Vancomycin population pharmacokinetics and optimized dosing in critically ill infants and children.

BACKGROUND

Vancomycin dosing remains challenging due to narrow therapeutic window and pharmacokinetic (PK) variability in critically ill children. This study aims to characterise pediactric vancomycin PK, identify significant covariates, and optimise dosing strategies.

METHODS

Vancomycin plasma concentrations from patients aged 1 month to 12 years (n=50) were retrospectively retrieved from therapeutic drug monitoring (TDM) records and combined with data from a prospective PK study (n=9). A population PK model was developed using non-linear, mixed-effects modelling software (Monolix®). External validation was performed using an independent retrospective cohort (n=59). Monte Carlo simulations were used to assess optimal dosing regimens targeting an AUC24-48 of 400-600 mg·h/L.

RESULTS

A two-compartment model with first-order elimination best described the PK data. Typical population estimates for clearance and central volume of distribution (standardized to 70 kg) were 5.04 L/h and 31.5 L, respectively. In addition to age and weight, the estimated glomerular filtration rate was a significant predictor of vancomycin clearance. External validation demonstrated acceptable population-level bias with mean and median prediction error of 17.7% and 12.7%, but limited precision with a root mean square error (RMSE) of 52.7%. All predefined validation criteria were met when individual predictions were used, supporting adequate model performance after incorporating individual-specific information.

CONCLUSIONS

This study presents a population PK model for vancomycin in critically ill infants and children using combined prospective and retrospective data. The model can support personalized vancomycin dosing by improving target attainment and reducing toxicity across a range of patient ages and renal function.