Optimal patient selection for anticoagulant therapy in sepsis: an evidence‐based proposal from Japan

SummarySelecting an appropriate target population is essential to maximize survival benefits of anticoagulant therapy against sepsis. Our meta‐analysis of three populations with sepsis and nationwide observational study in Japan showed that anticoagulants improved mortality only in sepsis‐induced disseminated intravascular coagulation (DIC) but not in non‐DIC. This divergent effect was physiologically explained by host‐protective immune responses of local thrombosis, which are mandatory in the early stage of sepsis. Meanwhile, the lack of definitive evidence for survival benefit provided by several trials of sepsis‐induced DIC indicated that this condition was probably not the best target of anticoagulants. Our multicenter cohort study including only patients with sepsis‐induced DIC showed a survival benefit from recombinant thrombomodulin only in patients with high disease severity. Thus, we believe that the population with sepsis and DIC and high disease severity is the optimal target for anticoagulant therapy. Anticoagulant therapy without appropriate target selection should be avoided because of the increased risk of bleeding with no survival benefit. Selecting an appropriate target population is essential to maximize survival benefits of anticoagulant therapy against sepsis. Our meta‐analysis of three populations with sepsis and nationwide observational study in Japan showed that anticoagulants improved mortality only in sepsis‐induced disseminated intravascular coagulation (DIC) but not in non‐DIC. This divergent effect was physiologically explained by host‐protective immune responses of local thrombosis, which are mandatory in the early stage of sepsis. Meanwhile, the lack of definitive evidence for survival benefit provided by several trials of sepsis‐induced DIC indicated that this condition was probably not the best target of anticoagulants. Our multicenter cohort study including only patients with sepsis‐induced DIC showed a survival benefit from recombinant thrombomodulin only in patients with high disease severity. Thus, we believe that the population with sepsis and DIC and high disease severity is the optimal target for anticoagulant therapy. Anticoagulant therapy without appropriate target selection should be avoided because of the increased risk of bleeding with no survival benefit. Blood coagulation disorders, which invariably occur in patients with sepsis, play a key role in causing organ dysfunction and subsequent death. As a defense against these poor outcomes, anticoagulant therapies are expected to be beneficial based on the pathophysiology of sepsis. Despite the numerous studies focusing on anticoagulant therapies against sepsis, these therapies continue to remain a matter of dispute. The main purpose of this mini‐review was to summarize the findings of recent related articles along with our own research and to propose our strategy regarding patient selection for anticoagulant therapies in sepsis. Various conditions of critical illness are known to cause coagulation disorders; however, all populations with critical illness are not the optimal target of anticoagulant therapy. Recently, Allingstrup et al. conducted a systematic review and meta‐analysis of antithrombin use in critically ill patients, including those with sepsis, trauma, burns and obstetric diseases, and in premature infants and reported no significant survival benefit of antithrombin concentrate 1.Allingstrup M. Wetterslev J. Ravn F.B. Møller A.M. Afshari A. Antithrombin III for critically ill patients: a systematic review with meta‐analysis and trial sequential analysis.Intensive Care Med. 2016; 42: 505-20Crossref PubMed Scopus (68) Google Scholar. One explanation for the null effect of antithrombin in their analysis might be attributed to the type or heterogeneity of the patients. Our assessment of the effectiveness of anticoagulant therapy should be carefully focused on the target population of the intervention because the pathophysiological heterogeneity caused by the types of underlying disease intensely influences the efficacy of these therapies. Coagulation disorders induced by critical illness can be subdivided into several phenotypes according to the types of underlying disease. For example, the fibrinolytic phenotype is frequently observed in the acute phase of severe trauma and in patients with hematopoietic malignancies. The fibrinolysis/fibrinogenolysis and consumption coagulopathy induced by tissue‐type plasminogen activator in these patients can lead to severe bleeding disorders. In contrast, thrombotic phenotypes characterized by plasminogen activator inhibitor‐1‐induced inhibition of fibrinolysis and inadequate control of coagulation are linked to the late phase of trauma and sepsis and can result in microvascular thrombosis and subsequent multiple organ dysfunction syndrome. Therefore, the efficacy of anticoagulant therapy should be examined not in the overall heterogeneous population of critical illness (as in the case of the Allingstrup et al. meta‐analysis 1.Allingstrup M. Wetterslev J. Ravn F.B. Møller A.M. Afshari A. Antithrombin III for critically ill patients: a systematic review with meta‐analysis and trial sequential analysis.Intensive Care Med. 2016; 42: 505-20Crossref PubMed Scopus (68) Google Scholar), but in each underlying disease separately. Thus, could sepsis, one of the major underlying diseases of disseminated intravascular coagulation (DIC), be an optimal target for anticoagulant therapy? To answer this clinical question, many randomized controlled trials (RCTs) of anticoagulant therapies in sepsis were conducted over the past few decades. However, insufficient evidence was available from these trials to support the effectiveness of anticoagulant therapy in improving mortality. The lack of evidence on survival benefit from these RCTs led the Surviving Sepsis Campaign Guidelines 2016 not to recommend the use of any anticoagulant agents as adjunctive therapy for sepsis 2.Rhodes A. Evans L.E. Alhazzani W. Levy M.M. Antonelli M. Ferrer R. Kumar A. Sevransky J.E. Sprung C.L. Nunnally M.E. Rochwerg B. Rubenfeld G.D. Angus D.C. Annane D. Beale R.J. Bellinghan G.J. Bernard G.R. Chiche J.D. Coopersmith C. De Backer D.P. et al.Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock: 2016.Intensive Care Med. 2017; 43: 304-77Crossref PubMed Scopus (3706) Google Scholar. Also, van der Poll et al. claimed that the universal use of anticoagulants in sepsis patients cannot be recommended because of the risk of bleeding in their sharply defined review 3.van der Poll T. Opal S.M. Should all septic patients be given systemic anticoagulation? No.Intensive Care Med. 2017; 43: 455-7Crossref PubMed Scopus (17) Google Scholar. Meanwhile, much evidence from observational studies and post‐hoc subgroup analyses of RCTs has suggested that anticoagulant therapy might reduce mortality if used in the specific subset of sepsis. For example, the post‐hoc subgroup analysis of the KyberSept trial suggested that favorable treatment effects of antithrombin were observed only in sepsis patients with predicted mortality between 30% and 60% according to the baseline SAPS II score 4.Wiedermann C.J. Hoffmann J.N. Juers M. Ostermann H. Kienast J. Briegel J. Strauss R. Keinecke H.O. Warren B.L. Opal S.M. KyberSept InvestigatorsHigh‐dose antithrombin III in the treatment of severe sepsis in patients with a high risk of death: efficacy and safety.Crit Care Med. 2006; 34: 285-92Crossref PubMed Scopus (136) Google Scholar. Also in another post‐hoc analysis of the same KyberSept trial, the treatment effects of antithrombin were associated with a significant reduction in mortality only in the patients suffering from sepsis‐induced DIC but not in the non‐DIC patients 5.Kienast J. Juers M. Wiedermann C.J. Hoffmann J.N. Ostermann H. Strauss R. Keinecke H.O. Warren B.L. Opal S.M. KyberSept investigatorsTreatment effects of high‐dose antithrombin without concomitant heparin in patients with severe sepsis with or without disseminated intravascular coagulation.J Thromb Haemost. 2006; 4: 90-7Crossref PubMed Scopus (309) Google Scholar. Recently, we conducted a meta‐analysis of RCTs focusing on three specific populations with sepsis (overall sepsis, sepsis‐induced coagulopathy and sepsis‐induced DIC) 6.Umemura Y. Yamakawa K. Ogura H. Yuhara H. Fujimi S. Efficacy and safety of anticoagulant therapy in three specific populations with sepsis: a meta‐analysis of randomized controlled trials.J Thromb Haemost. 2016; 14: 518-30Crossref PubMed Scopus (115) Google Scholar. Although there were no significant reductions in mortality in the overall sepsis population and the population with sepsis‐induced coagulopathy, we observed significant reductions in mortality (risk ratio, 0.72; 95% confidence interval [CI], 0.62–0.85) in the population with sepsis‐induced DIC. We also conducted a multicenter nationwide observational study including 2663 patients with severe sepsis in Japan 7.Yamakawa K. Umemura Y. Hayakawa M. Kudo D. Sanui M. Takahashi H. Yoshikawa Y. Hamasaki T. Fujimi S. Japan septic disseminated intravascular coagulation (J‐Septic DIC) study group. Benefit profile of anticoagulant therapy in sepsis: a nationwide multicentre registry in Japan.Crit Care. 2016; 20: 229Crossref PubMed Scopus (67) Google Scholar. After adjustment for imbalances using propensity score methods, survival effects of anticoagulant therapy were observed only in the patients with DIC (adjusted hazard ratio [HR], 0.609; 95% CI, 0.456–0.814) but not in the patients without DIC. Additionally, a current revisited meta‐analysis by Wiedermann et al. 8.Wiedermann C.J. Antithrombin concentrate use in disseminated intravascular coagulation of sepsis: Meta‐analyses revisited.J Thromb Haemost. 2018; Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar updated the Allingstrup et al. meta‐analysis by replacing all 2314 patients of the KyberSept trial with 229 confirmed DIC patients reported in a post hoc study of the trial 5.Kienast J. Juers M. Wiedermann C.J. Hoffmann J.N. Ostermann H. Strauss R. Keinecke H.O. Warren B.L. Opal S.M. KyberSept investigatorsTreatment effects of high‐dose antithrombin without concomitant heparin in patients with severe sepsis with or without disseminated intravascular coagulation.J Thromb Haemost. 2006; 4: 90-7Crossref PubMed Scopus (309) Google Scholar. Consequently, antithrombin concentration was significantly associated with a lower risk of death in sepsis‐induced DIC in the revisited analysis. These different effects on mortality between overall sepsis and sepsis‐induced DIC can be partly explained by the recent findings on the pathophysiology of sepsis. In certain situations during sepsis, local thrombosis may function as antimicrobial matrices to mediate a physiological host‐protective process against pathogens called 'immunothrombosis' and thus be of major importance as an intrinsic effector of innate immunity 9.Engelmann B. Massberg S. Thrombosis as an intravascular effector of innate immunity.Nat Rev Immunol. 2013; 13: 34-45Crossref PubMed Scopus (1038) Google Scholar. In contrast, the ongoing excessive activation of inflammation might result in uncontrolled activation of thrombosis, which occurs as an initial physiological stage as sepsis‐induced DIC develops. The overwhelmed process of thrombosis results in the development of thrombotic disorders and can no longer assist in host protection. Therefore, anticoagulant therapies might be meaningful in the population with sepsis‐induced DIC, whereas in non‐DIC patients with sepsis, they may cause the contradictory effect of inhibiting host‐defensive thrombosis, which would help to capture and ensnare pathogens circulating in the blood, and thus fail to improve mortality. According to these lines of evidence, we thus believe that DIC is the second key pathology in which anticoagulant therapy could achieve maximum efficacy, in addition to sepsis. However, is the presence of DIC, or the absence of it, enough to select targeted sepsis patients for anticoagulant therapy, or are there more specific populations other than just DIC patients with sepsis who can benefit from the interventions? In our multicenter observational study described above 7.Yamakawa K. Umemura Y. Hayakawa M. Kudo D. Sanui M. Takahashi H. Yoshikawa Y. Hamasaki T. Fujimi S. Japan septic disseminated intravascular coagulation (J‐Septic DIC) study group. Benefit profile of anticoagulant therapy in sepsis: a nationwide multicentre registry in Japan.Crit Care. 2016; 20: 229Crossref PubMed Scopus (67) Google Scholar, survival benefits associated with anticoagulant therapies were also found in patients with high severity of illness (SOFA score, 13–17; adjusted HR, 0.601; 95% CI, 0.451–0.800), suggesting this to be another key condition that can help to determine the optimal target for anticoagulant therapy. Thus, we emphasize that the overall population with sepsis‐induced DIC may not always receive a maximum mortality benefit from anticoagulant therapy. In fact, several RCTs targeting patients with sepsis‐induced DIC found null effects of anticoagulant therapies against death 10.Gando S. Saitoh D. Ishikura H. Ueyama M. Otomo Y. Oda S. Kushimoto S. Tanjoh K. Mayumi T. Ikeda T. Iba T. Eguchi Y. Okamoto K. Ogura H. Koseki K. Sakamoto Y. Takayama Y. Shirai K. Takasu O. Inoue Y. et al.A randomized, controlled, multicenter trial of the effects of antithrombin on disseminated intravascular coagulation in patients with sepsis.Crit Care. 2013; 17: R297Crossref PubMed Scopus (108) Google Scholar, 11.Vincent J.L. Ramesh M.K. Ernest D. LaRosa S.P. Pachl J. Aikawa N. Hoste E. Levy H. Hirman J. Levi M. Daga M. Kutsogiannis D.J. Crowther M. Bernard G.R. Devriendt J. Puigserver J.V. Blanzaco D.U. Esmon C.T. Parrillo J.E. Guzzi L. et al.A randomized, double‐blind, placebo‐controlled, Phase 2b study to evaluate the safety and efficacy of recombinant human soluble thrombomodulin, ART‐123, in patients with sepsis and suspected disseminated intravascular coagulation.Crit Care Med. 2013; 41: 2069-79Crossref PubMed Scopus (220) Google Scholar. In our other multicenter retrospective cohort study, comprised entirely of 162 patients with sepsis‐induced DIC 12.Yoshimura J. Yamakawa K. Ogura H. Umemura Y. Takahashi H. Morikawa M. Inoue Y. Fujimi S. Tanaka H. Hamasaki T. Shimazu T. Benefit profile of recombinant human soluble thrombomodulin in sepsis‐induced disseminated intravascular coagulation: a multicenter propensity score analysis.Crit Care. 2015; 19: 78Crossref PubMed Scopus (50) Google Scholar, we stratified the patients into several strata of disease severity as determined by APACHE II and SOFA scores and evaluated the survival benefit of recombinant human thrombomodulin (rhTM) by propensity score‐adjusted methods. We found that rhTM administration was significantly associated with lower mortality in the high‐severity subset (APACHE II, 24 to 29; adjusted HR, 0.281; 95% CI, 0.093–0.850), whereas there were no effects on mortality in the other subsets. Additionally, in a systematic review and meta‐analysis of rhTM therapy for sepsis‐induced DIC, a meta‐regression analysis revealed a significant negative slope between the effect size of rhTM therapy and the baseline mortality rate in the original studies 13.Yamakawa K. Aihara M. Ogura H. Yuhara H. Hamasaki T. Shimazu T. Recombinant human soluble thrombomodulin in severe sepsis: a systematic review and meta‐analysis.J Thromb Haemost. 2015; 13: 508-19Crossref PubMed Scopus (110) Google Scholar. These findings clearly suggested that the survival effect of anticoagulant therapies might be different even among patients with sepsis‐induced DIC on the basis of disease severity (possibly as indicated by SOFA or APACHE II scores), and it is reasonable to prescribe anticoagulant therapy that is limited to the patients with high severity of illness. Thus, the third key component for selecting patients for anticoagulants might be high disease severity. Currently, there are several unresolved issues surrounding this topic. For example, which criteria should be used for diagnosing DIC? Which agents should be used to treat the optimal target patients discussed in this review? Unfortunately, only little evidence is available to answer these clinical questions. Therefore, new directions for future clinical research are needed. In conclusion, we believe that the patient population fulfilling all three factors of sepsis and DIC and high disease severity may be an optimal target for anticoagulant therapies (Fig. 1). In other words, initiating anticoagulant therapy without the appropriate selection of target patients should be avoided because of the increased risk of bleeding complications with no demonstrable survival benefit. This conclusion, however, is not based on definitive evidence provided by RCTs, and further RCTs of anticoagulant therapy focused on the potentially optimal population to receive this therapy need to be performed. Y. Umemura and K. Yamakawa together conceived the concept of and drafted the manuscript. K. Yamakawa was responsible for drafting, editing, and submission of the manuscript. Both authors read and approved the final manuscript. The authors state that they have no conflict of interest.