Background: Tuberculosis (TB) patients with a single nucleotide polymorphism (SNP) of the drug transporter gene SLCO1B1 (C463A) have reduced oral bioavailability (F) of rifampin. The impact of this pharmacogenetic variability on TB treatment outcomes is uncertain. A recent rifampin population PK model for TB meningitis (TBM) by Savic et al. (2015) provided the opportunity to examine the effect of pharmacogenetics on rifampin concentrations in CSF.
Methods: We evaluated the population PK model for rifampin during TBM treatment with published CSF rifampin data that had not been used in model development (D'Oliveira et al. 1972), with a two-compartment distribution model, first-order absorption and clearance, and a CSF effect compartment. We performed a Monte Carlo simulation of rifampin CSF concentrations in TBM patients (n=10,000), with rifampin dose by weight bands based on WHO guidelines, and minimum inhibitory concentrations (MICs) sampled from the wild-type distribution. Based on published data, we introduced the SLCO1B1 SNP into the model as a 36% reduction in oral bioavailability. We calculated the AUC0-24/MIC ratio on the 7th day of treatment, with a target AUC0-24/MIC of 271.
Results: We observed agreement between predicted and observed rifampin CSF concentrations (Figure 1). Across all TBM patients with the high-F SLCO1B1 genotype, a mean of 77% reached the target, compared with 50% among patients with low-F genotype. At an MIC of 0.25 mg/L, the AUC target in CSF was reached in 79% of patients with high-F genotype, compared with 15% of patients with low-F genotype (Figure 2).
Conclusion: We observed a difference in target attainment by SLCO1B1 genotype across a range of MICs below the breakpoint of 1.0 ug/mL. Next, we will simulate rifampin dosing strategies for TBM treatment, to inform clinical trial design.
Figure 1. Evaluation of Savic et al population PK model with previously published rifampin CSF data.
Legend: Triangle = Observations; Dash = 95th percentile predicted conc; Solid = 50th percentile predicted conc; Dot = 5th percentile predicted conc.
Figure 2. Probability of target attainment based on pharmacogenetics variability in SLCO1B1.
Legend: Solid = Low F genotype; Dash = High F genotype
K. Thakur, None
V. Ivaturi, None
C. Vinnard, None
See more of: Poster Abstract Session