A randomized, controlled, multicenter trial of the effects of antithrombin on disseminated intravascular coagulation in patients with sepsis

Patients with a diagnosis of DIC (JAAM DIC score ≥4) with sepsis and levels of antithrombin ranging from 50 to 80% were eligible for this study. Patients who met the following criteria were excluded: (1) less than 15 years of age; (2) a history of hematopoietic malignancy; (3) a history of liver cirrhosis classified as Child-Pugh grade C; (4) receiving concomitant treatment with chemotherapies or irradiation; (5) a history of known clotting disorders or receiving anticoagulant therapy; (6) in an early phase of trauma or burn injuries; and (7) a life expectancy of less than 28 days. Web-based randomization with an allocation ratio of 1:1 for the control and antithrombin groups was generated by the University Hospital Medical Information Network (UMIN) center. Neither the physicians nor the patients were blinded to the treatment assignment.

Immediately after the patients met the inclusion criteria, they were randomly assigned to either a group receiving antithrombin at a dose of 30 IU/kg (given over 60 minutes) per day for three days, or to the control group with no intervention. After randomization, antithrombin was promptly administered (day 0). The participating centers were allowed to freely select from the three available antithrombin concentrates used in our country (CSL Behring, Japan Blood Products Organization, Nihon Pharmaceuticals Co., Ltd.). Three days later, antithrombin was administered after the taking of a blood sample for evaluation in the morning. During the three days of antithrombin administration, the use of drugs that affect blood coagulation and fibrinolysis was contraindicated. These drugs are as follows: unfractionated heparin, low-molecular-weight heparin, danaparoid sodium, nafamostat mesilate, ulinastatin, and soluble thrombomodulin. However, the use of unfractionated heparin to flush vascular catheters, nafamostat mesilate for renal replacement therapy and gabexate mesilate (2 g/day fixed dose) was allowed. No patients received activated protein C. Packed red blood cell (PRBC) concentrate, platelet concentrate and fresh frozen plasma were transfused based on the 2008 recommendations of the Surviving Sepsis Campaign guidelines [19]. Substitution therapy for DIC with fresh frozen plasma was also allowed when the prothrombin time (activity) was <30%. Fibrinogen concentrate and prothrombin complex concentrate are not permitted for the treatment of DIC in this country.

SIRS, sepsis, severe sepsis and septic shock were defined according to the American College of Chest Physicians/Society of Critical Care Medicine consensus conference [20]. Tissue hypoperfusion was defined as blood lactate level ≥2.0 mmol/L. Arterial hypotension was considered to be present when the systolic blood pressure was <90 mmHg. The disease severity of the patients was evaluated according to the acute physiology and chronic health evaluation (APACHE) II score determined at the time of enrollment [21]. Organ dysfunction was assessed according to the sequential organ failure assessment (SOFA) score [22]. A DIC diagnosis was made based on the JAAM DIC diagnostic criteria [23, 24]. Overt DIC scores calculated based on the International Society on Thrombosis and Haemostasis (ISTH) scoring system were also used [25]. The fibrin/fibrinogen degradation product (FDP) was used as a fibrin-related marker for the ISTH criteria. No increase, moderate increase and strong increase were defined as FDP values of <10, 10 < FDP <25, and >25 mg/L, respectively. When the total score was ≥4 and ≥5, a diagnosis was established using the JAAM and ISTH criteria, respectively. The JAAM DIC scoring system is presented in Table 1. DIC recovery was defined as a JAAM DIC score on day 3 of less than 4. Major bleeding complications included intracranial bleeding and the transfusion of >6 U of PRBC concentrate (approximate volume, 1,200 mL) within 24 hours [7]. In our country, the volume of 1 U PRBC is about half of that in other countries.

The primary efficacy end point was recovery from JAAM DIC on day 3. The definition of ‘recovery’ has been described in the previous paragraph. The secondary efficacy end point was 28-day all-cause mortality. All end points were analyzed on an intention-to-treat basis. The following variables were evaluated or measured from day 0 (the day of inclusion) to day 28. (A) JAAM DIC score, ISTH DIC score and SOFA score; (B) platelet count, prothrombin time (ratio, international normalized ratio (INR), activity), fibrinogen, FDP and antithrombin; (C) lactate; (D) APACHE II score; (E) thrombin-antithrombin complex (TAT), soluble fibrin, plasmin-alpha2-plasmin inhibitor complex (PIC), D-dimer and plasminogen activator inhibitor-1 (PAI-1); and (F) outcome. The platelet count and global coagulation and fibrinolysis markers were measured at each hospital, and the molecular markers of coagulation were measured by an independent analysis institution (SRL Medisearch Inc., Tokyo, Japan). The time schedule of these evaluations in the trial is shown in Table 2. All adverse events including bleeding complications were observed and recorded until seven days after the start of the administration of antithrombin. Routine intensive care was not specified by the study protocol and was carried out according to the current practice in each center.

The sample size (100 patients for each arm) was calculated to detect a 20% reduction in expected mortality in the antithrombin arm based on an assumed control mortality of 40% with a two-sided alpha error of 0.05 and a statistical power of 0.8. The interim analysis was planned to examine 60 patients for each arm. The measurements are expressed as the mean ± standard deviation (SD) unless otherwise indicated. The IBM SPSS 20.0 for MAC OSX software program (IBM Japan, Tokyo, Japan) was used for the statistical analyses and calculations. Comparisons between two groups were made with unpaired Student’s t test or Mann-Whitney’s U test, and either the chi-square test or Fisher’s exact test was used when required. To compare the time courses of the two groups, two-way repeated measures of analysis of variance (ANOVA) were applied. Differences with a two-sided P value of <0.05 were considered to be statistically significant.

The primary efficacy population consisted of 60 patients who were randomly assigned in equal populations of 30 to either the control group or the antithrombin group; however, two patients in the antithrombin group were associated with protocol violations (failure to start antithrombin treatment). Ultimately, 30 patients in the control group and 30 patients in the antithrombin group were analyzed by the intention-to-treat method. The patients were well matched at study entry for age, sex and APACHE II, SIRS, SOFA and DIC scores (Table 3). The two groups were also well matched with respect to background diseases, sites of infection and microbial types (Additional files 1 and 2). The use of nafamostat mesilate for anticoagulation during renal replacement therapy and the requirements for gabexate mesilate and substitution of blood components were equally distributed between the two groups (Table 3).

The baseline antithrombin levels were equal between the two groups (Table 3 and Figure 1). These levels were unchanged in the control group until day 3. In contrast, the patients in the antithrombin group exhibited an elevated mean antithrombin level on day 3 of 107.6 ± 24.5% (Figure 1). As shown in Figure 2, the DIC scores, especially the ISTH overt DIC scores, in the antithrombin group significantly improved compared with those observed in the control group on day 3. In the antithrombin group, the DIC recovery rate was 53.3% (16/30), which was more than double that of the control group of 20.0% (6/30) (Figure 3). In contrast, the SOFA scores on day 3 (5.4 ± 3.3 vs. 5.3 ± 3.8, P = 0.665) and the 28-day and hospital mortality rates did not differ between the control and antithrombin groups (Table 3) (Figure 4). The use of gabexate mesilate did not affect the outcomes of the patients (Table 4). There were no significant differences in the SIRS criteria or global markers of coagulation and fibrinolysis between the control and antithrombin groups; however, the platelet counts in the antithrombin group were higher compared with those in the control group (Table 5). The administration of antithrombin did not affect the levels of TAT, soluble fibrin, PIC, D-dimer or PAI-1 on day 3 in the two groups (data not shown).

The incidence of minor bleeding complications did not differ between the control and antithrombin groups (Table 6). There were no overlaps of these bleeding complications among each patient. Neither major bleeding nor any other adverse events were observed in association with antithrombin treatment during the observation period.

Due to the strict setup conditions in the control group with no intervention, small numbers of patients were included. Therefore, this study lacks statistical power. In addition, this study did not have a double-blind placebo-controlled design. The study population included patients with sepsis, severe sepsis and septic shock, but the ratio of patients with sepsis was high in both groups, which may have resulted in a lower mortality rate than expected, which may have led to the lack of efficacy of antithrombin observed in the second end point of the study. The possibility of bias cannot be confidently rejected, and it has been suggested that patients who do not start the allocated intervention should not be included in an intention-to-treat analysis [36]. Because of the small number of the patients included, two patients with this type of protocol violation may lead to a substantial bias in the interpretation of the results. Therefore, we added the results of the per-protocol analysis as Additional files (Additional files 3 and 4).