Search

The narrow therapeutic window of tacrolimus necessitates daily monitoring and predictive algorithms based on genetic and nongenetic factors. In this study, we constructed predictive algorithms for tacrolimus stable dose in a retrospective cohort of 1045 Chinese renal transplant recipients. All patients were genotyped for CYP3A4 20230T>C (rs2242480), CYP3A4 T>C (rs4646437), CYP3A5*3 6898A>G (rs776746), ABCB1 129T>C (rs3213619); ABCB1 c.1236C>T (rs01128503), ABCB1 c.2677G>T/A (rs2032582) and ABCB1 c.3435C>T (rs1045642) polymorphisms, and the effects of gene-gene and gene-environment interactions on the predictive accuracy of algorithm were evaluated. In wild-type CYP3A4 rs2242480 (TT) carriers, patients who took calcium channel blockers had lower tacrolimus stable doses than those without the concomitant medications (P < 1 × 10 -4 ). In contrast, there was no significant difference in mutant type patients. Similarly, the tacrolimus stable doses in wild-type CYP3A5 rs776746 carriers who had hypertension were higher than those without hypertension (P = 4.10 × 10 -3 ). More importantly, dose-predictive algorithms with interaction terms showed higher accuracy and better performance than those without interaction terms. Our finding suggested that wild-type CYP3A4 rs2242480 (TT) carriers should be more cautious to take tacrolimus when they are coadministrated with calcium channel blockers, and CYP3A5 rs776746 (AA) carriers may need higher tacrolimus dosage when they are in combination with hypertension., (© 2019, The American College of Clinical Pharmacology.)

Tacrolimus has a narrow therapeutic window and considerable variability in clinical use. Our goal was to compare the performance of multiple linear regression (MLR) and eight machine learning techniques in pharmacogenetic algorithm-based prediction of tacrolimus stable dose (TSD) in a large Chinese cohort. A total of 1,045 renal transplant patients were recruited, 80% of which were randomly selected as the "derivation cohort" to develop dose-prediction algorithm, while the remaining 20% constituted the "validation cohort" to test the final selected algorithm. MLR, artificial neural network (ANN), regression tree (RT), multivariate adaptive regression splines (MARS), boosted regression tree (BRT), support vector regression (SVR), random forest regression (RFR), lasso regression (LAR) and Bayesian additive regression trees (BART) were applied and their performances were compared in this work. Among all the machine learning models, RT performed best in both derivation [0.71 (0.67-0.76)] and validation cohorts [0.73 (0.63-0.82)]. In addition, the ideal rate of RT was 4% higher than that of MLR. To our knowledge, this is the first study to use machine learning models to predict TSD, which will further facilitate personalized medicine in tacrolimus administration in the future., Competing Interests: The authors declare no competing financial interests.

Background: Whether the loss-of-function allele CYP3A5*3 variant is associated with significantly impaired metabolism of cyclosporine A (CsA) in transplant patients is still controversial because of the lack of prospective, large-scale clinical studies performed among diversely ethnic populations., Objectives: This meta-analysis was designed to determine whether the CYP3A5*3 variant could affect CsA blood concentrations and the rate of acute rejection in renal transplant recipients., Methods and Results: All relevant publications were retrieved online from 1966 to March 2010, in which 14 studies were chosen, and 1821 renal transplant patients were enrolled. The results showed that there were significant differences in the CsA dose-adjusted trough concentration (C0) between the CYP3A5*3/*3 and CYP3A5*1/*1 carriers [weighted mean difference (WMD): 10.06 mug/l per mg/kg, 95% confidence interval (CI): 3.12-17.00, P=0.004] and between the non-CYP3A5*1 allele carriers and the CYP3A5*1 allele carriers (WMD: 8.32 mug/l per mg/kg, 95% CI: 3.16-13.49, P=0.002). In addition, a subgroup analysis stratified by ethnicity indicated that a significant difference in CsA dose-adjusted C0 was observed between the non-CYP3A5*1 allele carriers and the CYP3A5*1 allele carriers in Asian patients, but not in Caucasian patients. Moreover, a significant difference in the mean daily dose was observed between the non-CYP3A5*1 allele carriers and the CYP3A5*1 allele carriers (WMD: -0.19 mg/kg, 95% CI: -0.31 to -0.07, P=0.002). However, the meta-analysis suggested that there was little or no association of the CYP3A5*3 variant with the acute rejection rate in renal transplant patients treated with CsA [odds ratio=0.94, 95% CI: 0.57-1.54, P=0.80]., Conclusion: We concluded that the CYP3A5*3 variant could be associated, to a certain extent, with increased CsA dose-adjusted C0 in blood and reduced mean daily doses, but that this genetic variant allele seemed to have little effect on the acute rejection rate in renal transplant patients taking CsA.

1. The calcineurin inhibitor cyclosporine is widely used to prevent allograft rejection after solid organ transplantation. It has a narrow therapeutic index and shows considerable interindividual differences in its pharmacokinetics. Interindividual differences in the activity and expression of the metabolising enzymes cytochrome P450 (CYP) 3A4 and 3A5 and the multidrug efflux pump P-glycoprotein (P-gp) contribute considerably to cyclosporine pharmacokinetics. Variability in the activity of CYP3A4, CYP3A5 and P-gp could be considered to result from genetic polymorphisms encoding their genes. 2. The aim of the present study was to evaluate retrospectively the effects of genetic polymorphisms of CYP3A4, CYP3A5 and MDR1 on cyclosporine dose adjusted trough blood concentration during the early period after renal transplantation in Chinese patients. 3. One hundred and six renal transplant recipients in China were genotyped by polymerase chain reaction-restriction fragment length polymorphism for CYP3A4*18A, CYP3A5*3 and MDR1 C3435T. Cyclosporine whole blood levels were measured by fluorescence polarization immunoassay. Dose-adjusted trough blood concentrations (C(0)) were determined and compared among the different genotype groups. 4. The frequency of the CYP3A4*18A, CYP3A5*3 and MDR1 C3435T variant alleles were 0.005 (95% confidence interval (CI) 0.048, 0.0049), 0.783 (95% CI 0.781, 0.785) and 0.528 (95% CI 0.526, 0.531), respectively, and these alleles exhibited incomplete linkage disequilibrium. The median cyclosporine dose-adjusted C(0) in CYP3A5*1/*1 genotype subjects (n = 6) was 14.8 ng/mL per mg per kg (range 11.1-26.8 ng/mL per mg per kg), in CYP3A5*1/*3 patients (n = 34) it was 23.7 ng/mL per mg per kg (range 9.0-61.0 ng/mL per mg per kg) and for CYP3A5*3/*3 patients (n = 66) it was 26.4 ng/mL per mg per kg (range 9.8-85.8 ng/mL per mg per kg; P = 0.012, Kruskal-Wallis test). Accordingly, cyclosporine dose-adjusted C0 was larger in CYP3A5 non-expressors than expressors in the first week after renal transplantation. In addition, wild-type homozygotes (n = 21) for MDR1 C3435T had a slight but significantly lower dose-adjusted C0 compared with heterozygotes (n = 58): 17.7 (10.3-60.8) versus 26.4 (9.0-67.3) ng/mL per mg per kg, respectively (P = 0.014, Mann-Whitney U-test). 5. In conclusion, the present study shows that genetic polymorphisms in CYP3A5 may be responsible, in part, for the large interindividual variability of cyclosporine pharmacokinetics during the early phase after renal transplantation in Chinese patients. Patients with the CYP3A5*3 variant genotype require a low dose of cyclosporine to reach target levels compared with those with the CYP3A5*1 allele.