现在的位置: 首页研究点评, 进展交流>正文
[JAMA发表述评]:重组人可溶性血栓调节蛋白治疗脓毒症相关凝血病:另一项阴性脓毒症试验?
2019年06月11日 研究点评, 进展交流 暂无评论

Editorial May 19, 2019

Recombinant Human Soluble Thrombomodulin in Patients With Sepsis-Associated Coagulopathy: Another Negative Sepsis Trial?

Tom van der Poll

JAMA. Published online May 19, 2019. doi:10.1001/jama.2019.5792

Sepsis is a heterogeneous syndrome caused by an unbalanced host response to an infection resulting in organ dysfunction.1 Sepsis is associated with activation of the coagulation system, and 30% to 60% of patients have disseminated intravascular coagulation (DIC).2 Fulminant DIC is characterized by` activation of the coagulation system combined with an impaired function of endogenous anticoagulant mechanisms resulting in widespread microvascular thrombosis and concurrent hemorrhage secondary to consumption of clotting factors and platelets.3,4 Strong and bimodal interactions exist between coagulation and inflammation, wherein coagulation factors enhance inflammation and vice versa.4 Observational studies suggest a link between the severity of coagulopathy, organ dysfunction, and death in patients with sepsis. In addition, in several preclinical sepsis models, inhibition of coagulation reduces both systemic inflammation and mortality.3,4 These findings led to large randomized trials of several anticoagulant agents in patients with sepsis, most notably recombinant human activated protein C (APC), tissue factor pathway inhibitors, and antithrombin.5 None of these interventions improved survival, with the exception of APC in the PROWESS trial and those results were not upheld by subsequent trials.5

In this issue of JAMA, Vincent and colleagues report the long-awaited results of the SCARLET (Sepsis Coagulopathy Asahi Recombinant LE Thrombomodulin) phase 3 trial.6 SCARLET was designed to determine the effect of another anticoagulant, soluble human recombinant thrombomodulin (ART-123), on 28-day mortality in patients with sepsis-associated coagulopathy and concomitant cardiovascular failure, respiratory failure, or both. The study failed to demonstrate a reduction in the primary end point of 28-day all-cause mortality, raising the question of why the study was conducted given prior results with other anticoagulants. The answer is somewhat nuanced, but, in short, thrombomodulin has a different mode of action that offers important theoretical advantages.

The soluble human recombinant thrombomodulin evaluated in this study is composed of the extracellular portion of the transmembrane receptor thrombomodulin, which entails a C-type lectin domain and a region with 6 epidermal growth factor–like repeats.7 The epidermal growth factor–like repeats express a high-affinity binding site for the central coagulation protease thrombin. Thrombin complexed with thrombomodulin cannot cleave fibrinogen or activate factor V or factor XIII but rather accelerates the conversion from protein C to APC, a potent anticoagulant by virtue of its capacity to inactivate clotting factors Va and VIIIa.7 Thus, soluble human recombinant thrombomodulin has 2 anticoagulant modes of action: direct, through inhibition of the procoagulant activity of thrombin, and indirect by promoting protein C activation.

In addition, after binding by thrombomodulin, thrombin is unable to cleave protease-activated receptor-1 or -3 on cells, thereby losing its proinflammatory properties. Thrombin-thrombomodulin also accelerates the activation of thrombin-activatable fibrinolysis inhibitor, which inhibits fibrinolysis and inflammatory mediators, such as bradykinin and the anaphylatoxins C3a and C5a.7 Independent of thrombin binding, the C-type lectin-like domain of thrombomodulin has anti-inflammatory and cytoprotective effects, including inhibition of mitogen-activated protein kinase pathways and nuclear factor-κ activation, impediment of adhesion of neutrophils to endothelial cells, and inhibition of high-mobility group box 1 protein, an alarmin implicated in sepsis pathogenesis.7 The C-type lectin domain can also impede the binding of lipopolysaccharide, the proinflammatory component of the gram-negative bacterial cell wall, to its receptor CD14, thereby mitigating cell activation. Thus, thrombomodulin has multiple anticoagulant, anti-inflammatory, and cytoprotective effects that could, theoretically, be beneficial in patients with sepsis. Consistently, preclinical studies have documented favorable effects of soluble thrombomodulin in experimental sepsis associated with improvements of coagulation abnormalities, inflammation, and organ dysfunction.8

The first clinical indication that recombinant thrombomodulin might be a valuable adjunctive therapy for patients with sepsis came from a phase 3 trial conducted in Japan of 227 patients with DIC secondary to hematologic malignancy or infection.9 In this study, the effects of recombinant thrombomodulin were compared with heparin using resolution of DIC as a primary end point.9 DIC was resolved in 66.1% of patients in the recombinant thrombomodulin group compared with 49.9% of patients in the heparin group, but there was no effect on the secondary end point of mortality (21.9% vs 25.7%). These data led to the registration of recombinant thrombomodulin (ART-123) for the management of DIC in Japan, where it has been in clinical use since 2008. The first randomized placebo-controlled study that examined the effect of soluble human recombinant thrombomodulin in patients with sepsis (n = 741) was a phase 2b trial published in 2013, in which thrombomodulin had a nonsignificant effect on the 28-day mortality rate (17.8% vs 21.6% in the placebo group).10

The inclusion criteria of the SCARLET study by Vincent et al6 were based on a post hoc analysis of this 2013 trial that suggested the greatest potential effect of recombinant thrombomodulin on mortality may be among patients with sepsis and a prolonged international normalized ratio (INR), reduced platelet count, and at least 1 organ dysfunction.10 Thus, the SCARLET trial sought to include only patients with sepsis-associated coagulopathy, defined as an INR greater than 1.40 without other known etiology and a platelet range of 30 to 150 × 109/L or a decrease in platelet count greater than 30% within 24 hours. The study enrolled 800 patients who received the assigned intervention (395 in the recombinant thrombomodulin group and 405 in the placebo group) over approximately 5.5 years. Recombinant thrombomodulin did not improve the primary efficacy end point of the study; the 28-day all-cause mortality rate was 26.8% in the thrombomodulin group vs 29.4% in the placebo group (P = .32), an absolute risk reduction of 2.55% (95% CI, −3.68% to 8.77%). Recombinant thrombomodulin was shown to be biologically active through lower plasma levels of the coagulation activation markers D-dimer, prothrombin fragment F1.2, and thrombin-antithrombin complexes in patients who received thrombomodulin vs patients who received the placebo. Whether recombinant thrombomodulin influenced inflammatory responses was not determined.

Why did this trial demonstrate no important differences between groups? Despite its demonstrated biological activity, soluble thrombomodulin may simply not be beneficial or its effect may not be strong or broad enough to reduce mortality in patients with sepsis, who may experience an array of host response aberrations that are unrelated to the mechanism of action of this drug. There are other potential explanations. Recombinant thrombomodulin seemed to be most beneficial for patients who did not receive concurrent heparin (low dose for deep vein thrombosis prophylaxis or flushes); in this subgroup (n = 384), the mortality rate was 25.0% in patients treated with recombinant thrombomodulin vs 31.3% in patient who received the placebo, a difference of 6.25% (95% CI, −2.72% to 15.22%). A possible explanation could be that heparin can inhibit the interaction between thrombomodulin and thrombin,11 thereby interfering with the main mechanism of action of recombinant thrombomodulin. It is also possible that the trial was underpowered due to a mortality rate in the placebo group that was higher than the assumed value used for the sample size calculation. Notably, mortality rates in both treatment groups were higher during the second half of the trial, which was not discussed by the authors but could be related to recruitment of additional study sites in response to low enrolment of patients. Intensive care units from many different areas of the world eventually participated, with likely differences in standard of care and one-third of sites enrolling only 1 patient.

The authors presented findings from post hoc analyses that provide some insight into which subgroups of patients might potentially benefit from recombinant thrombomodulin treatment and could be evaluated in future trials. At the time of the first study drug administration, 18.3% of patients no longer met the INR inclusion criterion and 23.5% had normal platelet counts, which was most likely related to the long period (a 40-hour window was implemented after a protocol amendment was applied during the study) from a qualifying INR measurement until first dosing. Among patients with coagulopathy at the time of the first drug administration, the 28-day mortality rate was 82 of 307 patients (26.7%) in the recombinant thrombomodulin group vs 105 of 327 (32.1%) in the placebo group (a nonsignificant absolute difference of 5.40% [95% CI, −1.68% to 12.48%]). Although this analysis was not planned, the observation is consistent with the previous phase 2b study.10 Furthermore, recombinant thrombomodulin was also associated with nonsignificant reductions in mortality among patients with greater activation of the coagulation system, as indicated by plasma thrombin-antithrombin levels greater than 10 ng/mL (n = 489; 29.0% in the thrombomodulin group vs 35.7% in the placebo group) or protein C concentrations of 40% or lower (n = 510; 26.5% in the thrombomodulin group vs 32.8% in the placebo group). Thus, failure to identify patients who were more likely to respond favorably to recombinant thrombomodulin may partially explain the lack of efficacy for reducing mortality.

In the SCARLET trial, the use of recombinant thrombomodulin was not associated with an increased occurrence of serious major bleeding, although the absolute number of serious bleeding events was higher in the recombinant thrombomodulin group than the placebo group (5.8% vs 4.0%; nonsignificant difference). Previous studies also point to a relatively good safety profile of recombinant thrombomodulin.9,10,12 The use of other anticoagulants in patients with sepsis has been associated with increased bleeding.5 For example, the use of recombinant human APC was associated with a relative risk of severe bleeding of approximately 1.5 across different trials.5

What could be next for soluble human recombinant thrombomodulin in patients with sepsis? The results of the SCARLET trial by Vincent et al5 indicate that soluble human recombinant thrombomodulin (ART-123) has no demonstrable effect in reducing mortality in the study population evaluated. The manufacturer of the drug has posted the protocol for a second phase 3 trial (SCARLET-2) on clinicaltrials.gov (NCT03517501) with inclusion criteria that are similar to the original SCARLET study. If soluble thrombomodulin is studied in a subsequent trial of sepsis patients, it may be important to consider modifying these enrollment criteria. By enriching the study population of a future trial to include patients who may be more likely to benefit, especially by using narrow time windows between blood tests defining coagulopathy and first drug administration and possibly by excluding heparin as concurrent therapy, such a study may provide additional insights on the treatment of patients with sepsis-associated coagulopathy.

给我留言

您必须 [ 登录 ] 才能发表留言!

×
腾讯微博