Effect of ABCG2 c.421C>A variant on exposure to MPA
In estimating the effect of ABCG2 c.421C>A variant, we controlled for a number of potentially interfering factors (Table 3). Before matching, variant and wt patients differed considerably in respect to most of the matching variables (Table 4), while total exposure (AUCτ,ss) appeared higher and total body clearance (CLT/F,ss) appeared lower in variant carriers than in wt controls (Table 4). Frequentist and Bayesian estimates suggested by 45% (95%CI 10-92) and by 38% (95%CrI 7-81) higher total exposure, respectively (Table 5), and by 34% (10-52) and 29% (5-47) lower total body clearance, respectively (Table 5) in variant carriers than in wt controls. Eventually, 11/12 variant carriers were matched to 43/56 wt controls with excellent balance regarding most matching covariates except (d ≥0.1) for the donors’ ABCC2 1249 G>A genotype, body mass index and estimated creatinine clearance (Table 4): average concentration-time profiles were not much changed vs. raw data (Figure 1B), AUCτ,ss was still higher (d=0.824) and CLT/F,ss was lower (d=-0.559) in variant carriers than in wt controls (Table 4). With additional adjustment for suboptimally matched covariates, frequentist and Bayesian estimates indicated by 41% (95%CI 11-79) and by 39% (95%CrI 5-81), respectively, higher total exposure, and by 27% (7-42) and by 29% (5-46) lower total body clearance, respectively, in variant carriers than in wt controls (Table 5). Probability that the GMR for AUCτ,ss was >1.20 was 90.7% based on the frequentist analysis and it was 85.6% based on the Bayesian analysis (Figure 2).