Background: The objective is to study the population pharmacokinetics of lorazepam and midazolam in critically ill patients with acute renal failure who are treated with continuous venovenous hemofiltration (CVVH).
Methods: Twenty critically ill patients with acute renal failure on CVVH therapy were administered either lorazepam (n = 10) or midazolam (n = 10) by continuous infusion. CVVH was performed with an ultrafiltrate flow of 2 L/h with filtrate substitution in the predilution or postdilution mode. Blood flow through the 1.9-m 2 cellulose triacetate membrane filter was 180 mL/min. For 48 hours, multiple blood and ultrafiltrate samples were obtained for determination of concentrations of the drug and its metabolites.
Results: The pharmacokinetics of lorazepam is described best by a 1-compartment model. No significant covariates were identified. Total-body clearance was 6.4 L/h, and volume of distribution was 376 L. Ultrafiltration clearance was 0.31 L/h, equivalent to approximately 5% of total clearance. Average degree of plasma protein binding was 82.9% for lorazepam, with a sieving coefficient of 0.16 +/- 0.03. For lorazepamglucuronide, degree of plasma protein binding was 39.5%, and sieving coefficient was 0.48 +/- 0.07. The pharmacokinetics of midazolam is described best by a 1-compartment model. No significant covariates were identified. Total-body clearance was 8.5 L/h, and volume of distribution was 157 L. Clearance by ultrafiltration was 0.055 L/h, equivalent to approximately 0.7% of total clearance. Average degree of plasma protein binding was 95.8%, with a sieving coefficient of 0.04 +/- 0.03. For the metabolite 1-hydroxymidazolamglucuronide, average degree of plasma protein binding was 43.4%, with a sieving coefficient of 0.45 +/- 0.06.
Conclusion: Neither lorazepam nor midazolam is removed efficiently by CVVH. CVVH contributes significantly to the removal of the glucuronide metabolites lorazepamglucuronide and 1-hydroxymidazolamglucuronide.