Combination of Pharmacokinetic and Pathogen Susceptibility Information To Optimize Meropenem Treatment of Gram-Negative Infections in Critically Ill Patients

Abstract

Meropenem is one of the most frequently used antibiotics to treat life-threatening infections in critically ill patients. This study aimed to develop a meropenem dosing algorithm for the treatment of Gram-negative infections based on intensive care unit (ICU)-specific resistance data. Antimicrobial susceptibility testing of Gram-negative bacteria obtained from critically ill patients was carried out from 2016 to 2020 at a tertiary care hospital. Based on the observed MIC distribution, stochastic simulations (n = 1,000) of an evaluated pharmacokinetic meropenem model, and a defined pharmacokinetic/pharmacodynamic target (100%T_>4×MIC while minimum concentrations were <44.5 mg/L), dosing recommendations for patients with varying renal function were derived. Pathogen-specific MIC distributions were used to calculate the cumulative fraction of response (CFR), and the overall MIC distribution was used to calculate the local pathogen-independent mean fraction of response (LPIFR) for the investigated dosing regimens. A CFR/LPIFR of >90% was considered adequate. The observed MIC distribution significantly differed from the EUCAST database. Based on the 6,520 MIC values included, a three-level dosing algorithm was developed. If the pathogen causing the infection is unknown (level 1), known (level 2), known to be neither Pseudomonas aeruginosa nor Acinetobacter baumannii, or classified as susceptible (level 3), a continuous infusion of 1.5 g daily reached sufficient target attainment independent of renal function. In all other cases, dosing needs to be adjusted based on renal function. ICU-specific susceptibility data should be assessed regularly and integrated into dosing decisions. The presented workflow may serve as a blueprint for other antimicrobial settings.

Keywords: Gram negative; antimicrobial susceptibility testing; critically ill; dosing algorithm; meropenem.

FIG 1

Observed distribution of MIC measurements and EUCAST predicted distribution for Acinetobacter baumannii (left) and Pseudomonas aeruginosa (right). Bars indicate MIC values determined in critically ill patients of the University Hospital, LMU Munich, between 2016 and 2020; Points and whiskers indicate median prediction and 95% confidence interval for the EUCAST distribution based on 3,000 bootstraps. N, total number of isolates.

FIG 2

Cumulative fraction of response (CFR) or local pathogen-independent mean fraction of response (LPIFR) for continuous, prolonged, and short infusion regimens at daily doses of 1.5 g, 3 g, 6 g, and 9 g in function of creatinine clearance. All pathogens (red), Acinetobacter baumannii (green), and Pseudomonas aeruginosa (blue). CFR and LPIFR results were based on the defined target of the pharmacokinetic/pharmacodynamic index (meropenem concentration 100% of the time above 4× the MIC but not above the toxic threshold of 44.5 mg/L) and the distribution of MIC measurements of critically ill patients of the University Hospital, LMU Munich, from 2016 to 2020 and 1,000 stochastic simulations.

FIG 3

Dosing algorithm based on intensive care unit-specific pathogen information of a tertiary care hospital for meropenem in critically ill patients with Gram-negative infections. *, Resistance rate of  >10%, consider combination therapy. ⁺, Target of 100% time above 1× MIC. CI, continuous infusion; PSAE, Pseudomonas aeruginosa; ACB, Acinetobacter baumannii; S, susceptible; I, susceptible with increased exposure; R, resistant.