Assessing the Utility of a Genotype‐Guided Tacrolimus Equation in African American Kidney Transplant Recipients: A Single Institution Retrospective Study

Joy E Obayemi,Lauren Callans,Nikhil Nair,Hui Gao,Divya Gandla,Bao‐Li Loza,Sarah Gao,Maedeh Mohebnasab,Jennifer Trofe‐Clark,Pamala Jacobson,Brendan Keating
DOI: https://doi.org/10.1002/jcph.2461
2024-05-22
The Journal of Clinical Pharmacology
Abstract:Tacrolimus metabolism is heavily influenced by the CYP3A5 genotype, which varies widely among African Americans (AA). We aimed to assess the performance of a published genotype‐informed tacrolimus dosing model in an independent set of adult AA kidney transplant (KTx) recipients. CYP3A5 genotypes were obtained for all AA KTx recipients (n = 232) from 2010 to 2019 who met inclusion criteria at a single transplant center in Philadelphia, Pennsylvania, USA. Medical record data were used to calculate predicted tacrolimus clearance using the published AA KTx dosing equation and two modified iterations. Observed and model‐predicted trough levels were compared at 3 days, 3 months, and 6 months post‐transplant. The mean prediction error at day 3 post‐transplant was 3.05 ng/mL, indicating that the model tended to overpredict the tacrolimus trough. This bias improved over time to 1.36 and 0.78 ng/mL at 3 and 6 months post‐transplant, respectively. Mean absolute prediction error—a marker of model precision—improved with time to 2.33 ng/mL at 6 months. Limiting genotype data in the model decreased bias and improved precision. The bias and precision of the published model improved over time and were comparable to studies in previous cohorts. The overprediction observed by the published model may represent overfitting to the initial cohort, possibly limiting generalizability.
pharmacology & pharmacy
What problem does this paper attempt to address?