Retrospective Bayesian Reanalysis of Single Gentamicin Concentrations: A Neonatal Case Series.

Area of Science:

• Neonatal pharmacology
• Pharmacokinetics and drug monitoring
• Antibiotic dosing optimization

Background:

• Existing nomograms for gentamicin dosing in neonates are limited or outdated.
• Bayesian software shows promise for therapeutic drug monitoring but its utility for neonatal gentamicin is not established.
• Accurate gentamicin level prediction is crucial for effective and safe neonatal care.

Purpose of the Study:

• To evaluate the accuracy of Bayesian estimation for predicting peak (Cmax) and trough (Cmin) gentamicin concentrations from a single measured level in neonates.
• To compare the performance of Bayesian analysis using trough-only versus peak-only concentration data.
• To provide guidance on optimal timing for single-level gentamicin monitoring in neonates.

Main Methods:

• Retrospective analysis of gentamicin concentrations in 57 neonates.
• Utilized InsightRx® software for Bayesian estimation of Cmax and Cmin from single peak or trough levels.
• Assessed accuracy using mean difference and 95% limits of agreement between estimated and measured concentrations.

Main Results:

• Bayesian analysis of a single trough concentration yielded reliable Cmin estimates (MD: 0.03 mg/L, LOA: -0.17 to 0.13) but less accurate Cmax estimates.
• Analysis of a single peak concentration resulted in overestimation of both Cmax (MD: 1.2 mg/L, LOA: -0.58 to 3.0) and Cmin.
• Neonatal characteristics (median GA 34 weeks, PN 0 days) were noted.

Conclusions:

• Bayesian analysis of neonatal gentamicin trough levels is accurate for predicting minimum concentrations (Cmin) when drawn near the end of the dosing interval.
• Estimating maximum concentrations (Cmax) using this method is less reliable.
• Single-level gentamicin monitoring should prioritize trough levels drawn late in the dosing interval to ensure safety and efficacy.