902. Unreturned Pill Bottles in the 1489 and 1490 Clinical Trials: An Important Measure of Poor Adherence that Is Often Ignored in Pill Count Calculations

Abstract

Abstract Background Adherence to antiretroviral therapy is important for HIV suppression. In clinical trials, adherence is commonly measured by pill count; limitations are that unreturned pills may not have been taken and unreturned pill bottle data are omitted. This analysis focuses on the relationship between unreturned pill bottles as a measure of poor adherence and the overall effect on virologic success rates across all treatment groups. Methods Pill bottle return category (all bottles returned or ≥1 bottle unreturned) and return rates for participants across all treatment groups from two treatment-naïve INSTI clinical trials (Studies 1489 and 1490) were calculated. Association of bottle return category or rates with virologic events through week 144, including last on-treatment observation carried forward (LOCF) outcome, was determined; comparisons used Fisher’s exact or Wilcoxon rank sum test. Results Virologic suppression with ≥95% adherence by pill count can differ for those with unreturned pill bottles (Figure). In these studies, 60% of participants returned all their pill bottles through week 144; if one visit with ≥1 unreturned bottle was allowed, this percentage increased to 81%. The mean bottle return rate was 94% and did not differ by study, treatment arm, or sex. Failure to return pill bottles was significantly associated with lower suppression rates. Additionally, significant differences in pill bottle return rate (p < 0.01) were observed by week 144 LOCF outcome (95% vs 77% mean return rate for those with HIV RNA < 50 c/mL vs ≥50 c/mL), need for resistance testing (95% vs 77% return rate for those not tested vs tested), confirmed virologic failure (VF) (94% vs 90% return rate for those without VF vs with VF) and blip status (95% vs 92% return rate for those without blips vs with blips). HIV-1 Viral Loads of Two Participants with ≥95% Adherence by Pill Count through Week 144 Conclusion In these treatment-naïve INSTI clinical trials, failure to return pill bottles was associated with lower suppression rates. Although the calculated adherence rates in these studies was relatively high (median ≥95%), these calculations did not account for unreturned pill bottles. We believe that assessing adherence by both pill count and pill bottle return rate may provide a more complete picture of adherence in clinical trials. Disclosures Rima K. Acosta, BS, Gilead Sciences, Inc. (Employee, Shareholder) Grace Q. Chen, BS, Gilead Sciences, Inc. (Employee) Hailin Huang, PhD, Gilead Sciences, Inc. (Employee, Shareholder) Hui Liu, PhD, Gilead Sciences, Inc. (Employee, Shareholder) Kirsten L. White, PhD, Gilead Sciences, Inc (Employee, Shareholder)