Background:  Atazanavir (ATV) pharmacokinetics (PK) are influenced by transporters and enzymes that are regulated by the pregnane -X- receptor (PXR; NR1I2). PXR is a ligand-activated receptor, but its expression is also correlated with CYP3A4 in liver in the absence of enzyme inducers. The PXR 63396C>T single nucleotide polymorphism (SNP) alters PXR expression and CYP3A4 activity in vitro, and we showed previously an association with unboosted ATV plasma concentrations. Since ritonavir is thought to activate PXR, the aim of this study was to develop a population PK analysis to assess whether 63396C>T was also associated with boosted ATV clearance (CL) and simulated trough concentrations (C_trough).

Methods:  A population analysis was performed with 769 plasma samples from 352 patients receiving boosted ATV. Genotyping was conducted on 91.5% of the patients using standard methodology. Three-hundred-twenty-two of the blood samples were collected at random time points and 30 patients had a full concentration time profile at steady state. Non-linear mixed effects modelling was applied (NONMEM) to explore the effects of PXR 63396C>T and patient demographics. The model was validated by means of simulation and visual predictive check. Simulations of ATV PK profiles were performed to calculate 90% prediction intervals (90% PI) for different genotypes.

Results:  Of the 352 patients included in the study, 73% were male, 83% were receiving tenofovir and the median (range) body weight and age were 70.3 (46 to 110) kg and 41 (18 to 74) years, respectively. A one-compartment model with first-order absorption best described the data. Population clearance was 6.2 L/h with inter-patient variability of 18%. Homozygosity for the T allele was associated with a 26.8% higher CL that was statistically significant. Patient demographic factors had no effect on the CL. Simulated C_trough were lower in the TT group (90% PI = 0.04 to 2.54) compared to other patients (90% PI = 0.056 to 3.3); 21% of T homozygotes had simulated C_trough below the proposed MEC (0.15 mg/mL) compared to 14% of the remaining patients.

Conclusion:  These data show an association between PXR 63396C>T and boosted ATV CL. This polymorphism tags other SNP in putative response elements in the PXR regulatory region and is therefore likely to be associated with baseline PXR expression and baseline expression of its target genes (eg, CYP3A4 and ABCB1). These data add to our understanding of the pharmacogenetic determinants of variability in ATV PK.