CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

refill data predicts ARV detection in plasma and viral suppression (VS) in a routine care cohort in Cape Town, South Africa. Methods: HIV+ women who initiated TDF+XTC+EFV during pregnancy and achieved VS (≤50 cps/mL) were followed for up to 24 months. Plasma viral load and presence of ARV (>20 ARVs tested for using mass-spectrometry) were measured at multiple study visits. Patient-level routine pharmacy data were used to classify each visit as: having no ARVs in hand (i) today, (ii) for >30 days, or (iii) >90 days prior. Generalised estimating equations were used to calculate associations between ARV in hand, VS, and detectable ARV in plasma. Secondary analyses were restricted to a) women who stayed in one of three large clinics to minimize heterogeneity in routine data, and b) the first visit to calculate diagnostic characteristics. Results: Across 237 women and 417 visits (median 10 months on ART, IQR=7- 14) 46%were not VS. Any ARV was detected in plasma at 60% of visits, of which EFV was detectable in 98%, TFV in 65% and FTC in 73% of visits. Patients were classified as having no ARV in hand at 56% of visits, with 81% and 63% of these having no ARV in hand for >30 and >90 days, respectively. Absence of any ARV in plasma was strongly associated with viraemia (OR 70.6, 95% CI 35.7-139.6). No ARVs in hand today (OR 7.7, 95% CI 4.7-12.6), for >30 days (OR 15.0, 95% CI 6.5-25.3) and >90 days (OR 19.7, 95% CI 10.8-35.6) were also associated with viremia; similar associations were observed between drugs in hand and plasma ARV detection (Table). Associations with VS, but not plasma ARVs, strengthened when restricted to women who were in care at one of three large clinics. At the first visit, increasing time with no ARV in hand resulted in decreased sensitivity (VL 76% to 33%; plasma ARV 83% to 35%) and increased specificity (VL 67% to 92%; plasma ARV 68% to 92%). Conclusion: Although ARV detection in plasma was the best predictor of virologic outcomes, having ARV in hand was a strong predictor of VS and presence of ARV. Routine pharmacy data provides a feasible, inexpensive alternative objective measure of ART adherence for public sector programme evaluation in high-burden settings.

1083 VIRAL SUPPRESSION TRAJECTORIES AMONG HIGH-NEED PATIENTS IN LOW-BARRIER HIV CARE Julia C. Dombrowski 1 , Meena Ramchandani 1 , McKenna C. Eastment 1 , Matthew R. Golden 1 1 University of Washington, Seattle, WA, USA Background: Ending the HIV epidemic will require intensive efforts to sustain viral suppression among persons with HIV who have complex medical and social barriers to care. We previously showed that a clinic offering walk-in, incentivized care (the Max Clinic) improves viral suppression. Here we examine viremia trajectories among Max Clinic patients. Methods: We included patients with ≥180 days of observation time after enrollment during Dec 2014-Jan 2019, starting on the enrollment date and ending at the time of death, relocation, or July 31, 2019. We categorized patients into groups defined a priori based on knowledge of common viremia trajectories: 1) early consistent suppression [first viral load (VL)<200 copies/mL ≤6 months (mo) after enrollment; all subsequent VL<200]; 2) late consistent suppression (first VL<200 >6 mo after enrollment, all subsequent VL<200), 3) transient/intermittent suppression (≥1 VL<200, subsequent VL>200); and 4) no suppression (no VL<200). We compared the characteristics of patients in each group using χ2 tests for categorical variables and Kruskal-Wallis One-way ANOVA for continuous variables. Results: Among 167 patients with a median observation time of 27 mo [interquartile range (IQR): 16-39 mo], 69%were homeless or unstably housed at enrollment, 54% used methamphetamine, 51% injected drugs, and 32% had a diagnosed psychotic, bipolar or personality disorder. Most patients (59%) had transient/intermittent suppression, followed by early consistent suppression (26%), no suppression (10%) and late consistent suppression (5%). The groups differed by the median observation time, which was shorter in the no suppression (15 mo) and early consistent suppression (20 mo) groups than in the transient/intermittent suppression (34 mo) and late consistent suppression (37 mo) groups (p<0.001). The groups did not differ significantly by gender, race, ethnicity, housing status, substance use or depression/anxiety disorder diagnoses. Patients with psychotic, bipolar disorder or personality disorder were less likely to be in the late or no suppression groups (p<0.04). The median time from the first suppressed VL to a subsequent unsuppressed VL was 4 mo (IQR 2-10 mo). Conclusion: The vast majority of patients in the low-barrier clinic reach viral suppression even in the context of unstable housing, substance use, or severe mental illness, but most are intermittently unsuppressed. Even with low-barrier care and high-intensity support, most patients continue to have periods of viremia. 1084 ROUTINE PHARMACY REFILLS PREDICT WOMEN’S PLASMA ARV DETECTION AND VIRAL SUPPRESSION Tamsin K. Phillips 1 , Andrew D. Redd 2 , Steven J. Reynolds 2 , Nei-Yuan M. Hsiao 1 , Gary Maartens 1 , Elaine J. Abrams 3 , Landon Myer 1 1 University of Cape Town, Cape Town, South Africa, 2 Johns Hopkins University, Baltimore, MD, USA, 3 Columbia University, New York, NY, USA Background: Detection of antiretrovirals (ARV) is an objective adherence measure that predicts HIV treatment outcomes, however, routine ARV testing is currently not feasible in high-burden settings. We examined how pharmacy

Poster Abstracts

1085 HIGHER INCOME INEQUALITY IS ASSOCIATED WITH LOWER CUMULATIVE ANTIRETROVIRAL ADHERENCE Frances Vernon 1 , Mary Morrow 1 , Samantha MaWhinney 1 , Ryan P. Coyle 1 , Stacey Coleman 2 , Lucas Ellison 1 , Jia-Hua Zheng 1 , Lane R. Bushman 1 , Jennifer J. Kiser 1 , Omar Galarraga 3 , Peter L.Anderson 1 , Jose R. Castillo-Mancilla 1 1 University of Colorado Anschutz Medical Campus, Aurora, CO, USA, 2 Duke University, Durham, NC, USA, 3 Brown University, Providence, RI, USA Background: Tenofovir diphosphate (TFV-DP) in dried blood spots (DBS) is associated with viral suppression and predicts future viremia. However, its association with the social determinants of health (SDoH) in people living with HIV (PLWH) has not been evaluated. Methods: DBS for TFV-DP were prospectively collected from a clinical cohort of PLWH receiving tenofovir disoproxil fumarate (TDF)-based therapy (up to 3 visits over 48-weeks between 2014 and 2017). Zip code was collected at enrollment and matched with the relevant SDoH data from 2016 obtained from AIDSVu (aidsvu.org). SDoH data included household income, percent living in poverty, education level and income inequality (the latter was quantified using the Gini coefficient, where 0 and 1 represent absolute income equality and inequality,

CROI 2020 408