Population pharmacokinetic/pharmacodynamic modeling of linezolid in critically ill adult patients

Abstract

Objective. To evaluate pharmacokinetics using therapeutic drug monitoring (TDM) data and develop a population pharmacokinetic/pharmacodynamic (PK/PD) model of linezolid for individualizing dosage regimens in critically ill adult patients in intensive care unit (ICU). Materials and Methods. A retrospective study of linezolid pharmacokinetics was conducted using TDM data (peak-trough strategy) and included data (n = 23) from patients treated in the ICU and receiving the standard linezolid dosing regimen of 600 mg intravenously twice daily. Plasma linezolid concentrations were measured by high-performance liquid chromatography with ultraviolet detection (HPLC-UV). Individual and population PK parameters of linezolid were estimated using the NPAG program from USC*PACK software (Pmetrics). Results. The trough concentration values obtained with the standard dosing regimen were within the recommended range (2–7 mg/L) in only 7 (28.0%) cases, with trough concentrations below 2 mg/L in 14 (56.0%) cases and above 7 mg/L in 4 (16.0%) cases. The coefficient of variation was estimated as 55.1% for Cmax and 126.4% for Ctrough. Based on the TDM results (a total of 50 linezolid measured concentration), the parameters of the linear one-compartment PK model were identified. The median elimination rate constant (Kel) was estimated as 0.136 1/h (mean: 0.167 1/h), the median apparent volume of distribution (V) was 42.3 L (mean: 49.5 L), the median half-life (T1/2) was calculated as 4.2 h (mean: 5.2 h), and the median total clearance (CL) was 6.64 L/h (mean 8.0 L/h). The obtained estimations of population PK parameters of linezolid are consistent with those previously published in the literature. For estimated PK parameters, interindividual variability was within 43–63% range. PK/PD modeling results showed that for the assumed MIC value of 1 mg/L, the estimates of the main PK/PD parameters (AUC0–24/MIC) were within the recommended target ranges in the majority of patients (except for 4 patients). For MIC = 2 mg/L, only 6 (26.1%) patients and for MIC = 4 mg/L only 2 (8.7%) patients had their AUC0–24/MIC values above 100. Conclusions. When treating critically ill patients, due to significant PK changes associated with disease severity and intensive care unit procedures, as well as significant interindividual variability, personalized linezolid dosing should preferably be guided by TDM. The use of specialized Bayesian population modeling software will enable individualization based on the primary PK/PD target (AUC0–24/MIC), which requires relatively complex calculations, even for rare TDM measurements. It will also allow for concentration monitoring to begin as soon as possible after therapy initiation, without waiting for steady-state distribution to be established. This approach helps improve the efficacy and safety of linezolid therapy