Population pharmacokinetics and pharmacodynamics of cefpirome in critically ill patients against Gram-negative bacteria

Objectives To develop a population pharmacokinetics model for cefpirome in ICU patients, to assess pharmacokinetic-pharmacodynamic profiles vs. MIC distribution of likely ICU pathogens, and to assess their expected cumulative fraction of response (CFR). Design and setting Prospective observational study in a multidisciplinary ICU. Measurements and results Twelve patients received 2 g cefpirome intravenously over 12 h. Thirteen blood samples were taken on two occasions. Demographic and creatinine clearance data were collected. Based on the final covariate model obtained using NONMEM, Monte Carlo simulations were undertaken to simulate free-drug concentrations for two administration methods: intermittent bolus administration (IBA) and continuous infusion (CI) with a loading dose of 0.5 g. Concentration-time profiles were evaluated by the probability of achieving free-drug concentrations above the MIC for more than 65% of dosing interval. Using MIC distributions from the EUCAST programme the CFR for each method was evaluated. A three-compartment model with zero-order input best described the concentration-time data. The CFR for Escherichia coli and Klebsiella spp. was greater than 97% in all IBA and CI doses but for Pseudomonas aeruginosa, and Acinetobacter spp. achieved target concentrations of 56% and 46%, respectively. High-dose CI cefpirome (6 g/day) for P. aeruginosa and Acinetobacter spp. was required to achieve CFR of 89%. Conclusion Measured creatinine clearance appears to be a good marker of cefpirome clearance and potentially could be used to individualise cefpirome therapy. When given as IBA or CI for E. coli and Klebsiella spp., cefpirome should be successful. Cefpirome fails to achieve the bactericidal target even when administered at high-doses such as 6 g/day for P. aeruginosa and Acinetobacter spp. Prospective clinical studies are needed to conclusively validate these findings.