Integrating pharmacogenetic and clinical factors to predict the C0/D/W-based tacrolimus phenotype in kidney...

Tacrolimus shows substantial interindividual pharmacokinetic variability, complicating dose individualization in renal transplantation. The tacrolimus trough concentration-to-dose-to-weight ratio (C0/D/W) has been proposed as a simple surrogate of tacrolimus bioavailability, yet the extent to which pharmacogenetic and clinical factors explain this phenotype in stable adult kidney transplant recipients remains unclear.

We conducted a single-centre retrospective study including 77 adult renal transplant recipients treated with prolonged-release tacrolimus formulations (Advagraf® or Envarsus®) and clinically stable. Tacrolimus exposure phenotype was defined using C0/D/W tertiles: the lowest tertile was classified as the high-metabolism phenotype and the highest tertile as the low-metabolism phenotype. Genotyping was performed using a customized OpenArray panel. The predictive performance of CYP3A5 rs776746 alone was assessed using diagnostic accuracy metrics. Multivariable logistic regression models integrating clinical covariates and genotype-based exposure scores were developed. Model performance was evaluated using McFadden's pseudo-R² and Receiver Operating Characteristic analysis with bootstrap internal validation.

The high- and low-metabolism groups comprised 39 and 38 patients, respectively. CYP3A5 expresser genotypes (*1 carriers) identified high-metabolism phenotype with a sensitivity of 48.7% and a specificity of 89.5%, yielding an overall accuracy of 68.8%. In multivariable analysis, CYP3A5 rs776746 score, age, and post-transplant period were associated with metabolism phenotype. The primary parsimonious model demonstrated good discrimination (apparent AUC 0.83; 95% CI 0.73-0.92), with an optimism-corrected AUC of 0.80 and a McFadden pseudo-R² of 0.26. Exploratory analyses including additional pharmacogenetic variants such as ABCB1 rs1045642 provided only marginal improvement in model fit and did not justify increased model complexity.

CYP3A5 genotype alone provides high specificity but limited sensitivity for predicting tacrolimus C0/D/W-based metabolism phenotype. Incorporation of key clinical variables, including age and time since transplantation, improves phenotype discrimination and provides a more comprehensive characterization of tacrolimus exposure variability. Additional pharmacogenetic markers showed limited incremental value. These findings support a combined genetic-clinical approach for tacrolimus dose individualization, although external validation in larger cohorts is warranted.