CROI 2024 Abstract eBook

Abstract eBook

Poster Abstracts

870

Pharmacogenetic Associations With HIV-1 Virologic Suppression Among Patients with TB/HIV in Brazil Felipe Ridolfi 1 , Gustavo Amorim 1 , David W. Haas 1 , Maria B. Arriaga 1 , Cody Staats 1 , Marcelo Cordeiro-Santos 2 , Afrânio L. Kritski 3 , Marina C. Figueiredo 1 , Bruno B. Andrade 1 , Timothy R. Sterling 1 , Valeria C. Rolla 4 1 Vanderbilt University, Nashville, TN, USA, 2 Fundação de Medicina Tropical Doutor Heitor Vieira Dourado, Manaus, Brazil, 3 Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil, 4 Instituto Nacional de Infectologia Evandro Chagas, Rio de Janeiro, Brazil Background: Human genetic variants can affect TB and HIV drug pharmacokinetics (PK), which may affect risk for toxicity or treatment failure. Here we evaluated associations between pre-specified genetic variants and HIV virologic suppression (VS) among patients treated for TB and HIV. Methods: We included TB/HIV RePORT-Brazil participants who initiated standard TB treatment [2 months of isoniazid/rifampicin (or rifabutin)/ pyrazinamide/ethambutol, then 4 months or more of isoniazid/rifampicin (or rifabutin)], and who also received antiretroviral therapy (ART) during this period. The outcome was HIV-1 VS (for this study we considered as <50 HIV-1 RNA c/mL) after at least 2 weeks of ART. Regimens were categorized as containing an integrase strand transfer inhibitor (INSTI) or non-nucleoside reverse transcriptase inhibitor (NNRTI). We genotyped UGT1A1 rs887829 that affects dolutegravir (DTG) and raltegravir (RAL) PK, and CYP2B6 rs3745274, rs28399499, rs4803419 that affect efavirenz (EFV) PK; all have defined normal, intermediate, and poor metabolizer groups. Genotyping was by MassARRAY iPLEX Gold. We compared outcome proportions (Fisher's test) and time-to-VS (survival analysis, Wilcoxon-Gehan test). Results: Among 194 TB/HIV participants included, 88 (45%) achieved VS. RAL was the most frequent INSTI (n=88, 88%), and EFV the most frequent NNRTI (n=76, 99%; and one participant used etravirine). In the INSTI group, similar proportions of VS were achieved for UGT1A1 normal (n=16, 39%) and intermediate (n=17, 41%) genotypes. Among participants receiving EFV, those who achieved VS were more likely to be CYP2B6 intermediate metabolizers (n=23, 70%). There were inconclusive associations comparing the proportions of VS among INSTI- and EFV-based ART, and based on CYP2B6 and UGT1A1 genotypes (Table 1). Furthermore, no consistent associations were found comparing the time-to-VS among ART regimens and genotypes. Conclusion: In this cohort of patients treated for TB/HIV, genetic variants that affect ART PK were not significantly associated with likelihood of VS.

869

Estimating Optimal Anti-TB Drug Concentrations in a Prospective, Observational Cohort in Brazil Gustavo Amorim 1 , David W. Haas 1 , Marina C. Figueiredo 1 , Cody Staats 1 , Marcelo Cordeiro-Santos 2 , Afrânio L. Kritski 3 , Brian C. Hachey 1 , Bruno B. Andrade 1 , Timothy R. Sterling 1 , Valeria C. Rolla 4 , for the RePORT-Brazil Consortia 1 Vanderbilt University, Nashville, TN, USA, 2 Fundação de Medicina Tropical Doutor Heitor Vieira Dourado, Manaus, Brazil, 3 Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil, 4 Oswaldo Cruz Foundation - Fiocruz, Rio de Janeiro, Brazil Background: Tuberculosis (TB) treatment is highly effective, but in real-world settings, >10% of TB patients experience drug toxicity or treatment failure. Current target drug levels of standard anti-TB medications (isoniazid (INH), rifampin (RIF), pyrazinamide (PZA), and ethambutol (EMB)) are based on expected levels two hours after dosing. We assessed a large Brazilian cohort of pulmonary TB patients undergoing anti-TB therapy for concomitant optimal drug levels associated with minimal toxicity and maximum effectiveness. Methods: Plasma drug levels from participants enrolled in Regional Prospective Observational Research in Tuberculosis (RePORT)-Brazil were used to estimate therapeutic drug range for all four anti-TB drugs. Pharmacokinetic (PK) analyses using drug concentrations and time since last anti-TB dose were performed using Extreme Gradient Boosting. The PK model was built using 10-fold cross-validation, adjusting for sex, BMI, HIV positivity, smoking and alcohol use, substance use, HbA1c, ancestry informative genetic markers, directly observed therapy assigned at baseline, and NAT2 acetylator status. Maximum drug concentrations (C max ) computed from individual PK profiles were tested for correlation with TB treatment outcomes: toxicity (Grade ≥ 3 adverse events) and treatment failure or recurrence. Safety and effectiveness bounds were defined as drug concentrations that would lead to probabilities of toxicity or treatment failure/recurrence of no more than 5%, respectively. Confidence intervals were computed after 999 bootstrap resamples. Therapeutic drug range was defined as concentrations that were both safe and effective. Results: There were 966 plasma samples from 459 participants. Overall, 11 (2.4%) experienced toxicity after month 1, and 13 (2.8%) had treatment failure or TB recurrence. Log-transformed C max for RIF was associated with greater odds of toxicity: Odds Ratio (OR) = 12.9 [95% CI=3.5-47.1]. Log-transformed C max for INH, EMB, and PZA were associated with decreased odds of failure/recurrence: OR = 0.5 [0.2-0.9], OR = 0.6 [0.4-1.0], and OR = 0.4 [0.2-0.8], respectively. Therapeutic drug ranges for all four anti-TB drugs are shown in the Table. Conclusion: Our findings suggest target Cm a x values, particularly for INH and RIF, that differ somewhat from currently recommended targets. Further studies with less variation in PK profiles are still needed to compute optimal targets for concentration ranges.

Poster Abstracts

871

Pharmacokinetics of Isoniazid Metabolites During Pregnancy and Postpartum Brandon Klein 1 , Zixuan Wei 1 , David Nerguizian 1 , Amita Gupta 2 , Adriana Weinberg 1 , Grace Montepiedra 3 , Mary Morrow 1 , Samantha MaWhinney 1 , Philippa Musoke 4 , Linda Aurpibul 5 , Gaerolwe Masheto 6 , Farah Abdelmawla 1 , Lane Bushman 1 , Peter L. Anderson 1 , Kristina M. Brooks 1 1 University of Colorado Anschutz Medical Campus, Aurora, CO, USA, 2 The Johns Hopkins University, Baltimore, MD, USA, 3 Harvard TH Chan School of Public Health, Boston, MA, USA, 4 Makerere University, Kampala, Uganda, 5 Chiang Mai University, Chiang Mai, Thailand, 6 Botswana Harvard AIDS Institute Partnership, Gabarone, Botswana Background: IMPAACT P1078 evaluated the safety of isoniazid (INH) preventative therapy initiated antepartum (AP) or postpartum (PP) in women with HIV (WWH). Hepatotoxicity occurred at high rates (~6-7%) across both arms during the PP period and a higher risk of adverse pregnancy outcomes was also identified among N-acetyltransferase type 2 (NAT2) slow acetylators. These adverse outcomes may be due in part to toxic INH metabolites, such as hydrazine (Hz), which have not been previously evaluated in pregnancy. INH can either be converted to acetylisoniazid (AcINH) via NAT2 and then acetylhydrazine (AcHz), or to isonicotinic acid (INA) via amidases, with both pathways contributing to Hz

CROI 2024 269