Using genetic and clinical factors to predict tacrolimus dose in renal transplant recipients

Aims: Tacrolimus has a narrow therapeutic window and shows significant interindividual difference in dose requirement. In this study we aim to first identify genetic factors that impact tacrolimus dose using a candidate gene association approach, and then generate a personalized algorithm combining identified genetic and clinical factors to predict individualized tacrolimus dose. Materials & methods: We screened 768 SNPs in 15 candidate genes in metabolism, transport and calcineurin inhibition pathways of tacrolimus, for association with tacrolimus dose in a discovery cohort of 96 patients. Results: Four polymorphisms in CYP3A5 and one polymorphism in CYP3A4 were identified to be significantly associated with tacrolimus stable dose (p < 8.46 × 10-5). The same SNPs were identified when dose-normalized trough tacrolimus concentration was analyzed. The CYP3A5*1 allele was associated with significantly higher stable dose, bigger dose increase, higher risk of being underdosed and lower incidence of post-transplant hyperlipidemia. ABCB1 polymorphisms were not associated with stable dose. No significant difference was found between CYP3A5 expressers and nonexpressers in incidence of acute rejection and time to first rejection. Age, ethnicity and CYP3A inhibitor use could predict 30% of tacrolimus dosing variability. Adding the identified genetic polymorphisms to the algorithm increased the predictability to 58%. In two validation cohorts of 77 and 64 patients, the algorithm containing both genetic and clinical factors produced correlation coefficients of 0.63 and 0.42, respectively. This algorithm gave a prediction of the stable doses closer to the actual doses when compared with another algorithm based only on the CYP3A5 genotype. Conclusion: CYP3A5 genotype is the most significant genetic factor that impacts tacrolimus dose among the genes studied. This study generated the first pharmacogenomics model that predicts tacrolimus stable dose based on age, ethnicity, genotype and comedication use. Our results highlight the importance of incorporating both genetic and clinical, demographic factors into dose prediction.