Background: Atazanavir-associated hyperbilirubinemia can cause premature discontinuation of atazanavir and avoidance of its initial prescription. We used genomewide genotyping and clinical data to characterize determinants of atazanavir pharmacokinetics and hyperbilirubinemia in AIDS Clinical Trials Group protocol A5202.
Methods: Plasma atazanavir pharmacokinetics and indirect bilirubin concentrations were characterized in HIV-1-infected patients randomized to atazanavir/ritonavir-containing regimens. A subset had genomewide genotype data available.
Results: Genomewide assay data were available from 542 participants, of whom 475 also had data on estimated atazanavir clearance and relevant covariates available. Peak bilirubin concentration and relevant covariates were available for 443 participants. By multivariate analysis, higher peak on-treatment bilirubin levels were found to be associated with the UGT1A1 rs887829 T allele (P=6.4×10(-12)), higher baseline hemoglobin levels (P=4.9×10(-13)), higher baseline bilirubin levels (P=6.7×10(-12)), and slower plasma atazanavir clearance (P=8.6×10(-11)). For peak bilirubin levels greater than 3.0 mg/dl, the positive predictive value of a baseline bilirubin level of 0.5 mg/dl or higher with hemoglobin concentrations of 14 g/dl or higher was 0.51, which increased to 0.85 with rs887829 TT homozygosity. For peak bilirubin levels of 3.0 mg/dl or lower, the positive predictive value of a baseline bilirubin level less than 0.5 mg/dl with a hemoglobin concentration less than 14 g/dl was 0.91, which increased to 0.96 with rs887829 CC homozygosity. No polymorphism predicted atazanavir pharmacokinetics at genomewide significance.
Conclusion: Atazanavir-associated hyperbilirubinemia is best predicted by considering UGT1A1 genotype, baseline bilirubin level, and baseline hemoglobin level in combination. Use of ritonavir as a pharmacokinetic enhancer may have abrogated genetic associations with atazanavir pharmacokinetics.