Aims: Some population pharmacokinetic models have been developed using height to explain some of the interindividual variability in tobramycin pharmacokinetics in cystic fibrosis patients. However, their predictive performance when extrapolated to other clinical centres is unclear. Therefore, the aim of this study was to externally evaluate the predictability of tobramycin population pharmacokinetic models with an independent dataset and perform simulations using previously recommended height-based dosing regimens.
Methods: A literature search was conducted through the PubMed database to identify relevant population pharmacokinetic models. Tobramycin plasma concentration data from April 2014 to November 2019 were retrospectively collected from the Institut universitaire de cardiologie et de pneumologie de Québec, Canada. External evaluations were performed using NONMEM® v7.5 and RStudio® v1.3.1073. Monte Carlo simulations were performed to evaluate the probability of target attainment of Cmax /MIC ratios for several dosing regimens.
Results: The validation dataset included 27 patients and 143 concentration samples. Three models were evaluated. Only the ones by Crass et al. and Alghanem et al. performed satisfactorily in terms of prediction-based diagnostics with MDPE values of -3.4% and 29.3% and MDAPE values of 19.0 and 29.5%, respectively. In simulation-based evaluations, both pcVPC and NPDE showed no evidence of model misspecification. Our simulations suggest that patients treated with a once-daily dose of 3.4 mg/cm should produce peak and trough levels consistent with current guidelines.
Conclusion: Our results show that the models by Crass et al. and Alghanem et al. are appropriate for simulation-based applications to aid individualized dosing in our population and that height-based dosing regimens could be considered in cystic fibrosis patients.
Keywords: NONMEM; cystic fibrosis; population pharmacokinetics; therapeutic drug monitoring; tobramycin.
© 2021 British Pharmacological Society.