Clinical Laboratory–Based Assay Methodologies May Underestimate and Increase Variability of Vancomycin Protein Binding in Hospitalized Patients

Abstract

To evaluate and compare the percent protein binding of vancomycin in hospitalized patients by using a clinical or research laboratory-derived assay methodology, and to evaluate potential patient characteristics accounting for alterations in protein binding. Prospective noninterventional cohort study. Single-center tertiary care medical center. A total of 55 hospitalized adults who were receiving vancomycin for a suspected or documented infection between August and November 2011 and required therapeutic drug monitoring. Vancomycin protein-binding studies were conducted by using ultracentrifugation of 63 blood samples collected from the 55 patients for therapeutic drug monitoring as part of clinical practice. Total and free drug concentrations were assayed in the research laboratory by using high-performance liquid chromatography (HPLC) and in the clinical laboratory by using fluorescence polarization immunoassay (FPIA). Multivariate linear regression analysis was performed to identify patient variables that were predictive of vancomycin protein binding. The average protein binding was statistically significantly lower and more variable when assayed by FPIA compared with HPLC (mean±SD 47.3±13.0% vs 54.6±9.5%, p<0.001). Multivariate analyses showed that after controlling for days of vancomycin therapy, patients in the intensive care unit (ICU) had a protein binding value that was 8.4% lower than non-ICU patients (p=0.005). Using research laboratory-based HPLC methodology, we identified an average vancomycin protein binding of 54.6% with considerably less variability than reported in the literature using clinical-based assay methodologies. Further, we identified patient factors that may likewise have an impact on this value. Future studies of vancomycin protein binding should consider use of a nonclinical assay to minimize methodological-induced variability.