Pharmacokinetic modelling and Bayesian estimation-assisted decision tools to optimize vancomycin dosage in neonates: only one piece of the puzzle. (2nd September 2019)

• Record Type: Journal Article
• Authors: Allegaert, Karel; Flint, Robert; Smits, Anne
• Is Part Of: Expert opinion on drug metabolism and toxicology. Volume 15:Number 9(2019)
• Page Start: 735
• Page End: 749
• Publication Date: 2019-09-02
• Subjects: Vancomycin -- therapeutic drug monitoring -- newborn -- pharmacokinetics -- Bayesian model-supported targeting -- coagulase-negative staphylococcus; Drugs -- Toxicology -- Periodicals; Drugs -- Metabolism -- Periodicals; 615.7
• DOI: 10.1080/17425255.2019.1655540
• Languages: English
• ISSNs: 1742-5255

Abstract

ABSTRACT: Introduction : Vancomycin is commonly administered to neonates, while observational data on therapeutic drug monitoring (TDM, trough levels) suggest that vancomycin exposure and dosage remain substandard. Area covered : Data on vancomycin pharmacokinetics (PK) and its covariates are abundant. Consequently, modeling is an obvious tool to improve targeted exposure, with a shift from TDM trough levels to area under the curve (AUC24h ) targets, as in adults. Continuous administration appeared as a practice to facilitate AUC24h target attainment, while Bayesian model-supported targeting emerged as a novel tool. However, the AUC24h /MIC (minimal inhibitory concentration) target itself should consider neonate-specific aspects (bloodstream infections, coagulase-negative staphylococci, protein binding, underexplored causes of variability, like assays, preparation and administration inaccuracies, or missing covariates). Expert opinion : To improve targeted exposure in neonates, initial vancomycin prescription should be based on 'a priori model-based individual dosing' using validated dosing regimens, followed by further tailoring by dosing optimization applying Bayesian estimation-assisted TDM. Future research should focus on the feasibility to integrate these tools (individualized dosing, Bayesian models) in clinical practice, and to perform PK/PD studies in the relevant animal models and human neonatal setting (coagulase-negative staphylococci, bloodstream infections).