CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

463 PHARMACOLOGIC MEASURES OF PrEP ADHERENCE AMONG HIGH-RISK MSM IN HPTN 067 Jennifer Velloza 1 , Monica Gandhi 2 , Craig W. Hendrix 3 , Peter Bacchetti 2 , Pamela Murnane 2 , James P. Hughes 1 , Maoji Li 4 , Marcel Curlin 5 , Timothy H. Holtz 6 , Sharon Mannheimer 7 , Mark A. Marzinke 3 , K. Rivet Amico 8 , Estelle Piwowar-Manning 3 , Susan H. Eshleman 3 , Robert M. Grant 9 1 University of Washington, Seattle, WA, USA, 2 University of California San Francisco, San Francisco, CA, USA, 3 Johns Hopkins Hospital, Baltimore, MD, USA, 4 Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 5 Oregon Health and Sciences University, Portland, OR, USA, 6 CDC, Atlanta, GA, USA, 7 Columbia University, New York, NY, USA, 8 University of Michigan, Ann Arbor, MI, USA, 9 Gladstone Institute of Virology and Immunology, San Francisco, CA, USA Background: The effectiveness of oral emtricitabine (FTC)/tenofovir (TFV) disoproxil-fumarate pre-exposure prophylaxis (PrEP) is highly dependent on adherence. Given limitations of self-reported adherence, pharmacologic measures are useful for understanding patterns of adherence and identifying predictors of consistent PrEP use. Methods: We analyzed data from HPTN 067, a trial of intermittent and daily PrEP completed in 2014. We included men who have sex with men (MSM) in Bangkok, Thailand, and Harlem, United States. Participants were randomly assigned to daily, time-driven (one dose twice/week plus one dose after sex), or event-driven (one dose before and after sex) oral PrEP regimens. Study visits occurred at weeks 0, 4, 12, and 24 post-randomization to assess FTC and TFV levels. Plasma and hair samples measured short and long-term drug exposure, respectively. Electronic pill bottle data (Wisepill™) were collected weekly. Potential predictors of adherence were measured at baseline (e.g., demographics, alcohol use) and longitudinally (e.g., sexual behavior). We estimated Pearson correlation coefficients among measures and assessed predictors of log-transformed plasma and hair drug levels using linear mixed models. Pharmacologic measures below the detection limit were set equal to that limit prior to log transformation. Results: Among the 350 randomized MSM at the two sites, half had completed college (N=179; 51.1%) and the median age was 31 years (IQR: 25–38 years). At baseline, 20.6% reported heavy alcohol use (N=72) and the median number of sex partners in the prior three months was 4 (IQR: 2–8). Across all arms and follow-up periods, FTC and TFV hair concentrations were moderately correlated with plasma concentrations and Wisepill™ data (Pearson coefficients >0.28). In multivariate models, being enrolled at the Harlem site, being in the time- or event-driven arms, and having less than college education were associated with lower hair FTC/TFV levels, while heavy alcohol use was associated with higher hair levels (Table). In models evaluating plasma levels, similar results were seen for site and study arm, but older age and greater number of sex partners were significantly associated with higher plasma drug levels. Conclusion: In HPTN 067, plasma and hair drug concentrations and Wisepill™ data correlated with one another and served as complementary measures of PrEP adherence. Site, study arm, education, age, alcohol use, and sexual behavior influenced patterns of short and long-term PrEP adherence.

462 EFAVIRENZ LEVEL IN HAIR PREDICTS VIROLOGIC RESPONSE BETTER THAN LEVEL IN BLOOD Laura Dickinson 1 , Marco Siccardi 1 , Kathryn Anastos 2 , Mardge H. Cohen 3 , Deborah Gustafson 4 , Gerald Sharp 5 , Stephen J. Gange 6 , Seble Kassaye 7 , Peter Bacchetti 8 , Ruth Greenblatt 8 1 University of Liverpool, Liverpool, UK, 2 Montefiore Medical Center, Bronx, NY, USA, 3 Stroger Hospital of Cook County, Chicago, IL, USA, 4 SUNY Downstate Medical Center, Brooklyn, NY, USA, 5 NIAID, Bethesda, MD, USA, 6 Johns Hopkins University, Baltimore, MD, USA, 7 Georgetown University, Washington, DC, USA, 8 University of California San Francisco, San Francisco, CA, USA Background: Levels of ARV in blood reflect recent exposure (hours to days), but the outcomes of cART are based on exposure over months. We hypothesized that levels of efavirenz (EFV) in hair, which reflects exposure occurring over several weeks would better predict virologic response than single blood levels. Methods: Intensive PK (iPK), sparse blood and hair samples (1 cm of scalp) were collected from participants of the Women’s Interagency HIV Study who reported current use of EFV. EFV levels were measured in blood and hair by validated liquid chromatography tandemmass spec method. Population PK models were developed including clinical covariates previously shown to influence exposure (area under the curve, AUC). AUC was estimated from sparse blood levels and time of last dose using NONMEMmethods and compared directly to level in hair (also an indicator of exposure) using random effects multivariate logistic regression. Results: AUC by iPK was typically about 50% greater than that estimated from sparse sample, though the values were correlated (median ratio of AUC by iPK to AUC by sparse = 1.47, correlation coef =0.83, p<.0001, n=92). When AUC estimated from sparse blood samples were compared to levels measured in hair as predictors of virologic suppression (in models of 1071 observations made of 315 individuals that adjusted for adherence, decade of age, ethnicity and individual) log level in hair was a statistically significant predictor of suppression, while AUC estimated from sparse samples was not (Table below). Conclusion: Exposure to EFV estimated from hair samples better predicted virologic response than exposure estimated from single blood specimens and time since dose. These findings support the utility of measures of long term exposure in predicting treatment responses, which may be important for research and in the management of patients who are at high risk for treatment failure.

Poster Abstracts

CROI 2018 165