Disseminated intravascular coagulation is caused by multiple causes, with different pathophysiological mechanisms and diverse clinical manifestations. Severe infection, trauma, pathological obstetrics, malignant tumor and poisoning, etc., can all lead to coagulation dysfunction and gradually progress to DIC (11). Therefore, DIC is the result of a deterioration in coagulation function. In the process of coagulation activation caused by severe infection, endothelial cells change from anticoagulant surface to pro-coagulant surface, the production and release of coagulation factors increase, and platelets are activated, resulting in hypercoagulable state and microvascular thrombosis, which further affects tissue perfusion and leads to organ dysfunction. Therefore, it is thrombotic DIC (12, 13). However, DIC caused by trauma is due to loss and dilution of hemostatic substrates such as coagulation factors and platelets, as well as the activation of fibrinolysis, so the clinical manifestation is bleeding type (13, 14). Therefore, DIC, similar to ARDS and sepsis, is not a disease, but a clinical syndrome. The clinical manifestations of DIC caused by different etiologies vary greatly, so the treatments are also different. Therefore, it is necessary to establish new diagnostic criteria according to the pathophysiological changes of DIC.

The clinical manifestations of DIC, including hypercoagulation and hypocoagulation, thrombus and bleeding, are not clearly defined and often coexist (11). For example, in the lipopolysaccharide (LPS)-induced rat model of sepsis, bleeding was observed at 10 min after LPS injection. The bleeding area expanded at 30 min. Thrombus was formed in the microvessel at 60 min, and thrombosis aggravated at 120 min, resulting in the cessation of blood flow in the distal side of the vessel, indicating that thrombosis and bleeding coexisted (15). The same is true in the rat model of trauma, the microthrombus was seen in the mesenteric venule at 30 min. The thrombus dissolved at 180 min. However, extravasation of red and white blood cells (micro bleeding) increased (3). The causes of DIC determines the clinical manifestations, some are mainly bleeding, some are mainly thrombosis, but in most cases both exist simultaneously. However, there is no direct evidence of microthrombosis, which is easy to be ignored clinically (16).

Hypercoagulation and hypocoagulation, coagulation and fibrinolysis, compensation and decompensation evolve dynamically in DIC. In trauma-induced DIC, thrombin production is increased due to post-traumatic tissue damage and bleeding, leading to hypercoagulation. With the progression of disease and the increase of blood loss, the endothelial cell damage worsens and it changes to hypocoagulation. With the gradual cessation of bleeding, combined with the supplementation of clotting substrates, and the inhibition of fibrinolysis by the production of thrombin-induced plasminogen activator inhibitor-1 (PAI-1), which leads to the conversion to hypercoagulation. It reflects the evolution between hypercoagulation and hypocoagulation, and between coagulation and fibrinolysis (17). In sepsis-induced DIC, the invasion of pathogen is first recognized by the innate immune system, leading to activation of endothelial cells and immune cells to form local clots that prevent the spread of the pathogen, known as “immunothrombosis” (18, 19). At this point, the body’s coagulation function is compensatory. As the disease progresses, coagulation activation is further aggravated, leading to widespread thrombosis in the microcirculation, tissue ischemia and hypoxia, and then organ dysfunction. This is a decompensated state (12, 20, 21).

The central pathophysiological changes of DIC caused by various etiologies are endothelial cells injury and systemic coagulation activation, which reflects the characteristics of “disseminated” (22). This is what distinguishes it from local thrombotic disease.

Endothelial cells line the luminal surface of all blood vessels, and are therefore the first barrier that separates blood and tissue. Endothelial cell injury promotes coagulation activation and DIC. Loss of homeostasis of the coagulation system leads to microthrombosis, tissue hypoperfusion and organ dysfunction. Thus, DIC is considered to be an initiating factor of MODS rather than one of the organs (23).

The current diagnostic criteria rely solely on clinical features and laboratory test results, and cannot reflect pathophysiological changes, especially endothelial cells damage (33). The differential diagnosis of acute DIC is very important, and it may be difficult to distinguish it from other diseases by clinical and laboratory parameters alone. For example, heparin-induced thrombocytopenia (HIT) may cause thrombosis and thrombocytopenia. Thrombotic thrombocytopenic purpura (TTP) leads to thrombosis, organ dysfunction and thrombocytopenia. Antiphospholipid syndrome (APS) can result in thrombosis, thrombocytopenia and prolonged activated partial thromboplastin time (APTT). Severe form of APS may lead to rapid onset of multiple organ dysfunction. Liver cirrhosis can cause bleeding tendency, prolonged PT, prolonged APTT and thrombocytopenia. All of the above may meet the current DIC diagnostic criteria, but they are not DIC and need to be differentiated (34, 35). Early local bleeding in patients with aortic aneurysm leads to prolonged PT, increased D-dimer (D-D), and thrombocytopenia, but coagulation activation is localized without systemic endothelial cell damage. It mimics DIC, but is not DIC (3). Therefore, the current DIC diagnostic criteria that rely solely on clinical manifestations and laboratory findings are not appropriate.

Clinical features, as parameters of diagnostic criteria, cannot fully reflect the actual situation. The signs of fulminant purpura or hemorrhagic embolism, or hemodynamic instability resulting from large blood vessel embolism may be evaluated clinically. However, the microcirculatory embolism is not easy to detect clinically and may delay diagnosis, especially in sepsis-induced DIC.

Diagnosis is for proper treatment, so the diagnostic criteria need to be highly sensitive and specific. At present, the diagnostic criteria rely on laboratory test results, which are not sensitive. Many studies have explored biomarkers for early prediction of coagulation dysfunction, most of which are related to endothelial cell damage (8, 10, 36). However, multiple systems such as coagulation and inflammation interact, so it is key to include biomarkers that reflect multi-system changes in DIC diagnostic criteria (37). Currently, most biomarkers are not widely tested clinically, limiting their inclusion in the diagnostic criteria.

Disseminated intravascular coagulation caused by different etiologies has different pathophysiology, so different diagnostic criteria should be adopted. Sepsis, acute respiratory distress syndrome (ARDS) and DIC are all clinical syndromes with great heterogeneity and need to be phenotyped in order to adopt similar treatments for patients with the same pathophysiology (38, 39). At present, there have been some attempts to classify DIC, such as “coagulation phenotype” and “fibrinolytic phenotype” according to coagulation or fibrinolytic changes. The former is represented by sepsis, which is mainly characterized by thrombosis, tissue hypoperfusion and organ dysfunction. Coagulation activation and suppressed fibrinolysis are common features in sepsis. PAI-1 is significantly increased, and fibrinolytic degradation products such as D-D and FDP are slightly elevated. Anticoagulation is the main treatment for this type. DIC with enhanced fibrinolysis can be seen in trauma, and the main clinical manifestations are hemorrhage. It is characterized by hyperfibrinolysis, which manifests as a significant increase in D-D and FDP, and almost no increase in PAI-1. Antifibrinolytic therapy is required in this type. DIC is also divided into “acute” and “chronic” according to the course. There are differences in etiology, course of disease, clinical manifestations, laboratory tests, treatments and prognosis (16, 40). Diseases that cause acute DIC include infection, trauma, pathological obstetrics, etc. Onset is usually within 7 days. The clinical manifestations are mild in the early stage and gradually progress in the middle and late stage. Microcirculation disorders and organ dysfunction are common, and most coagulation tests suggest a decompensated state. The main treatment is to control the etiology and improve coagulation function. The prognosis is poor. Chronic DIC can be seen in malignant tumors, pregnancy process, etc. The course of the disease is usually more than 14 days, and there is no microcirculation disturbance and organ failure. Laboratory tests often reveal a compensatory state. Combined treatment with anticoagulation and antifibrinolysis is effective, and the prognosis is good (3). However, the current classification is not based on pathophysiology, and the personalized diagnostic criteria for DIC based on etiologies are more suitable and helpful.

The parameters in the current diagnostic criteria cannot fully reflect the actual coagulation function. At present, most of the DIC diagnostic criteria include coagulation and fibrinolysis parameters and platelet counts, but they cannot represent the overall picture of actual coagulation function. For example, platelet counts are inconsistent with function, and evaluating platelet counts alone is not comprehensive. In addition, platelet count in severe patients is affected by a variety of factors, such as bone marrow suppression, destruction of extracorporeal circulation and immune response, etc., which can lead to thrombocytopenia, but do not represent abnormal coagulation function (41, 42). Fibrinolytic parameters such as D-D and FDP included in the diagnostic criteria cannot fully represent the severity of coagulation dysfunction. For example, fibrinolysis is suppressed in sepsis-induced DIC, so the levels of D-D and FDP underestimate the severity of coagulation activation (43, 44). Fibrinogen (Fib), as an acute reaction protein, is mostly elevated in sepsis and does not reflect changes in coagulation and fibrinolysis function (12). Currently, ISTH is the most commonly used DIC diagnostic criteria in sepsis, which is obviously inappropriate due to the inclusion of Fib.

Personalized DIC diagnostic criteria needs to be re-established according to different causes. No “one-size-fits-all criteria” for DIC!

With the in-depth study on the pathophysiology of DIC, it is gradually realized that TF-mediated activation of extrinsic coagulation pathway and the subsequent coagulation dysfunction are the main factors contributing to the progression of DIC (45). In the process of DIC, in addition to coagulation activation, the anticoagulation pathway is also damaged, which aggravates coagulation dysfunction. Interactions between platelets, clotting factors, endothelial cells and neutrophils amplify the coagulation cascade. Therefore, cell-mediated activation of the coagulation system is the basis for initiation, expansion and spread of a series of coagulation cascades (46). Bidirectional regulation of inflammation and coagulation system promotes the onset of DIC (47). Thrombin-antithrombin (TAT) and PAI-1, as key molecules in the pathophysiology of DIC, are altered in the early stage of DIC and may be used as biomarkers for the diagnosis of DIC (33).

With the deepening understanding of clinical syndromes such as sepsis and DIC, it has been recognized that the failure of clinical studies on sepsis is probably due to the high heterogeneity of patients, and therefore it is necessary to identify the phenotypes (48, 49). With the development and application of artificial intelligence, Seymour et al. (50) divided sepsis patients into α, β, γ, and δ phenotypes by machine learning method as early as 2019, among which δ phenotype showed the most obvious coagulation changes, mainly manifested by increased TAT, D-D and PAI-1. The mortality rate was also highest in the δ phenotype. In 2021, Kudo et al. (51) divided sepsis patients into dA, dB, dC, and dD phenotypes by machine learning method. Cluster dA had the most severe coagulopathy with high D-D and FDP levels, the most severe organ dysfunction, and the highest mortality. The results of the two studies are similar, suggesting that patients with severe coagulation activation have more organ dysfunction and higher mortality, which is consistent with the previous understanding that “DIC is an initiating factor of MODS.”

Disseminated intravascular coagulation is a dynamic process, which requires the diagnostic criteria for the early stage of DIC. In 2019, ISTH proposed that the process from SIC to DIC develops gradually, and DIC caused by sepsis needs to adopt a “two-step” diagnosis scheme (52), that is, sepsis patients with thrombocytopenia should be first screened using SIC scoring system, and early anticoagulant treatments should be carried out for patients meeting the SIC criteria. The overt-ISTH DIC score was evaluated as the second step. DIC patients diagnosed by overt-ISTH criteria should be treated with bundle therapy including anticoagulants. A differential diagnosis should be performed in sepsis patients who do not meet SIC or overt-ISTH criteria. The “two-step” approach represents a continuum of SIC and DIC, where SIC criteria is highly sensitive and more suitable for early screening, followed by early intervention to prevent progression.

In 2018, Chang (53) proposed that DIC should be reinterpreted according to “two-activation theory of the endothelium.” Infection, trauma, pathological obstetrics and other etiologies first activate the complement system and promote the generation of C5b-9 as a terminal membrane attack complex, resulting in endothelial injury and activation, and thus causing endotheliopathy. Subsequently, two important molecular events occur: activation of inflammatory pathway and microthrombotic pathway. The former triggers the release of cytokines, causing “inflammatory storm.” The latter mediates platelet activation and endothelial exocytosis of ultra-large von Willebrand factor (ULVWF), which is anchored to endothelial surface and recruits activated platelets to form microthrombus composed of platelet-ULVWF complexes. This process leads to thrombocytopenia and disseminated intravascular microthrombosis (DIT), resulting in MODS. Therefore, it is suggested that “DIC” should be redefined as “endotheliopathy-associated DIT.” This is a new understanding of the pathophysiological mechanism, but it still needs further verification.

In order to establish reasonable DIC diagnostic criteria, the most important thing is to fully understand its pathophysiological mechanisms, such as the damaging effects and mechanisms of damage-associated molecular patterns (DAMPs) including histones (54), and the interaction between the pro-coagulant substances released after endothelial dysfunction, which remain to be further explored. It seems that it is imperative to redefine DIC based on different etiologies and pathophysiological mechanisms, such as “sepsis-induced coagulopathy,” “trauma-induced coagulopathy,” “cancer-induced coagulopathy,” “heatstroke-induced coagulopathy,” “pregnancy-related coagulopathy,” et al. Of course, personalized diagnostic criteria based on pathophysiology are also needed (55). The new DIC diagnostic criteria should have the following characteristics: sequential diagnostic criteria for early detection of DIC, the parameters included in the diagnostic criteria can reflect the understanding of pathophysiology, clear cut-off values of the parameters, clinical applicability, global unified (currently Japan and western countries tend to apply different diagnostic criteria) and be able to identify therapeutic patients. Since DIC is a complex syndrome caused by multiple etiologies, the combination of artificial intelligence stratification and the inclusion of sensitive biomarkers will help redefine DIC and identify phenotypes.

Although the diagnosis of DIC is important, it is even more important to provide early warning of its onset and stratify DIC patients so that early intervention can be carried out in high-risk groups to prevent its progression. With the development of artificial intelligence, it may be the future direction to collect clinical information and biological samples, combine with bioinformatics to screen relevant biomarkers and identify the correlation with clinical information, so as to warn the onset of DIC or incorporate into diagnostic criteria, and then screen targeted patients for DIC anticoagulant therapy.