Serum linezolid concentrations are reduced in critically ill patients with pulmonary infections: A prospective observational study.

Abstract

The concentration-time profile of linezolid varies considerably in critically ill patients. Question of interest is, if the site of infection influences linezolid serum concentrations. 68 critically ill patients, treated with linezolid, were included. The concentration-time-profile for linezolid was determined using maximum a-posteriori predictions. A trough concentration (Cmin) between 2 and 10 mg/L was defined as the target. A generalized linear model (GLM) was established to evaluate potential covariates. The indications for linezolid therapy were in descending order: peritonitis (38.2%), pneumonia (25.0%), infectious acute respiratory distress syndrome (ARDS) (19.1%), and other non-pulmonary infection (17.7%). 27.2 and 7.9% of Cmin were subtherapeutic and toxic, respectively. In the GLM, ARDS (mean: -2.1 mg/L, CI: -3.0 to -1.2 mg/L) and pneumonia (mean: -2.2 mg/L, CI: -2.8 to -1.6 mg/L) were significant (p < 0.001) determinants of Cmin. Patients with ARDS (mean: 2.3 mg/L, 51.2% subtherapeutic, 0.0% toxic) and pneumonia (mean: 3.5 mg/L, 41.5% subtherapeutic, 7.7% toxic) had significantly (p < 0.001) lower Cmin than those with peritonitis (mean: 5.5 mg/L, 14.4% subtherapeutic, 9.3% toxic) and other non-pulmonary infection (mean: 5.2 mg/L, 3.3% subtherapeutic, 16.5% toxic). Linezolid serum concentrations are reduced in patients with pulmonary infections. Future studies should investigate if other linezolid thresholds are needed in those patients due to linezolid pooling in patients´ lung.