Predicting tacrolimus concentrations in children receiving a heart transplant using a population pharmacokinetic model.

Objective: Immunosuppressant therapy plays a pivotal role in transplant success and longevity. Tacrolimus, a primary immunosuppressive agent, is well known to exhibit significant pharmacological interpatient and intrapatient variability. This variability necessitates the collection of serial trough concentrations to ensure that the drug remains within therapeutic range. The objective of this study was to build a population pharmacokinetic (PK) model and use it to determine the minimum number of trough samples needed to guide the prediction of an individual's future concentrations.

Design setting and patients: Retrospective data from 48 children who received tacrolimus as inpatients at Primary Children's Hospital in Salt Lake City, Utah were included in the study. Data were collected within the first 6 weeks after heart transplant.

Outcome measures: Data analysis used population PK modelling techniques in NONMEM. Predictive ability of the model was determined using median prediction error (MPE, a measure of bias) and median absolute prediction error (MAPE, a measure of accuracy). Of the 48 children in the study, 30 were used in the model building dataset, and 18 in the model validation dataset.

Results: Concentrations ranged between 1.5 and 37.7 μg/L across all collected data, with only 40% of those concentrations falling within the targeted concentration range (12 to 16 μg/L). The final population PK model contained the impact of age (on volume), creatinine clearance (on elimination rate) and fluconazole use (on elimination rate) as covariates. Our analysis demonstrated that as few as three concentrations could be used to predict future concentrations, with negligible bias (MPE (95% CI)=0.10% (-2.9% to 3.7%)) and good accuracy (MAPE (95% CI)=24.1% (19.7% to 27.7%)).

Conclusions: The use of PK in dose guidance has the potential to provide significant benefits to clinical care, including dose optimisation during the early stages of therapy, and the potential to limit the need for frequent drug monitoring.