Population Pharmacokinetics And Rational Dosing Of Aminoglycosides In Critically Ill Patients

Objective:Early and adequate antibiotic treatment is critical for the clinical outcome of critically ill patients.Tobramycin and gentamicin are representative drugs of aminoglycosides used for the anti-infective treatment in critically ill patients.However,the population pharmacokinetic（PPK）characteristics and rational dosing of aminoglycosides in this patient population are unclear.This study intends to build PPK models for tobramycin and gentamicin in critically ill patients based on real-world clinical data and to explore rational dosing of tobramycin and gentamicin in critically ill patients and CRRT patients by Monte Carlo simulation approach.Methods:（1）Data mining and processing was carried out by using R^?software to perform from a Medical Information Mart for Intensive Care Ⅲ（MIMIC Ⅲ）database that is publicly available internationally.The dosing records（e.g.,dose,administration time,and infusion time）,concentration measurements,demographics（e.g.,age and weight）,and clinical variables（e.g.,serum creatinine）of tobramycin and gentamicin in critically ill patients were extracted from the database.Since there is no CRRT patients in MIMIC Ⅲ,a literature search for CRRT patients dosed with tobramycin or gentamicin was made through Pub Med,and the relevant dosing records,concentration measurements,and covariates information were extracted and collated.（2）The collected concentration-time data of tobramycin and gentamicin in critically ill patients and CRRT patients were used to establish PPK models by non-linear mixed-effects modeling approach using the NONMEM^?software（version 7.3）.（3）Based on the constructed PPK models and predefined PK/PD target and toxicity risk threshold,Monte Carlo simulation were performed to investigate the applicability of the Hartford nomogram in critically ill patients and the rational dosing regimen of aminoglycosides in critically ill CRRT patients.Results:（1）Based on the MIMIC Ⅲ database,140 critically ill patients with complete dosing records and concentration measurements of tobramycin were obtained,and 72 critically ill patients with complete dosing records and concentration measurements of gentamicin were obtained.Based on the data mining of retrieved Pub Med literature,14CRRT patients with complete dosing records,concentration measurements and CRRT parameters of gentamicin were obtained.（2）A two-compartment model was optimal for describing the PK characteristics of tobramycin in critically ill patients,and the correlation between patients’creatinine clearance and drug clearance followed a power model.One-compartment model was optimal for describing the PK characteristics of gentamicin in both critically ill patients and CRRT patients,and the correlation between creatinine clearance and drug clearance in critically ill patients was also followed by the power model.The drug clearance in the CRRT critically ill patient consisted of the endogenous clearance and the CRRT extracorporeal clearance.（3）Monte Carlo simulations demonstrated that the Hartford nomogram of aminoglycosides can be used as an initially empirical dosing regimen for“normal”critically ill patients with a pathogen MIC of≤1mg/L,but it was not suitable for an MIC of≥2 mg/L.A combination of 7 mg/kg gentamicin q24h and 40 m L/kg/h CRRT dose can provide desirable efficacy and low risk of toxicity for CRRT patients.Conclusions:This study successfully established PPK models of tobramycin and gentamicin in critically ill patients and CRRT patients and clarified the rational dosing regimens for critically ill patients.Our study can provide useful information for the rational use of aminoglycosides for clinicians.