Risk factors for target non-attainment during empirical treatment with β-lactam antibiotics in critically ill patients

Purpose: Risk factors for β-lactam antibiotic underdosing in critically ill patients have not been described in large-scale studies. The objective of this study was to describe pharmacokinetic/pharmacodynamic (PK/PD) target non-attainment envisioning empirical dosing in critically ill patients and considering a worst-case scenario as well as to identify patient characteristics that are associated with target non-attainment. Methods: This analysis uses data from the DALI study, a prospective, multi-centre pharmacokinetic point-prevalence study. For this analysis, we assumed that these were the concentrations that would be reached during empirical dosing, and calculated target attainment using a hypothetical target minimum inhibitory concentration (MIC), namely the susceptibility breakpoint of the least susceptible organism for which that antibiotic is commonly used. PK/PD targets were free drug concentration maintained above the MIC of the suspected pathogen for at least 50 % and 100 % of the dosing interval respectively (50 % and 100 % f T >MIC). Multivariable analysis was performed to identify factors associated with inadequate antibiotic exposure. Results: A total of 343 critically ill patients receiving eight different β-lactam antibiotics were included. The median (interquartile range) age was 60 (47-73) years, APACHE II score was 18 (13-24). In the hypothetical situation of empirical dosing, antibiotic concentrations remained below the MIC during 50 % and 100 % of the dosing interval in 66 (19.2 %) and 142 (41.4 %) patients respectively. The use of intermittent infusion was significantly associated with increased risk of non-attainment for both targets; creatinine clearance was independently associated with not reaching the 100 % f T >MIC target. Conclusions: This study found that-in empirical dosing and considering a worst-case scenario-19 % and 41 % of the patients would not achieve antibiotic concentrations above the MIC during 50 % and 100 % of the dosing interval. The use of intermittent infusion (compared to extended and continuous infusion) was the main determinant of non-attainment for both targets; increasing creatinine clearance was also associated with not attaining concentrations above the MIC for the whole dosing interval. In the light of this study from 68 ICUs across ten countries, we believe current empiric dosing recommendations for ICU patients are inadequate to effectively cover a broad range of susceptible organisms and need to be reconsidered. © 2014 Springer-Verlag and ESICM.