CROI 2015 Program and Abstracts

Abstract Listing

Poster Abstracts

Conclusions: This is the first study to ever examine the correlation between concentrations of PrEP medications in hair samples and DBS, matrices in which long-term exposure to PrEP are captured. Strong correlations were observed between drug levels in hair and DBS. Further study is needed to compare the elimination kinetics and variability of TFV and FTC in hair versus TFV-DP and FTC-TP in DBS to guide the use of these measures for quantifying cumulative versus recent adherence in PrEP trials and real-world settings.

TUESDAY, FEBRUARY 24, 2015 Session P-H2 Poster Session

Poster Hall

2:30 pm– 4:00 pm Pharmacogenomics 515 UGT1A1 Genotype Predicts Bilirubin-Related Discontinuation of Atazanavir/Ritonavir Saran Vardhanabhuti 1 ; Heather J. Ribaudo 1 ; Raphael J. Landovitz 2 ; Igho Ofotokun 3 ; Jeffrey L. Lennox 3 ; Judith S. Currier 2 ; Lana M. Olson 4 ; DavidW. Haas 4 1 Harvard School of Public Health, Center for Biostatistics in AIDS Research, Boston, MA, US; 2 UCLA Center for Clinical AIDS Research & Education, Los Angeles, CA, US; 3 Emory University School of Medicine, Atlanta, GA, US; 4 Vanderbilt University School of Medicine, Nashville, TN, US Background: Atazanavir (ATV) increases direct bilirubin (bili) levels which may cause jaundice. Concern about this deters initial prescribing of ATV. AIDS Clinical Trials Group (ACTG) protocol A5257 showed inferior tolerability of ATV/ritonavir (/r) versus both darunavir (DRV)/r and raltegravir (RAL), all with tenofovir/emtricitabine, as initial HIV therapy. We characterized genetic and clinical determinants of premature ATV/r discontinuation for hyperbilirubinemia through 96 weeks, and peak total bili through 24 weeks in A5257. Methods: Reason for discontinuation was based on site investigator report. Genotypes for rs887829 were by HumanCore Exome Chip. Associations for bili-related ATV/r discontinuation were assessed with a competing risks framework and Cox proportional hazard models. Linear regression models quantified peak total bili associations with UGT1A1 rs887829, baseline indirect bili, and baseline hemoglobin. Results: Primary analyses included 481 subjects (211 Black, 183 White, 87 Hispanic) who initiated randomized ATV/r and consented for genetic testing; secondary analyses also included 491 subjects initiating DRV/r and 478 initiating RAL. In the ATV/r group, 14% of subjects were homozygous for rs887829 T/T (19% of Blacks, 8% of Whites, 16% of Hispanics). ATV/r was discontinued for any toxicity or non-toxicity reason in 30%, 8% bili-related. Bili-related ATV/r discontinuation was associated with rs887829 T/T ( Figure ). Positive predictive values of rs887829 T/T for bili-related ATV/r discontinuation were 20% in Blacks; 60% in Whites, 29% in Hispanics; Negative predictive values were 97%, 95% and 97%, respectively. With rs887829 T/T, median peak total bili was 4.9 mg/dL (IQR=2.0-7.2 mg/dL) in Blacks, 4.6 mg/dL (IQR=1.3-7.4 mg/dL) in Whites, and 3.4 (IQR=0.9-4.7 mg/ dL) in Hispanics. In multivariable models, rs887829 T/T and baseline bili were independently associated with peak total bili, but only rs887829 T/T was associated with bili-related ATV/r discontinuation. In subjects with rs887829 C/C genotypes, ATV/r tolerability was non-inferior to DRV/r, but was still inferior to RAL.

Poster Abstracts

Conclusions: Without rs887829 T/T, likelihood of bili-related ATV/r discontinuation was low. With rs887829 T/T, likelihood of bili-related ATV/r discontinuation varied by race/ ethnicity, despite similar bili levels. Genetic testing to avoid ATV/r prescribing with rs887829 T/T would improve ATV/r tolerability and reduce likelihood of bili-related ATV/r discontinuation.

338

CROI 2015