Thrombocytopenia and thrombosis in disseminated intravascular coagulation (DIC)

Case 1: On his 50^th birthday, this patient with known advanced hepatitis C was forcibly taken by his best friends to an oyster raw bar to eat raw oysters. Despite his pleas that he hated raw oysters, he swallowed a few. After a few hours of partying, he felt ill, developed chills, fevers, and confusions and was taken to the Emergency Room where staff considered him inebriated and placed him in a quiet, dark room. In fact his ethanol level returned at 180 mg/dL. An hour later he developed large ecchymoses all over his body. His PT was 16.7 seconds; his PTT 68 seconds; platelet count 48,000/μL; pH was 7.08, and analysis for D-dimer tests returned markedly positive. Blood cultures were drawn and broad spectrum antibiotic therapy started. The ecchymoses worsened, epistaxis began, lactic acidosis worsened, and the results of all the prior blood tests deteriorated. What is going on? What can be expected? What can we do for this patient?

Case 2: A 60-year-old man with a prolonged history of esophagitis and now esophageal obstruction was found to have adenocarcinoma of the distal esophagus and uneventfully underwent esophagectomy, relieving obstruction but having positive margins for tumor. A few days later, his right leg began to ache and became swollen; appropriate studies revealed a substantial deep vein thrombosis (DVT) in that leg. He was treated with heparin infusion and warfarin therapy was initiated. Six days later when his INR was 2.7 for the second consecutive day, heparin therapy was stopped and he was discharged. Within hours, the right leg slowly became significantly worse, and now his left arm, at the site of a recent IV line, quickly became painful and swollen. He was readmitted. His platelet count, which had been 300,000/μL preoperatively and had remained at that level at discharge, was 160,000/μL on readmission. What might be going on and what can we offer this man?

Disseminated intravascular coagulation (DIC) is a clinical diagnosis tempered by a few readily available laboratory tests. Supplemental studies can be ordered to solidify the diagnosis or in special research situations. Classic, acute DIC was best clinically described by Mant and King 30 years ago.¹ It typically induces hemorrhagic manifestations as predominated in Case 1. Chronic (and often occult) DIC was landmarked by the 1977 paper by Sack et al.² This type of DIC often is more thrombotic, as exemplified in Case 2. Either clinical situation remains extremely dangerous, both from the seriousness of the specific underlying disorder that initiates and promulgates DIC, but also from the hemorrhagic and/or thrombotic tendencies that are characteristic of all DIC. The historical approach to our understanding of DIC has been reviewed.³ 

DIC does not occur in isolation, as it is essentially always the result of some underlying problem, which typically is rather obvious (Table 1 ). When one reviews the clinical conditions that are associated with DIC, one sees a veritable litany of very ill patients in nearly every category of inhospital medical care. Stepping back from this list, one can deduce that what is in common with these heterogeneous etiologies is cell injury and/or death such that coagulation is initiated in vivo in physiological response to the production of either thromboplastic agents directly (in cases in which tissue has been damaged to the point that its thromboplastic contents are released) or indirectly when monocytes and/or endothelial cells are physiologically stimulated to produce and secrete cytokines in response to injury thus initiating the extrinsic pathway of coagulation. Monocytes and endothelial cells that do not constitutively elaborate tissue factor (TF) do so when stimulated by pathogens that activate endothelial cells into elaboration and secretion of TF and von Willebrand factor (VWF). What separates this physiologic hemostasis from pathologic DIC is whether the prothrombotic substances released by the product of magnitude and duration of the stimulus outstrips the clearance of these prothrombotic substances by the reticuloendothelial system, physiologic inhibitors of coagulation, and the fibrinolytic system.

Should the coagulation system be stimulated to the point that thrombin circulates free from all physiologic inhibitors, circulating thrombin is then free to convert fibrinogen into fibrin systemically and to activate platelets, as well as activate factors V, VIII, and XIII, thus completing coagulation. Yet thrombin, via the thrombin:thrombomodulin (TM) complex, activates protein C, which then further activates the fibrinolytic system while inhibiting factors Va and VIIIa. The fibrinolytic system is also activated via endothelial release of tPA, which in turn activates plasmin. Plasmin, once free of its inhibitors, is then free to degrade systemically factors V, VIII, XIII, as well as any nascent fibrin clots and even fibrinogen itself, generating the characteristic fibrin degradation products (FDPs). An appreciation of these multiple pathways, often acting at cross purposes to one another, is what allows one to fathom the laboratory and clinical perturbations of DIC.

This pathologic perturbation of a physiologic process produces thrombosis (“an unwanted clot at an unwanted time in an unwanted place”), particularly in the microcirculation. This accounts for the inappropriate deposition of fibrin and platelets in the microcirculation, with subsequent organ ischemia resulting in the characteristic hypofibrinogenemia and thrombocytopenia seen in the circulating blood. Accordingly, the term “consumptive coagulopathy” is an accurate alternative name for DIC.

There is a spectrum of thrombosis and hemostasis in all cases of DIC; by definition, each or both may be encountered. One of the more common causes of acute DIC is sepsis. This form of DIC is characterized by a trend toward more bleeding than thrombosis, particularly as observed at the bedside. The characteristic thrombosis that one might see in septic-generated DIC is purpura fulminans (PF). An example of an occult thrombosis is adrenal vein thrombosis with its inevitable subsequent adrenal hemorrhage, producing the so-called Waterhouse-Friderichsen syndrome. One of the more common chronic causes of chronic DIC is tumor-initiated DIC. This hypercoagulability of cancer in its pluperfect form is called Trousseau syndrome (Table 2 ). Despite their abbreviated half-lives, circulating levels of fibrinogen and platelets are often normal, if not even slightly increased, in this slower type of DIC. Therefore, hemorrhage is not as characteristic of cancer-initiated DIC as is thrombosis with overt manifestations as DVT and pulmonary embolism (PE) as well as thrombosis in the central nervous system CNS and the abdominal organs.

DIC is frequent in trauma patients, particularly those having impaired cardiac output to include shock and its accomplices, acidosis and hypothermia. That these are synergistic in promulgation of this coagulopathy has been repeatedly shown in the trauma literature.⁴ There is a progression from dilutional coagulopathy during the resuscitation of trauma patients to the eventual development of frank DIC unless and until hypothermia and acidosis can be corrected by prompt resuscitation of the patient. Hemostasis works neither at low pH nor with hypothermia.

Mant and King¹ demonstrated plasma from the vast majority of patients suffering from acute classic DIC had abnormalities in prothrombin times (PT) and partial thromboplastin times (PTT) as well as thrombocytopenia and decreased fibrinogen levels. Nearly all these patients also had significant elevations of FDPs. Ironically, a shortened PTT has been observed early on or even prior to DIC in some cases.^5,6 The International Society of Thrombosis and Haemostasis (ISTH) Subcommittee on DIC promulgated a scoring system revolving around these readily available tests⁷ (Table 3 ) and validated this simplified approach,⁸ thereby negating the need for more complex DIC batteries. Even this ISTH-DIC laboratory test protocol is not without some misgivings, particularly if deployed in incorrect clinical situations (such as ambulatory patients with neither bleeding nor thrombosis) and mimics of DIC (such as dilutional coagulopathy). Three patients in these three different pathophysiologic conditions may present with identical laboratory parameters, as outlined in Table 3 . Although all three patients have similar laboratory abnormalities, more information is garnered by observing the patient and understanding the pathophysiology, which allows the clinician to distinguish Case A (chronic, stable, ambulatory severe hepatic insufficiency) from Case B (dilutional coagulopathy), and from true DIC as represented in Case C. Clearly, bedside information is extremely important, not only in the correct diagnosis but in prognostication and, more importantly, what treatment opportunities, if any, are revealed by determining the correct underlying process.

Many have attempted to augment these simple, readily available tests in efforts to not only further define DIC but to generate subclassifications and prognostic formulations. That these efforts have actually added to our understanding, let alone altered our treatment or prognosis, has not been proved. In classic acute DIC one would expect and often will find perturbations in concentrations of plasma anti-thrombin III, plasminogen, and alpha-2-plasmin inhibitor. Recently it has been demonstrated that the von Willebrand factor cleaving protease (ADAMTS-13) is markedly decreased in sepsis-related DIC and portends renal failure.⁹ These tests are more expensive, less readily available, and often return too late to significantly alter the approach in this dynamic and rapidly evolving illness.

Accordingly, one can now produce a short focused laboratory approach supporting the clinical diagnosis of DIC (Table 4 ). These tests serve to bolster one’s high pretest probability in a patient with unexplained bleeding, thrombosis, or often both. This set of laboratory tests is recommended.

Recognizing the strong tendency toward having objective, reproducible and numerical quantification systems in clinical studies, the ISTH-DIC scoring system is frequently employed. This is a legitimate effort. Some have divided DIC into “non-overt” and “overt” DIC. The question is whether this attempt to standardize definitions is of sufficient utility to overcome the vast heterogeneity among all patients who manifest the final common pathway we call DIC. For instance, one might theorize three different patients, all of whom who may be septic from pneumococcus. They may each have DIC and may even have similar ISTH-DIC scores but clearly are all associated with different prognoses either with or without treatment; even the treatment of each patient differs. An example might be a young college student, who, following what was thought to be a viral pneumonia, actually developed pneumococcal pneumonia and now has a parapneumonic empyema that on occasion leaks, causing brisk DIC. A second case, also characterized as pneumococcal sepsis, would be a middle-aged man who underwent splenectomy for staging of Hodgkin disease three decades ago. He has been well, but now suddenly develops chills, fevers, and prostration from post-splenectomy fulminant pneumococcal sepsis. The third case could be a 70-year old man with chronic obstructive pulmonary disease and multiple myeloma who now has pneumococcal sepsis. The notion that a numerical score derived from such a heterogenous group of patients could serve all needs at all times is somewhat stretched. Admittedly, it would be very difficult to accrue enough similar patients with each precise type of DIC to generate a more understandable system currently.

The cause of death in DIC patients according to most clinical reviews is more of a function of the underlying cause, particularly if that cause is resistant to therapeutic measures. Pathologic bleeding and pathologic thrombosis conspire against the patient. Multiorgan dysfunction syndrome (MODS) is a frequent consequence of DIC and is usually due to bleeding into organs or thrombotic alterations in various organs to include the hepatic, cardiac, central nervous, renal, and pulmonary systems.

A basic tenet involving one’s approach to DIC is that, if an underlying condition is responsible for triggering, promulgating, and continuing the pathologic free circulation of both plasmin and thrombin, one’s treatment is directed at reversing that cause. All patients should be resuscitated aggressively to optimize maintenance of fluid status, blood pressure, temperature, and pH. Those patients whose DIC triggers are most readily approached (typified by the obstetrical catastrophes) do very well, whereas those with chronic irreversible underlying diseases that are more difficult to resolve (such as sepsis in a granulocytopenic cancer patient with liver disease) are more difficult to treat.

Since the very beginnings of our unraveling of DIC, there has been vigorous discussions of whether replenishing a patient’s plasma coagulation factors and platelets do more damage than not. On the one hand, one wishes to keep enough procoagulant factors and platelets to afford hemostasis, yet any “extra” has been accused of “fanning the fire.” Most suggest that fibrinogen levels should be kept above 50 mg/dL yet less than 100 mg/dL, whereas platelet counts might be maintained in the area of 50,000/μL, as these are held to be minimally effective levels. The source for fibrinogen usually employed is cryoprecipitate (on the order of 10 “units” of cryoprecipitate with each unit containing about 200 mg of fibrinogen). Fresh frozen plasma (FFP) is rarely recommended but frequently used (Table 5 ).

At first thought, if antithrombin III is the primary inhibitor of circulating thrombin and circulating thrombin is a defining component of acute DIC, it would seem intuitive that antithrombin III replacement would have efficacy, but such has not proved to be the case in human trials. Infusion of preparations of activated protein C (APC) has shown some success, particularly in decreasing mortality with severe sepsis (Table 6 ).

The use of heparin in DIC remains controversial. In many studies, to include Mant and King’s early study,¹ death from hemorrhage appeared to be increased by the aggressive use of heparin. Antifibrinolytic agents, such as EACA, may be advocated and are usually employed only late in cases characterized by extreme ongoing hemorrhage. Because of the ever-present circulation of thrombin, thrombin inhibition by the prerequisite infusion of heparin is strongly advised before the infusion of EACA in order to minimize a thrombotic catastrophe.

On one hand that the ISTH-DIC score is really measuring something of value is supported by the data from the PROWESS study of use of APC versus placebo in sepsis-related DIC,¹⁰ which showed a very clear and present rise in mortality as a function of ISTH-DIC score. On the other hand, studies such as Aoki et al,¹¹ when investigating the efficacy of human APC in patients with DIC of all types, showed that the infusion of APC resulted in statistically significant improvement in selected DIC battery scores compared with heparin; however, there was no significant difference (P = .3) in the rate of complete recovery of DIC within those two groups.

The PROWESS investigators also noted that the ISTH-DIC scoring system consciously or unconsciously is used more to evaluate for hemorrhagic risk than thrombotic risk in DIC.¹⁰ At times there was a correlation with clinical hemorrhage, while at other times there was not. They noted that there are fewer laboratory means to quantify thrombotic risk, and at present thrombosis is determined almost exclusively by clinical findings that lack both sensitivity and specificity. Therefore, it is possible that any of the investigative metrics used to study a treatment for DIC has little chance of proving a causative change in hemorrhagic risk and even less chance that the studied treatment has an effect on thrombotic risk.

If DIC represents both hemorrhagic and thrombotic risks, what is the clinician to do? First, of course, is to identify the underlying cause and approach that as aggressively as possible. Second is to estimate the relative evidence favoring hemorrhage and favoring thrombosis using one’s best clinical judgment (some suggestions are listed in Table 7 ). Third, consider supportive therapy with blood and blood products (some suggestions are listed in Table 5 ). The status of some potentially available remedies in DIC management, none of which are to be viewed as standard of care as yet, are discussed in Table 6 .

For Case 1, the clinical diagnosis (later confirmed by blood and stool cultures) was that the patient had acquired Vibrio vulnificus septicemia from ingestion of raw seafood. Antibiotic therapy was initiated. He rapidly developed purpura fulminans and, because this skin necrosis was rapidly advancing, it was elected to administer heparin by infusion at a rate to give therapeutic plasma levels of 0.4 to 0.6 u/mL but without published evidence this therapy would prove effective. He worsened as MODS progressed and he expired from cardiovascular collapse. Autopsy showed microvascular changes consistent with DIC, advanced hepatic cirrhosis, and adrenal infarction/hemorrhage.

For Case 2, it was clear that despite (or because of?) heparin prophylaxis this patient was rapidly developing new DVTs and “relapsed” on warfarin therapy. Our working diagnosis was Trousseau syndrome. Because it was possible (low probability) that he had HIT, he was initially treated with fondaparinux 7.5 mg SQ until his screening ELISA test for HIT returned negative. Fondaparinux was continued for ease of outpatient management. His thromboses rapidly responded and he has remained on that fondaparinux dose for the past year without clear evidence of tumor recurrence.