The objective of this study was to describe the effect of therapeutic drug monitoring (TDM) and dose adjustments of β-lactam antibiotics administered to critically ill patients undergoing continuous renal replacement therapy (CRRT) in a 30-bed tertiary intensive care unit (ICU). β-Lactam TDM data in our tertiary referral ICU were retrospectively reviewed. Clinical, demographic and dosing data were collected for patients administered β-lactam antibiotics while undergoing CRRT. The target trough concentration range was 1-10× the minimum inhibitory concentration (MIC). A total of 111 TDM samples from 76 patients (46 male) with a mean ± standard deviation age of 56.6 ± 15.9 years and weight of 89.1 ± 25.8 kg were identified. The duration of antibiotic therapy was between 2 days and 42 days. TDM identified a need for dose modification of β-lactam antibiotics in 39 (35%) instances; in 27 (24%) samples, TDM values resulted in decreasing the prescribed dose of β-lactam antibiotic whereas an increase in the prescribed dose occurred in 12 (11%) cases. In patients treated for hospital-acquired pneumonia and primary or secondary bacteraemia, the dose was required to be decreased in 10/25 (40%) and 7/46 (15%) cases, respectively, to attain target concentrations. β-Lactam TDM is a useful tool for guiding drug dosing in complex patients such as those receiving CRRT. Although over one-third of patients manifested concentrations outside the therapeutic range, most of these CRRT patients had excessive β-lactam concentrations.

This is the first population pharmacokinetic study of vancomycin in patients receiving PIRRT and we observed large pharmacokinetic variability. Empirically, weight-based doses that are appropriate for the duration of PIRRT, should be selected and supplemented with therapeutic drug monitoring.

In the context of such variable pharmacokinetics, a guideline approach to dosing remains elusive because of insufficient available data and, therefore, use of therapeutic drug monitoring should be considered advantageous where possible.

No abstract