Résumé
Prolonged infusion of β-lactam antibiotics as either extended (over at least 2 hours) or continuous infusion is increasingly applied in intensive care units around the world in an attempt to optimise treatment with this most commonly used class of antibiotics, whose effectiveness is challenged by increasing resistance rates. The pharmacokinetics of β-lactam antibiotics in critically ill patients is profoundly altered secondary to an increased volume of distribution and the presence of altered renal function, including augmented renal clearance. This may lead to a significant decrease in plasma concentrations of β-lactam antibiotics. As a consequence, low pharmacokinetic/pharmacodynamic (PK/PD) target attainment, which is described as the percentage of time that the free drug concentration is maintained above the minimal inhibitory concentration (MIC) of the causative organism (fT>MIC), has been documented for β-lactam treatment in these patients when using standard intermittent bolus dosing, even for the most conservative target (50% fT>MIC). Prolonged infusion of β-lactams has consistently been shown to improve PK/PD target attainment, particularly in patients with severe infections. However, evidence regarding relevant patient outcomes is still limited. Whereas previous observational studies have suggested a clinical benefit of prolonged infusion, results from two recent randomised controlled trials of continuous infusion versus intermittent bolus administration of β-lactams are conflicting. In particular, the larger, double-blind placebo-controlled randomised controlled trial including 443 patients did not demonstrate any difference in clinical outcomes. We believe that a personalised approach is required to truly optimise β-lactam treatment in critically ill patients. This may include therapeutic drug monitoring with real-time adaptive feedback, rapid MIC determination and the use of antibiotic dosing software tools that incorporate patient parameters, dosing history, drug concentration and site of infection. Universal administration of β-lactam antibiotics as prolonged infusion, even if supported by therapeutic drug monitoring, is not yet ready for "prime time", as evidence for its clinical benefit is modest. There is a need for prospective randomised controlled trials that assess patient-centred outcomes (e.g. mortality) of a personalised approach in selected critically ill patients including prolonged infusion of β-lactams compared with the current standard of care.