Background: Optimization of the timing of appropriate antibiotics is crucial to improve the management of patients in severe sepsis and septic shock. Vancomycin is commonly used empirically in cases of nosocomial infections in critically ill patients. Therefore, early optimization of vancomycin pharmacokinetics is likely to improve outcomes.
Objective: To evaluate a pharmacokinetic program to predict serum vancomycin concentrations in accordance with administered dose, weight, height, and creatinine clearance in a critically ill population.
Methods: We conducted a prospective observational single-center study in a 45-bed intensive care unit (ICU). All patients hospitalized in the ICU requiring intravenous treatment with vancomycin for a suspected infection were enrolled. The modalities of vancomycin therapy and the monitoring of serum concentrations were left to the discretion of the treating clinician. We compared the measured serum vancomycin concentrations with those predicted by the MM-USCPACK program and analyzed the factors influencing the prediction.
Results: Fifty-four intravenous vancomycin courses were administered in 48 critically ill patients over the 3-month study. The precision was considered acceptable, based on a relative precision equal to 8.9% (interquartile range 3.5-18.9%) and the relative bias for all predictions was equal to -1.3%. Overall, 77.3% of predictions were within 20% of observed concentrations; factors correlating with a poorer prediction were a change in renal function, obesity, and the magnitude of organ dysfunction on initiation of vancomycin (expressed by a Systemic Organ Failure Assessment score >11).
Conclusions: The MM-USCPACK program is a useful and reliable tool for prediction of serum vancomycin concentrations in patients hospitalized in ICU and likely reflects the close monitoring of renal function in this setting. For some patients (more severely ill, obese, or significant change in renal function during vancomycin therapy), predictions were less precise.