CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

not other symptoms. Outcome was available in 57 participants: within 30 days a complete (61.4%) or partial (33.3%) improvement in symptoms was reported. Headache (p=0.039) and sleep disorders (p=0.083) were associated with complete resolution of symptoms. Patients were switched to raltegravir (30.3%), elvitegravir/cobicistat (28.8%), darunavir/cobicistat (25.8%) or rilpivirine (12.1%)-containing regimens. Partial/complete regression of NPS was observed in 66.7%/33.3%(DRV/c), 43.5%/52.2% (RA), 12.5%/62.5% (RPV) and 11.1%/72.2% (EVG/c) (Chi2 1.99, p=0.158). Conclusion: In our cohort study 11.7% of participants stopped DTG due to NPS: they were mostly sleep disorders and headache and a full regression of both was observed after switching to other drugs. In most cases a complete resolution of symptoms was observed; the incomplete resolution in almost one third of participants suggests alternative reasons. 443 LONG-TERM ADHERENCE MONITORING OF EMTRICITABINE IN HAIR BY MASS SPECTROMETRY IMAGING Elias Rosen 1 , WilliamM. Gilliland 1 , Nicole White 1 , Amanda Poliseno 1 , Heather M. Prince 1 , Angela Kashuba 1 1 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Background: Adherence to antiretroviral therapy is critical for effective treatment and prevention of HIV. Incorporation of drug in hair creates a long- term record of adherence behavior, and mass spectrometry imaging (MSI) using infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI) offers a means of monitoring changes in this record as an objective measure of adherence. Here we benchmark longitudinal profiles of emtricitabine (FTC) in hair strands using IR-MALDESI MSI following directly observed therapy (DOT). We also identify adherence thresholds and develop a tool for real-time evaluation of patient adherence in a clinical setting. Methods: Benchmarking was performed in hair samples cut close to the scalp from 12 volunteers undertaking 28-day phases of daily and then differentiated (0x, 1x, or 3x/wk; n=4 in each group) tenofovir+FTC dosing as part of a DOT study (NCT03218592). The proximal 2cm (~2 months growth) of hair strands (n=4) were collected on day 28 of each phase and fixed to glass slides with double-sided tape before analysis with an IR-MALDESI source coupled to a Thermo QE+mass spectrometer. MSI data were processed with MSiReader, and longitudinal profiles of FTC over time were generated using customMatlab software. Quantification of the response was performed based on calibration from blank strands incubated in an FTC solution (lower limit of quantitation: 0.27 ng/mg hair; relative standard deviation (RSD): 20%). Results: FTC was measured in strands from 11 volunteers (1 reported recent hair salon treatment). A representative IR-MALDESI image (Fig. A; daily-to-3x/ wk dosing) shows distinct and localized bands of FTC in each strand associated with proximal differentiated dosing and distal daily dosing periods. Delineating these periods across all study samples, we evaluated average IR-MALDESI response associated with each dosing frequency (Fig. B). While interindividual variability in hair accumulation was observed (daily dosing RSD = 69%), an adherence cutpoint with high sensitivity (81%) and selectivity (100%) was derived from these data based on a receiver operating characteristic curve (Fig. C). A cutpoint-based analysis tool was developed to classify daily adherence in FTC longitudinal profiles (Fig. D). Conclusion: Longitudinal dose granularity for FTC can be visualized in hair strands by IR-MALDESI MSI. This approach provides a non-invasive long-term, daily adherence report for patients and clinicians, applicable to both treatment and prevention efforts.

444 ADHERENCE BY DBS, SELF-REPORT, AND PILL COUNT IN YOUNG ADULTS WITH PERINATAL HIV Nadia Nguyen 1 , Reuben Robbins 1 , Jose R. Castillo-Mancilla 2 , Peter L. Anderson 2 , Elaine J. Abrams 3 , AndrewWiznia 4 , Ira B. Wilson 5 , Curtis Dolezal 1 , Seth Kalichman 6 , Rehema Korich 1 , Jeannette Raymond 1 , Claude A. Mellins 1 1 New York State Psychiatric Institute, New York, NY, USA, 2 University of Colorado Denver, Denver, CO, USA, 3 ICAP at Columbia University, New York, NY, USA, 4 Albert Einstein College of Medicine, Bronx, NY, USA, 5 Brown University, Providence, RI, USA, 6 University of Connecticut, Storrs, CT, USA Background: Concentrations of tenofovir-diphosphate (TFV-DP) in dried blood spot (DBS) can objectively measure ART adherence and predict viral suppression. We explored if TFV-DP, self-reported adherence, and unannounced phone pill counts were associated with viral suppression needed for individual health (HIV viral load [VL]<20 copies/mL) and treatment as prevention (TasP) in the context of Undetectable=Untransmittable (U=U, VL<200) in young adults living with perinatal HIV infection (YAPHIV). Methods: We quantified adherence using TFV-DP and emtricitabine- triphosphate (FTC-TP) in DBS, self-report, and pill counts, and concomitant VL in YAPHIV from New York City, 18-28 years, receiving tenofovir-based regimens. Self-reported adherence and pill counts were assessed using validated measures. Mean and median TFV-DP levels (by regimen: tenofovir disoproxil fumarate/emtricitabine [TDF]/FTC, tenofovir alafenamide [TAF]/FTC), self-reported adherence, and pill counts were estimated, stratified by VL level (<20 copies/mL or <200 copies/mL). Differences in mean adherence levels were assessed with t-tests. Predictive accuracy of each measure to predict VL<200 copies/mL was compared using Receiver Operator Characteristic analysis. ‘White coat adherence’ (lowest TFV-DP with quantifiable FTC-TP) was assessed. Results: Of 34 YAPHIV on TDF/FTC, 41% had VL<20 (mean TFV-DP 1293 fmol/ punch [95% CI 1059, 1580]), and 59% had VL<200 (mean TFV-DP 658 fmol/ punch [95% CI 356, 1218]) (Table). Of 27 on TAF/FTC, 52% had VL<20 (mean TFV-DP 2250 fmol/punches [95% CI: 1446, 3500]), and 63% had VL<200 (mean TFV-DP 1996 fmol/punches [95% CI (1340, 2974]). White coat adherence was detected in 1 person. Of YAPHIV with all three adherence measures (TFV-DP level, self-report, pill counts; n=42), TFV-DP from TAF/FTC (0.97 [95% CI: 0.89, 1.00; p=0.001]) had the highest area under the curve for predicting VL<200, followed by TFV-DP from TDF/FTC (0.87 [95% CI: 0.68, 1.00; p=0.007]), pill count (0.72 [95% CI: 0.53, 0.90; p=0.03]), and self-report (0.69 [95% CI: 0.53, 0.86; p=0.05]). Conclusion: TFV-DP concentrations required to prevent transmission (VL<200 copies/mL) are lower than for individual health (VL<20 copies/mL) and different for TDF vs. TAF, suggesting differential thresholds for TasP/U=U vs. individual health, while self-reported adherence and pill counts were less sensitive. All adherence measures were associated with HIV VL<200, making the choice of a specific method dependent on context, preference, and available resources.

Poster Abstracts

CROI 2020 155