Predictive Performance of Pharmacokinetic Model-Based Virtual Trials of Vancomycin in Neonates: Mathematics Matches Clinical Observation

Bu-Fan Yao,Yue-E Wu,Bo-Hao Tang,Guo-Xiang Hao,Evelyne Jacqz-Aigrain,John van den Anker,Wei Zhao
DOI: https://doi.org/10.1007/s40262-022-01128-z
2022-01-01
Clinical Pharmacokinetics
Abstract:Background and Objective Vancomycin is frequently used to treat Gram-positive bacterial infections in neonates. However, there is still no consensus on the optimal initial dosing regimen. This study aimed to assess the performance of pharmacokinetic model-based virtual trials to predict the dose–exposure relationship of vancomycin in neonates. Methods The PubMed database was searched for clinical trials of vancomycin in neonates that reported the percentage of target attainment. Monte Carlo simulations were performed using nonlinear mixed-effect modeling to predict the dose–exposure relationship, and the differences in outcomes between virtual trials and real-world data in clinical studies were calculated. Results A total of 11 studies with 14 dosing groups were identified from the literature to evaluate dose–exposure relationships. For the ten dosing groups where the surrogate marker for exposure was the trough concentration, the mean ± standard deviation (SD) for the target attainment between original studies and virtual trials was 3.0 ± 7.3%. Deviations between − 10 and 10% accounted for 80% of the included dosing groups. For the other four dosing groups where the surrogate marker for exposure was concentration during continuous infusion, all deviations were between − 10 and 10%, and the mean ± SD value was 2.9 ± 4.5%. Conclusion The pharmacokinetic model-based virtual trials of vancomycin exhibited good predictive performance for dose–exposure relationships in neonates. These results might be used to assist the optimization of dosing regimens in neonatal practice, avoiding the need for trial and error.
What problem does this paper attempt to address?