Antimicrobial therapy in critically ill patients: Improving clinical outcomes using a translational pharmacological approach

Access to files

•   FULL-TEXT.PDF (pdf)

Abstract

Pulmonary infections in critically ill patients are common, frequently lethal and treatment may be complicated by bacterial resistance. Piperacillin-tazobactam (PTZ) is a broad-spectrum β-lactam antibiotic, frequently used for pulmonary infections. Lung antibiotic concentration reflects target site concentrations in patients with pneumonia. Critically ill patient’s exhibit marked pharmacokinetic (PK) variability. PTZ exposures resulting in maximal bacterial killing and prevention of emergence of drug resistance are not known. Administration of PTZ by extended infusions (EI) or using Bayesian dosage optimisation, instead of a fixed bolus regimen, may improve clinical outcomes. Experimental work was conducted in an in vitro hollow fibre infection model (HFIM) using two densities of Pseudomonas aeruginosa. Experimental data was described by a mathematical model allowing identification of PTZ exposures associated with bacterial killing and suppression of the emergence of resistance. The population PK of PTZ in the plasma and lung of 17 critically ill patients was estimated. Monte Carlo simulation was used to explore the proportion of patients that achieve the plasma and lung PTZ exposures associated with bacterial killing and resistance suppression and to determine the effect of administration schedule. Finally, the population PK of PTZ in the plasma of 146 critically ill patients was estimated and used to construct computer software that can individualise PTZ dosing. Precision of the dosing software was assessed in 8 additional individuals. At low bacterial density a trough piperacillin:MIC ratio of 3.4 for bolus and 10.4 for EI regimens were able to suppress the emergence of resistance. At higher bacterial density all regimens were associated with growth of a resistant sub-population. Pulmonary piperacillin and tazobactam concentrations were unpredictable and negatively correlated to pulmonary permeability. Simulations revealed that EI, compared with bolus dosing, of PTZ is associated with a higher likelihood of bacteria killing. Similar probability of developing resistance was predicted with PTZ administration by EI and by bolus administration. Performance of the dose optimisation software was satisfactory.Current PTZ regimens are insufficient to treat pneumonia in approximately 14% of critically ill patients. Delivery of PTZ by EI may be a more effective method of administration for some patients with nosocomial infections. Individualised PTZ regimens, delivering a target piperacillin concentration, identified in a HFIM, are achievable and should improved clinical outcomes. Patients with a high bacterial burden may required alternative therapeutic strategies to maximize bacterial killing and prevent antimicrobial resistance.

Layman's Abstract

Pneumonia is a frequently fatal disease affecting patients in intensive care units. These patients are commonly treated with piperacillin-tazobactam, a broad spectrum antibiotic that is extensively used in the NHS. Currently all critically ill patients (except those in renal failure) are administered the same dose of piperacillin-tazobactam. However this results in considerable variability in antibiotic concentrations in the blood and lungs. Patients with low concentrations may be at increased risk of not responding to piperacillin-tazobactam or developing resistance to antibiotics. High levels of drugs increase the risk of side-effects. The first part of this study used a laboratory infection model to identify the amount of the drug required to kill bacteria and prevent development of antibiotic resistance. The second part of this study described the concentrations of antibiotic in the blood and lungs of patients in the intensive care unit. Mathematical modelling was used to show that changing from current bolus regimens (were the drug is administered over 30 minutes every 8 hours) to delivering piperacillin-tazobactam by infusion (were the drug is administered over 4 to 8 hours every 8 hours) may cure more patients but does not stop emergence of resistance. Finally computer software was developed which could be used to identify the optimal dose for individual patients associated with clinical cure and suppression of emergence of resistance. Computer simulation demonstrated the reliability of the dosing software.

Keyword(s)

Pharmacodynamics; Pharmacokinetics; Piperacillin-tazobactam

Bibliographic metadata

Institutional metadata

Record metadata