Abstract Body

Background: There are practical issues, including high cost, in implementing virologic monitoring of patients on antiretroviral therapy (ART) in resourcelimited settings. We evaluated factors that might guide less frequent virologic monitoring after successful virologic suppression on efavirenz (EFV)-based ART in such settings. Methodology: HIV-1 RNA was measured every 8 weeks in antiretroviral naive participants with nadir CD4 counts <300/mm3 who were randomized to EFV with lamivudine/zidovudine or emtricitabine/tenofovir in the international ACTG A5175/PEARLS trial, and who achieved HIV-1 RNA <1000 c/mL at 24 weeks without prior virologic failure or new AIDS event. Predictors of late virologic failure (late VF: confirmed HIV-1 RNA ≥1000 c/mL after 24 weeks) were evaluated by multivariate Cox models using stepwise selection from among the following variables: demographic; pre-ART CD4, HIV-1 RNA, AIDS, TB, hepatitis B, laboratory test results and body mass index; ART regimen; occurrence of Grade ≥3 laboratory results and signs/symptoms, ART change, self-reported adherence before week 24; and week 24 CD4 change and HIV-1 RNA (≤400 vs 401-999). Results: 911 of 1045 subjects (87%) had HIV-1 RNA <1000 at 24 weeks. During median follow-up of >3yr after week 24, 82 of the 911 (9%) experienced late VF. VF rate appeared higher between 24 to 48 weeks (2.4%) but was reasonably constant thereafter (~1.25% every 24 weeks). In multivariate analysis, increased late VF risk was significantly associated with lower pre-ART hemoglobin (adjusted hazard ratio [AHR]=1.18 per 1g/ dL lower; 95% CI: 1.04, 1.33; p=0.008), and both self-reported imperfect ART adherence (55/516 [10.7%] vs 27/413 [6.8%] for perfect adherence; AHR=1.72; CI: 1.08, 2.72; p=0.022) and absence of Grade ≥3 laboratory results prior to week 24 (77/800 [9.6%] vs 5/111 [4.5%] for those with Grade ≥3 results; AHR=2.53; CI: 1.02, 6.29; p=0.046). Sensitivity analyses restricted to subjects who did not change ART regimen, and to those without Grade ≥3 laboratories, prior to week 24 showed similar results. In time-updated analysis, self-reported imperfect adherence was a significant predictor of increased late VF risk. Conclusions: In this clinical trial setting, the rate of late VF after successful suppression on EFV-containing ART was very low suggesting that virologic monitoring strategies involving infrequent (eg annual) measurements, possibly targeting subjects reporting imperfect adherence and those with lower pre-ART hemoglobin, might be considered; the implications of this for accumulation of resistance mutations needs evaluating. The unexpected finding that subjects without Grade ≥3 laboratory results before 24 weeks had higher risk of late VF needs evaluating further in a larger study.