This study was performed as an addition to a double-blind, placebo-controlled, multicenter clinical phase III trial in patients with severe sepsis (the KyberSept trial) [11
]. In accordance with the Institutional Review Board of the University of Vienna, patients were included in the study and their written informed consent was obtained after adequate recovery.
Patients suffering from severe sepsis were assigned by a telephone randomization service to receive either AT (Aventis Behring GmbH, Marburg, Germany) (n = 17) or placebo solution (1% human albumin) (n = 16). The treatment group received 6,000 IU AT as a bolus infusion followed by a maintenance dose of 250 IU/hour over four days. Standard therapy, such as antimicrobial therapy, respiratory or hemodynamic support, and fluid administration, was not influenced by the study and was at the physicians' discretion.
Patients with suspected severe sepsis were enrolled if they fulfilled the following criteria within a six hour time window prior to randomization: clinical evidence of sepsis with a suspected source of infection, body temperature (rectally or core) >38.5°C or <35.5°C, and leukocyte count >10,000/μl or <3,500/μl. Additionally, three out of the following six criteria had to be met: heart rate >100 beats/min, tachypnea >24 breaths/min or mechanical ventilation because of septic indication, systolic blood pressure <90 mmHg despite sufficient fluid replacement or the need for vasoactive agents to maintain systolic blood pressure >90 mmHg, thrombocytopenia with platelet counts <100,000/ml, elevated lactate levels or metabolic acidosis (that is to say, pH <7.3 or base excess above -10 mmol/l) not secondary to respiratory alkalosis, and oliguria with urine output <20 ml/hour despite sufficient fluid replacement. The exclusion criteria were, among others [11
], pregnancy and/or breastfeeding, presence of a condition other than sepsis anticipated to be fatal within 28 days, history of hypersensitivity to the study medication, treatment with an AT concentrate within the previous 48 hours, treatment with heparin exceeding 10,000 IU/day low-molecular-weight heparin, known bleeding disorders or ongoing massive surgical bleedings, nonsteroidal anti-inflammatory drugs in anti-inflammatory doses within the previous two days, platelet count <30,000 ml, pre-existing dialysis-dependent renal failure, end-stage liver disease, or enrollment in another clinical trial within the previous 30 days.
Routine laboratory measurements were carried out as usual. The coagulation profile was assessed at least three times a day during treatment with the study medication. Routine AT measurements during the first 14 days of the study were not performed, to ensure the double-blinded fashion of the study. Samples for determination of AT (activity and antigen) were drawn before the bolus infusion, 24 hours after the start of bolus infusion, and then daily. These samples were centrifuged at 4°C and stored at -80°C for central measurements. The AT activity was provided by Aventis Behring GmbH after trial finalization. The platelet count, plasma fibrinogen levels, prothrombin time, and activated partial thromboplastin time were assessed at baseline and three times daily.
In addition to standard coagulation tests we performed two thromboelastograph scans before the bolus infusion and then daily (TEG®
; Haemoscope, Skokie, IL, USA), each using 300 μl whole blood recalcified with 40 μl of 0.645% CaCl2
. TEG measurements were performed after incubating the blood samples with heparinase, to exclude heparin effects on TEG. The antibody fragment abciximab (ReoPro®
; Centocor, Leiden, The Netherlands) was added to one assay to evaluate platelet function [14
]. TEG permits a reliable, global assessment of hemostatic function correlating to routine coagulation tests and, most importantly, to postoperative blood loss and the incidence of thrombotic complications [5
]. Liquid whole blood transmits little or no torque in TEG, producing no amplitude on the TEG tracing even in blood samples with high viscosity [17
]. As the blood clots, fibers composed of fibrin and platelets form, producing an increasing amplitude. Platelets provide a phospholipid surface for coagulation reactions in the standard TEG tracing and thus promote the formation of fibrin [17
]. Furthermore, platelets bind to fibrinogen and modulate the viscoelastic properties of the clot via the platelet surface receptor glycoprotein IIb/IIIa [15
]. The glycoprotein IIb/IIIa receptor can be sufficiently inhibited by the antibody-fragment abciximab. Fibrinogen is soluble until thrombin binds to the central region, which produces proteolysis at the N-terminus, releasing fibrinopeptide A and B and fibrin monomer. This release process exposes other regions of the molecule to interact with other activated fibrin molecules for polymerization of the fibrin network. The end point of this cascade is fibrin.
The reaction time (r
), coagulation time (k
), alpha angle (α), maximum amplitude (MA), and abciximab MA were assessed. Values for r
are expressed in millimeters – as the chart speed is 2 mm/min, the time in minutes is equal to the distance in millimeters divided by two (normal ranges: r
= 10–19 mm, k
= 4–10 mm). The α value measures the speed of fibrin build-up and cross-linking, which resembles speed of clot strengthening (normal range: α = 44–56°). The MA measures the maximal clot strength, which is dependent on platelet function and, to a lesser extent, on fibrinogen level (normal range: MA = 50–64 mm). Whereas the correlation between standard MA and fibrinogen levels is usually weak, the modification of TEG with the antibody fragment abciximab results in a good correlation between fibrinogen levels and abciximab-modified MA [15
]. Because the resulting MA of the abciximab-modified TEG is an estimation of the contribution of fibrinogen to clot strength, the difference of the standard MA and the abciximab-modified MA primarily reflects platelet function [15
]. The TEG tracing of hypercoagulable blood typically shows a shorter reaction time, with a higher MA and a steeper α value than normal.
An unpaired Mann–Whitney nonparametric test with Bonferroni correction for multiple testing was performed to compare for differences between the AT and placebo group. P < 0.05 was considered statistically significant, and all data are presented as the mean ± standard deviation or as the median (minimum–maximum) unless otherwise indicated.