External validation of population pharmacokinetic models for meropenem in critically ill adult patients

While several meropenem population pharmacokinetic (popPK) models have been developed, few have undergone external validation, a crucial step to confirm their robustness and applicability in real-world settings.

This study aimed to conduct an external evaluation of published meropenem popPK models to assess their predictive performance and determine their suitability for implementing Model-Informed Precision Dosing strategies in the intensive care unit (ICU) setting.

The external validation dataset consisted of data retrospectively collected from a tertiary university hospital. Eight published popPK models were selected from the literature and bias and inaccuracy were calculated. Predictive performance was assessed in two subpopulations: continuous renal replacement therapy (CRRT) and non-CRRT patients, with further stratification by body mass index.

Eight popPK models were evaluated with an independent dataset of 30 ICU patients and 48 samples. The Ulldemolins et al. model exhibited the lowest bias for population-level predictions in the overall CRRT cohort. In the overall non-CRRT cohort, the models by Chung et al. and Lan et al. demonstrated excellent population and individual prediction performance. Within the obese subpopulation, the Shekar et al. model showed the lowest bias and inaccuracy in the CRRT cohort, while the models by Li et al., Lan et al., and Chung et al. performed best in the non-CRRT cohort.

Published meropenem popPK models exhibit considerable variability in predictive performance when validated in an external dataset of ICU patients failing to generalize across broader patient populations. These findings underscore the need for external validation with independent datasets to ensure reliable performance across diverse populations.