Dose Optimization of Amikacin in the Emergency Department: A Population Pharmacokinetics Simulation Study.

Abstract

Background: In adult patients with sepsis or septic shock admitted to the emergency department, a single intravenous 15 mg/kg amikacin dose provides inadequate pharmacokinetic-pharmacodynamic target attainment at the locally reported minimum inhibitory concentration (MIC) of 2 mg/L and the European Committee on Antimicrobial Susceptibility Testing clinical breakpoint for Enterobacterales of 8 mg/L.

Objectives: To provide an amikacin dosing strategy with a clinically acceptable probability of target attainment (PTA) for all patients.

Methods: Stochastic simulations were performed using a two-compartment population pharmacokinetics model of amikacin (NONMEM 7.5). PTA was evaluated for various dosing strategies across a range of virtual patients' body weight, body mass index, serum total protein, serum sodium, fluid balance, and estimated glomerular filtration rate according to the Chronic Kidney Disease Epidemiology Collaboration equation (eGFRCKD-EPI), at the locally reported MIC of 2 mg/L and the clinical breakpoint of 8 mg/L. The pharmacokinetic-pharmacodynamic targets were a 24-hour area under the concentration-time curve (AUC24h)/MIC of ≥80 and a 24-hour postdose concentration (C24h) of < 3 mg/L for efficacy and safety, respectively.

Results: The PTA for the clinical breakpoint of 8 mg/L was <90% with standard 15 mg/kg dosing, across all patient characteristics. A flat 1500-mg dose achieved ≥90% PTA for the entire population at a MIC of 2 mg/L. However, at the clinical breakpoint of 8 mg/L, a flat 3500-mg dose provided ≥90% PTA only when the eGFRCKD-EPI was <96 mL/min/1.73 m2. The C24h was similar for 1500 mg and 15 mg/kg dosing, whereas 3500 mg resulted in a higher C24h.

Conclusions: A flat dose is recommended over weight-based dosing. However, selecting a 1500-mg or 3500-mg dose may compromise either efficacy (MIC 2 mg/L) or safety (clinical breakpoint 8 mg/L), posing a dilemma. Clinical validation is warranted.