Antithrombin III in patients with severe sepsis: a pharmacokinetic study

Objectives: To evaluate the safety, pharmacokinetics, and the practicabilit y of two different antithrombin III (AT III) high-dose regimens in patients with severe sepsis. Design: Prospective, open, randomized, 2 parallel groups, multinational cli nical trial. Setting: Eleven academic medical center intensive care units (ICU) in Austr ia, Belgium, Denmark, Germany, Norway and Sweden. Patients: Thirty-three patients with severe sepsis who received standard su pportive care and antimicrobial therapy, in addition to the administration of AT III. Interventions: Patients received an intravenous loading dose of 6,000 IU AT III followed by either intermittent: bolus infusions of 1,000 IU AT III ev ery 4 h or a continuous infusion of 250 IU AT III/h for 4 days, resulting i n a total dose for both dosage regimens of 30,000 IU AT III. Measurements: All patients were evaluated for safety and all but one for ph armacokinetics. Results and conclusions: The administration of AT III was safe and well tol erated. The overall 28-day all-cause mortality was 30 % (43 % intermittent bolus infusions, 21 % continuous infusion). The mean probability of dying a ccording to the SAPS II was 48 %. The difference in mortality between both groups was within the range of chance. AT III plasma levels were elevated f rom low baseline levels to above 120 % soon after onset of AT III therapy a nd remained at these levels for the treatment phase of 4 days. Functional a nd immunologic levels of AT III corresponded very well. With an overall med ian volume of distribution of 4.51 (range: 2.4-6.51), AT III only moderatel y extended beyond plasma. The overall median elimination half-life was 18.6 h (range: 5.1-37.4). Overall, median response was 1.75 % per IU/kg (range: 1.15-2.8). The variability of elimination parameters was quite noteworthy (CV = 41-59 %), whereas distribution-related parameters showed a moderate variability ( CV = 24 %). In spite of this variability, both high-dose IV regimens reliab ly provided AT III levels above 120 % for ail but one patient. An increased mortality was observed for patients with a distribution volume exceeding 4 .51 (or a response < 1.7 % per IU/kg). AT III distribution volumes above 4.51 might indicate a capillary leak phen omenon The continuous infusion regimen was slightly preferred by the invest igators with regard to practicability.