Vancomycin dosing in neonates: enhancing outcomes using population pharmacokinetics and simulation

Abstract

Introduction: Optimizing vancomycin dosing in neonates is a critical yet complex goal. Traditional trough concentration-based dosing strategies correlate poorly with therapeutic efficacy and often fail to account for the significant renal function variability and drug clearance in neonates. The 24-hour area under the concentration-time curve to minimum inhibitory concentration (AUC₂₄/MIC) ≥ 400 mg h/L has emerged as a superior pharmacodynamic target. Population pharmacokinetics (PopPK) models allow optimized dosing by incorporating neonatal-specific factors such as postmenstrual age (PMA), gestational age (GA), serum creatinine (SCr), and weight.

Objective: To develop optimized vancomycin dosing regimens for neonates that achieve an 80% probability of target attainment (PTA) for an AUC₂₄/MIC ≥ 400 mg h/L across diverse clinical cohorts and simulated neonatal populations.

Methods: Real-world data from three international centers (Belgium, New Zealand, USA), including 610 individuals and 2399 vancomycin concentrations, were used to externally evaluate a previously published PopPK model (NONMEM^®). Missing data, including body weight, were imputed using Amelia II version 1.7.3 for R, while Zelig for R integrated multiple imputed datasets. A virtual population of 10,000 neonates was independently generated using MATLAB to simulate clinical scenarios considering covariates such as PMA, GA, SCr, body weight, and imputed body length.

Results: Simulations showed that PMA and SCr were key covariates that significantly improved PTA, particularly in preterm neonates. Preterm neonates achieved PTAs of 80% with daily doses of 30 or 40 mg/kg/day, while term neonates required 15 mg/kg every 8 hours or 20 mg/kg every 12 hours. The simulations demonstrated that these optimized dosing strategies achieved an 80% PTA for AUC₂₄/MIC ≥ 400 mg h/L in the virtual neonatal population. For neonates with PMA < 29 weeks and SCr > 0.6 mg/dL, including SCr as a covariate increased the likelihood of achieving the target from 65% to 87%.

Conclusion: Incorporating developmental factors like PMA and SCr into vancomycin dosing strategies achieved robust and clinically relevant outcomes. The optimized regimens achieved an 80% PTA for the AUC₂₄/MIC target for preterm and term neonates. These findings offer a scalable framework for improving neonatal vancomycin pharmacotherapy across diverse populations and clinical settings.

Keywords: dosing regimens; neonatology; population pharmacokinetics; simulations; the probability of target attainment; vancomycin.

Figure 1

Histograms of covariates of hypothetical preterm and term neonates. Histograms demonstrated the distributions of weight, SCr, HT, PMA, GA, and PNA in the hypothetical preterm (n=5,000) and term (n=5,000) neonates. The correlation coefficients with p-values were also included, indicating well-conserved correlations among covariates. The complete statistical summary was described in Supplementary Table S1 .

Figure 2

Visual predictive check (VPC) plots of external evaluation datasets 1–3. External evaluation of the population pharmacokinetic model (n=1000) using external datasets (a: Salt Lake City dataset, b: Leuven dataset, c: Dunedin dataset). Observed concentrations were plotted using an opened circle. The dashed lines represent the 5th and 95th percentiles, and the solid line represents the 50th percentile of the observations. Shaded areas are the corresponding model-based 95% confidence intervals for the median and the prediction intervals.

Figure 3

Box plots of steady-state AUC₂₄ (mg.hr/L) of the published dosing regimens. Box plots depicting average steady-state AUC₂₄ simulated from the published dosing regimens. The red dashed line at 400 mg.hr/L represents the target AUC₂₄/MIC ≥400, given MIC value of 1 µg/mL. From left to right: Age-based dosing regimens were PMA and PNA-based from Neofax 2011 (Hey, 2011); Age-based Neonatal Formulary 6 and 7 were PMA-based dosing regimens (Pomerantseva and Ilicheva, 2012; Ainsworth, 2014); Renal function-based dosing regimens were adapted from Red Book 2012 (American Academy of Pediatrics, 2012), Nelson’s Pediatric Antimicrobial Therapy (Germovsek et al., 2019), and Capparelli et al (Johnson et al., 1997); Weight-based dosing regimens were adapted from Red Book 2009 (American Academy of Pediatrics, 2009).

Figure 4

Simulations of steady-state AUC₂₄ using vancomycin daily doses of 20–45 mg/kg. Simulations of the average AUC₂₄ (mg.hr/L) at the steady state of preterm (a) and term (b) neonates. Dosing regimens for vancomycin intermittent IV infusion were 10, 15, 20, 30, and 40 mg/kg, and the dosing intervals were 8, 12, and 24 hours. Red lines at 400 mg.hr/L represent the target AUC₂₄ ≥400 mg.hr/L with the assumption of the MIC of 1 µg/mL.