Vancomycin (VAN) protein binding in plasma is influenced by illness and age; hence, doses titrated according to total concentrations are fraught. In this study, model-estimated free VAN concentrations (EFVC) were compared with assumed free VAN concentrations (AFVC) in neonates, children, and adults in the intensive care unit and those on dialysis. Patient cohorts were identified from the hospital database. Demographics, clinical characteristics, total VAN concentrations, and laboratory variables were obtained from electronic health records. EFVC was derived from 6 models identified in the literature. For all models, total VAN concentration was the most important predictor; other predictors included albumin, total protein, and dialysis status. The AFVC was calculated as 50% of the total concentration (ie, assumption of 50% bound). Differences between EFVC and AFVC in adults were insignificant; however, differences in pediatric intensive care unit patients, according to 2 different models, were significant: mean ± SD = 4.1 ± 1.58 mg/L and 4.7 ± 2.46 mg/L ( P < 0.001); the percentages within the free VAN trough range = 30.4% versus 55.1% and 30% versus 55.1%; and the supratherapeutic percentages = 65.2% versus 31.9% and 66.7% versus 31.9%, respectively. In neonates, the difference between EFVC and AFVC was mean ± SD = 6.9 ± 1.95 mg/L ( P < 0.001); the percentages within the free VAN trough range for continuous and intermediate dosing were 0% versus 81.3% and 14.3% versus 71.4%, and the supratherapeutic percentages were 100% versus 6.25% and 71.4% versus 0%, respectively. The fraction of free unbound VAN is higher in sick children and neonates than in adults. Therefore, total VAN concentrations do not correlate with the pharmacologically active free VAN concentrations in the same manner as in adults. Adjusting VAN doses in neonates and children to target the same total VAN concentration as the recommended therapeutic range for adults may result in toxicfree concentrations.
Read full abstract