CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

Conclusion: Ignoring mortality and disengagement from overestimates adherence to ART among all patients initiating ART and significantly exaggerates the effectiveness of treatment programs. IPCWmethods aimed at outreach data are broadly applicable, particularly also to viral suppression, which is associated with death and disengagement from care and which may also be substantially overestimated by excluding individuals not in care.

515 PATTERNS OF WISEPILL DEVICE USE IN AFRICAN ADULTS TAKING ANTIRETROVIRAL THERAPY (ART)

Catherine Orrell 1 , Nicholas Musinguzi 2 , Norma C. Ware 3 , Anna Cross 1 , Bosco M. Bwana 2 , Stephen Asiimwe 4 , Gideon Amanyire 2 , David R. Bangsberg 5 , Jessica E. Haberer 6 1 Desmond Tutu HIV Foundation, Cape Town, South Africa, 2 Mbarara University of Science and Technology, Mbarara, Uganda, 3 Harvard University, Cambridge, MA, USA, 4 Kabwohe Clinical Research Center, Kabwohe, Uganda, 5 Oregon Health and Sciences University, Portland, OR, USA, 6 Massachusetts General Hospital, Boston, MA, USA Background: ART adherence measured with electronic devices closely predicts virological outcomes. Use of such devices may not be consistent across settings or populations. Device non-use may explain some discrepancy between adherence data and virological outcomes, as has been suggested elsewhere. Other factors may play a role. Methods: This research was part of the Measuring Early Treatment Adherence (META) study examining ART adherence in early vs. late stage HIV disease, with sites in Cape Town, South Africa (SA) and south-western Uganda (UG) (NCT02419066). Real-time electronic monitoring, using the Wisepill®, was used to measure adherence. All participants were on TFV/3TC/EFV. For those with viral suppression (<400 copies/ml) at month 12, we plotted monthly adherence over a year after ART initiation by site and used linear regression modelling to explore predictors of average adherence. Independent variables were baseline age and CD4 count, site and number of >7day treatment interruptions. Those lost to follow up were considered to be virological failures. Results: Of 904 enrolled individuals, Wisepill data was available for 397 (94%) in SA and 472 (98%) in UG. Individuals in SA were older [mean age 34 years (sd 10) vs. 31 years (sd 9), p<0.001] with lower CD4 cell counts [median 370 (IQR 124-444) vs. 401 (IQR 155-490), p<0.001] at baseline. Women were in the majority at both sites (70% vs 68%, p=0.613). Viral suppression for those who had reached month 12 to date was achieved in 283/355 (80%) individuals from SA and 384/434 (88%) from UG. Adherence in suppressed individuals in SA was measured as 13-26% poorer than in UG across the observed year (Figure). Site (p<0.001) and longer gaps in adherence (non-linear, p<0.001) influenced the difference in adherence among individuals with viral suppression. The model accounted for 70% of the variance in adherence [F(7,659)=218.28, p<0.01, R2=0.70]. Conclusion: Many factors contribute to differences in adherence among those achieving viral suppression. Patterns of device use may differ by site and population. The site level effect may indicate device non-use due to stigma/ disclosure, suggested by qualitative findings (presented separately). Further support in the use of adherence monitoring devices in some populations and/ or device modifications to optimise use may be needed. Additionally, given the remaining ambiguity in the model, other potentially influential factors, such as inflammation, warrant investigation.

Poster Abstracts

514 USING INPATIENT DOT TO IDENTIFY TRUE ART FAILURE AND IMPROVE TREATMENT OUTCOMES Nicole E. Winchester 1 , Safia S. Kuriakose 2 , Frank Maldarelli 3 , Michael Proschan 1 , Yolanda Mejia 2 , Holly Ann Baus 1 , Megan Anderson 1 , Robin L. Dewar 2 , H. Clifford Lane 1 , Alice K. Pau 1 1 NIAID, Bethesda, MD, USA, 2 Leidos Biomedical Research, Inc, Frederick, MD, USA, 3 NCI, Rockville, MD, USA Background: Treatment failure (TF) remains a complex problem for some HIV+ patients (pts) despite potent ART. Most cases result from non-adherence, inadequate drug levels, or drug resistance. Regimen changes in response to TF may improve outcomes, however, they can also jeopardize future ART options and increase failure risk. In this observational study, we used an 8-day inpatient directly observed therapy (DOT) to assess the viral potency of the failing regimen before ART switches. Methods: Pts with multiple ART failures were eligible for enrollment if they failed at least 2 regimens with 2 last VL >1000 copies/mL. Efficacy of the current ART regimen was assessed using an 8-day inpatient DOT (DOT1) and VL was performed on days 1, 3, 5, and 8. A DOT Responder [DOT1-R] was defined as one who has a VL decline of ≥0.5 log 10 c/mL from screening to Day 8. Non-responders [DOT1-NR] were prescribed new regimens, if available, and returned for a 2 nd DOT (DOT2) to assess viral efficacy. Follow up occurred at wks 2, 4, 8, and 12, and then every 3 mos. HIV/ART education and adherence counseling were provided during DOT and at all follow-up visits by teammembers including medical, nursing, pharmacy, and social work staff per patient needs; patients were also contacted by phone between clinic visits. Results: 18 pts were enrolled; 16 completed DOT1. 1 pt had AKI at screening, had to stop TDF and did not have DOT1, 1 was lost to f/u before DOT. The remaining 16 pts comprise the analysis cohort and were 62.5%male; 87.5% black; mean age 43 yrs; mean CD4 80.3 cells/mm 3 ; median VL 4.39 log 10 c/mL. 8/16 (50%) had >0.5 log 10 c/mL decline with DOT1. The median VL changes for DOT1-R and DOT1-NR were -1.56 and -0.08 log 10 c/mL, respectively. DOT1-R, relative to DOT1-NR, had fewer yrs on ART (13.8 vs 19.9), fewer prior ARVs (9.1 vs 13.5), fewer DRM (9.5 vs 24.4), higher GSS scores (2.6 vs 0.2), and less education. 7 DOT1-NR underwent DOT2 on optimized ART. 5/7 (71%) of these pts responded to the new regimen with median VL change from day 1 to day 8 of -1.78 (-1.91 to -1.74) log 10 c/mL. 3 DOT1-NR had no active drugs as ART options. Conclusion: Using an 8-day inpatient DOT with intensive counseling, 50% of our pts with longstanding TF had substantial viral responses without ART changes. 5/7 (71.4%) DOT1-NR were virally suppressed after DOT2 with an optimized regimen. DOT is a valuable tool in identifying sub-optimal adherence as the cause of TF, reducing the frequency of unnecessary ART switches.

CROI 2018 187