1564Duration of Viral Suppression and objectively measured Antiretroviral Adherence as predictors of rebound Viremia in South Western Uganda

Background. Relationships between antiretroviral (ART) adherence and virologic suppression are critical to understanding risk of treatment failure and drug resistance. Methods. We used objective monitors to collect adherence data on adults taking ART during 2005-2011. CD4 and viral load were collected quarterly. Our primary outcome was rebound HIV-1 RNA viremia, defined by a detectable viral load following a previously undetectable viral load (<400 copies/mL). Our primary predictor was ART adherence, measured by: 1) average adherence, calculated by the number of pills taken divided by the number of pills prescribed; and 2) presence of three consecutive days of non-adherence. Secondary predictors were CD4 nadir and ART regimen. We used generalized estimating equations to fit regression models to assess for relationships between predictors and our outcome of interest. Results. 399 participants contributed 2,837 intervals during the follow-up. Median nadir CD4 was 125 and median duration of suppression was 1.1 years. The most common regimen was AZT/3TC/NVP, used by 58% of subjects. Risk of rebound viremia was higher during the first year of suppression than after in all adherence strata (3.5% vs 1.9% for average adherence >90%, 5.3% vs 2.1% for average adherence 60-90%, and 14.2% vs 7.3% for average adherence <60%, P<0.06 for all categories). Compared to periods with average adherence >90%, we found increased odds of rebound viremia for adherence <60% during the first year (AOR = 6.0, P <0.001) and after (AOR = 3.4 P =0.003); but increased odds for adherence 60-90% during the first year only (AOR= 1.9, P = 0.04 vs AOR = 0.8, P =0.60). One or more 72-hour gap was associated with increased odds of rebound viremia during the first year (AOR = 1.8, P = 0.05) and after (AOR = 2.2, P = 0.03). CD4 nadir was not predictive of rebound viremia, and there were no significant interactions in these relationships by regimen. Conclusion. In a cohort of PLWH on ART in rural Uganda, increasing duration of viral suppression was associated with reduced odds of viral rebound regardless of average adherence. Average adherence under 60% and one or more 72-hour treatment gaps are associated with increased risk of rebound viremia regardless of suppression time. Disclosures. All authors: No reported disclosures.