Urea kinetics and dialysis treatment time predict vancomycin elimination during high-flux hemodialysis

Background: Hemodialysis sessions with high-flux filters ask for a reconsideration of the kinetics of xenobiotics. The aim of this study was to analyze whether individual high-flux hemodialysis treatment parameters are of predictive value for dosing guidelines, with use of vancomycin as a model compound. Methods: Twenty-six patients receiving high-flux hemodialysis were studied prospectively. After an intravenous infusion of 1000 mg or 500 mg vancomycin, respectively, six to eight blood samples were collected within a period of 5 to 9 days, including one hemodialysis session. Serum vancomycin concentrations were measured by HPLC. Nonlinear mixed-effects modeling (NONMEM) was used to fit a two-compartment population pharmacokinetic model to the data of 20 patients; the data of the remaining six patients (group II) were used for a prospective evaluation of the model. Results: A linear relationship was found between vancomycin filter clearance (CLDV) and urea filter clearance (CLDBUN), derived from Kt/V (the product of urea clearance [K] and dialysis treatment time [t], standardized for the urea volume of distribution [V]). Mean (coefficient of variation) steady-state volume of distribution was 1.05 L/kg (22%), CLDV was 0.336.CLDBUN (13%), and residual interdialytic clearance was 2.25 ml/min (90%) in patients with creatinine clearance values (CLCR) below 2 ml/min and 2.25 ml/min + 0.59.CLCR (32%) in patients with CLCR values above 2 ml/min. The model predicted predialysis vancomycin concentrations before the first and the second postinfusion dialysis session in the six patients of group II, with a deviation of 1.8 +/- 1.0 mg/L and 0.8 +/- 0.5 mg/L, respectively. Conclusion: The described population pharmacokinetic model allows individualization of vancomycin dosing intervals in patients receiving hemodialysis, based on patient characteristics and urea kinetic modeling.