CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

1029 COMPARISON OF TFV-DP AND WISEPILL ADHERENCE AMONG YOUNG KENYAN WOMEN USING PrEP Maria Pyra 1 , Elizabeth A.Bukusi 2 , Nelly R. Mugo 2 , Kenneth Ngure 3 , Kevin Oware 2 , Catherine Kiptinness 4 , Katherine Thomas 5 , Lindsey Garrison 6 , Peter L. Anderson 7 , Jared Baeten 5 , Jessica E. Haberer 6 1 Howard Brown Health Center, Chicago, IL, USA, 2 KEMRI–Centre for Global Health Research, Kisumu, Kenya, 3 Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya, 4 Partners in Health Research and Development, Thika, Kenya, 5 University of Washington, Seattle, WA, USA, 6 Harvard Medical School, Boston, MA, USA, 7 University of Colorado Anschutz Medical Campus, Aurora, CO, USA Background: Understanding PrEP use and related barriers is dependent on good measures of PrEP adherence. We compared electronic monitoring with tenofovir-diphosphate (TFV-DP) assessed from dried blood spots (DBS) to assess consistency of both measures in a cohort of young women taking PrEP. Methods: Participants were 18-24 year old women at high risk (by VOICE risk score) for HIV in Thika and Kisumu, Kenya. Participants were encouraged to take PrEP for at least the initial 6 months with study visits at 1 month, 3 months, and then quarterly for 24 months. The primary adherence measure was a real-time electronic monitor (Wisepill), while a random sample (15%) of DBS collected from non-pregnant participants receiving PrEP was also tested for TFV-DP. Adherence was categorized as high (>85% or 6+ doses/week; >1,050 fmol/ punch), moderate (57-85% or 4-5 doses/week; 700-1,050 fmol/punch), and low (<57% or <4 doses/week; <700 fmol/punch) for electronic monitoring and DBS, respectively. Descriptive comparisons were made between DBS and corresponding Wisepill openings over the prior 30 days. Results: DBS results from 39 samples (representing 36 women) over the first 12 months of follow-up were available. Overall concordance between electronically monitored and DBS adherence was moderate at 59%. Of the 21 participants with low electronically monitored adherence (<4 doses/week), almost all (n=20) had TFV-DP<700 fmol/punch, with an average of 128 fmol/punch. Of the 11 with moderate electronically monitored adherence (4-5 doses/week), 4 (10%) had TFV-DP ≥700 fmol/punch and an average of 559 fmol/punch. Of the 7 participants with high electronically monitored adherence (≥6 doses per week), only 1 had the expected TFV-DP ≥1050 fmol/punch, with an average of 577 fmol/punch. Notably, among women with high electronically monitored adherence, average TFV-DP was 526 fmol/punch when hemoglobin was <11 g/ dL versus 616 fmol/punch when hemoglobin was >11 g/dL. Conclusion: Overall, these findings showmoderate consistency between both adherence measures, although the established TFV-DP thresholds may be high in this African population, especially among those with lower hemoglobin level. We found no evidence that women were taking PrEP by DBS without concurrent dosing by electronic monitoring; however, overreporting by electronic monitoring is possible. Additional studies are warranted to fully characterize both adherence measures for young women who are an important key population for HIV prevention. 1030 HIV RISK AND OBJECTIVELY MEASURED PrEP ADHERENCE IN YOUNG KENYAN WOMEN Jessica E. Haberer 1 , Nelly R. Mugo 2 , Elizabeth A.Bukusi 2 , Kenneth Ngure 2 , Kevin Oware 2 , Catherine Kiptinness 2 , Maria Pyra 3 , Katherine Thomas 3 , Lindsey Garrison 1 , Jared Baeten 3 1 Massachusetts General Hospital, Boston, MA, USA, 2 Kenya Medical Research Institute, Nairobi, Kenya, 3 University of Washington, Seattle, WA, USA Background: Pre-exposure prophylaxis (PrEP) is a highly effective means for preventing HIV acquisition. However, use among young women has generally been suboptimal and poorly characterized. We present objectively measured adherence and associated socio-behavioral factors among a cohort of young women. Methods: Participants were 18-24 year old women at high risk for HIV in Thika and Kisumu, Kenya. High risk was defined as a VOICE risk score >4; points (pts) were given for age <25 (2 pts), being single or not living with a primary sexual partner (2 pts), lacking financial or material support from a sexual partner (1 pt), a sexual partner having or potentially having other partners (2 pts), and alcohol use (1 pt). Participants were encouraged to take PrEP for at least 6 months and then counseled on continued use based on their preferences and HIV risk. Study visits occurred at 1 month, 3 months, and then quarterly for 24 months. Adherence was measured with a real-time electronic monitor (Wisepill) and summarized descriptively. Baseline predictors of high adherence were assessed by multivariable logistic regression analysis.

Results: A total of 347 women have been followed for 461 person-years (as of June 2019; study to end early 2020). At 1, 3, and 6 months of desired PrEP use, 35%, 14%, and 7% of participants took an average of 6 doses per week, respectively, while 50%, 24%, and 15% took an average of 5 doses per week. The only baseline factor significantly associated with high adherence (an average of 6 doses per week) over 6 months was the VOICE risk score: OR 0.53 (95% CI 0.33, 0.85) for each additional point (figure). Non-significant factors included in the model were age, number of current sexual partners, concern about PrEP, prior medication use, and intimate partner violence. Findings were similar when high adherence was defined as an average of 5 doses per week. Conclusion: Objectively measured PrEP adherence likely to be sufficient for protection against HIV was seen in a minority of participants and declined with time. Higher baseline HIV risk was associated with lower adherence in the first 6 months of use. These findings suggest limited prevention-effective adherence, although future analyses will assess the alignment of risk and adherence over time. Similar assessments in other PrEP cohorts may be useful for program evaluation; novel approaches are needed to help young women understand risk and the means to achieve effective HIV prevention.

Poster Abstracts

1031 USING A MOBILE APP AND DRIED BLOT SPOTS TO ASSESS ADHERENCE TO EVENT-DRIVEN PrEP Vita Jongen 1 , Elske Hoornenborg 1 , Mark A.Van Den Elshout 1 , Liza N. Coyer 1 , Udi Davidovich 1 , Henry J. De Vries 1 , Maria Prins 1 , Maarten F. Schim Van Der Loeff 1 , for the Amsterdam PrEP Project team in the HIV Transmission Elimination Amsterdam (H-TEAM) Initiative 1 Public Health Service Amsterdam, Amsterdam, Netherlands Background: Both daily and event-driven (i.e. before and after sex) pre- exposure prophylaxis (PrEP) regimens are effective against HIV acquisition. However, information about adherence to event-driven PrEP is scarce and predominantly based on self-reported data collected at 3-month intervals. We used a mobile-based diary application and intracellular tenofovir diphosphate (TFV-DP) levels to assess adherence among event-driven PrEP users participating in the Amsterdam PrEP demonstration project (AMPrEP) in the Netherlands. Methods: Participants could choose and switch between daily and event- driven PrEP regimens. Participants used a mobile application to record their sexual behaviour and pill use on a daily basis. We studied adherence by assessing (1) the number of condomless anal sex (CAS) acts covered by PrEP using data collected by the mobile application and (2) the correlation between TFV-DP concentrations (measured in dried blood spot (DBS) samples taken at the 3, 6 or 9, and 12 and 24 month visits; lower level of detection 12.5 fmol/ punch) and CAS, and between TFV-DP concentrations and self-reported pill use. Good adherence was defined as at least one tablet before a CAS act and one tablet within 48 hours of that CAS act. Results: Between September 2015 and February 2019, 139 of 376 (37.0%) AMPrEP participants used event-driven PrEP for at least 3 months. In this period, a total of 6,583 CAS acts were reported in the mobile application during event- driven PrEP use, of which 5,518 (83.8%) were covered by good PrEP adherence. Good PrEP adherence was more common around CAS acts with known (93.0%) and unknown (90.4%) casual partners, than with a steady partner (56.2%; p<0.001). Median TFV-DP concentration was 528 fmol/punch (IQR 232-900; levels ≥700 fmol/punch are correlated with use of at least 4 pills per week in the preceding 6 weeks) and higher TFV-DP concentration was associated with the number of self-reported CAS acts (β=0.15, 95% CI 0.11-0.19) and with the number of pills taken (β=0.08, 95% CI 0.06-0.09) in the 6 weeks before the DBS. Conclusion: In our PrEP demonstration study, the majority of reported CAS acts were covered by PrEP. Self-reported adherence to event-driven PrEP was very

CROI 2020 387