqman 1.0 or Roche Amplicor 1.5 at trial exit into national programs (January to March 2009).

Results: 2317 patients initiated ART with zidovudine/lamivudine plus tenofovir (n = 1717), abacavir (n = 300), or nevirapine (n = 300). By trial exit, 275 had died and 146 were lost to follow-up; 1455 (63%) were alive on first-line ART, a median 64 months (range 50 to 75) after ART initiation; 441 were on second-line ART (majority in second-line randomizations, so not considered here). A viral load result was available on 1164 (80%) first-line patients, with no evidence of systematic bias between those with and without a measurement. Viral load was <200 copies/mL in 80.2% (933), 200 to 999 copies/mL in 5.0% (58), 1000 to 9999 copies/mL in 9.0% (105), and ≥10,000 copies/mL in 5.8% (68). Viral load <200 copies/mL was marginally higher in LCM (483/587 [82.3%]) than CDM (450/577 [78.0%]) (difference 4.2%, 95%CI 0.0 to 8.9%, p = 0.07). Assuming pessimistically that viral load was ≥200 copies/mL at death and all second-line patients had viral load failure at switch, the overall estimated suppression rate at exit would be 54.8% (95%CI 52.5 to 57.0%). Multivariate logistic regression analysis found better viral load suppression in older patients (OR 1.32/10 years, p = 0.007), females (OR 1.36, p = 0.06), and patients with a higher pre-ART CD4 count (OR 1.39/100 cells/mm³, p = 0.02). Among 231 non-suppressed (≥200 copies/mL) patients, 144 (67.1%) had a current CD4 count ≥200 cells/mm³.

Conclusions:  A low rate of virologic failure was observed in this cross-sectional analysis of non-virologicly monitored patients after ~5 years of first-line ART, questioning the cost-effectiveness of routine virologic monitoring in resource-limited settings. The trend towards better viral load suppression in patients receiving CD4 monitoring likely reflects earlier switching to second-line ART in this group.