It is well known that patients with chronic liver disease often have an abnormal hemostasis test and bleed frequently from the gastrointestinal tract, therefore, it has become only too natural and consequential to link these two abnormalities and to imply that one is the cause of the other [1, 2]. The astute hepatologist is well aware that the anatomic and hemodynamic consequences of portal hypertension are the main determinants of gastrointestinal bleeding in severe liver disease . Yet, the impairment of hemostasis has long been considered an important cofactor. This is witnessed by the abundant literature on the value of hemostasis tests in the prediction of bleeding and on the clinical evaluation of hemostatic drugs as adjuvants in the prevention and control of bleeding. In this debate, I shall attempt to demonstrate how the recent adoption of hemostasis tests more capable than those used so far to reflect in vitro what is happening in vivo challenges the dogma that hemostasis is abnormal in patients with chronic liver disease. This reassessment of the problem has implications for treatment, particularly on whether or not it is warranted to look for hemostatic drugs as adjuvants in the treatment of portal hypertension and its hemorrhagic complications.
The past as a prologue
In each hematology or hemostasiology textbook, chronic liver disease is listed as an important cause of acquired bleeding. Chronic liver disease is indeed a cause of abnormal hemostasis tests, spanning from low count and abnormal function of blood platelets (often reflected by a prolonged skin bleeding time) to low plasma levels of coagulation factors (reflected by such prolonged coagulation screening tests as the prothrombin time and the activated partial thromboplastin time) (Table 1). There is also evidence that the fibrinolytic activity is heightened in chronic liver disease, with the implication that this system contributes to the derangement of hemostasis through the increased tendency of formed clots to lyse. Yet, the evidence that abnormalities of hemostasis tests are associated with an increased bleeding tendency is meager. Two small but innovative studies by Ewe  and Dillon et al.  do indeed observe that an array of hemostasis tests performed in peripheral blood correlated poorly with the actual duration of bleeding and the amount of blood loss measured directly at laparoscopy from the biopsy puncture. The authors of these articles concluded that abnormal bleeding after liver biopsy is a random event that cannot be predicted by the laboratory methods currently used to explore the hemostatic system. Other studies indicating little or no association between the risk of bleeding after liver biopsy and the degree of abnormal hemostasis tests are those of McVay and Toy  and Caturelli et al. . The only significant exception comes from Boberg et al. , who showed that, among 219 patients undergoing percutaneous liver biopsy, a prolonged skin bleeding time gave a 5-fold greater risk of significant post-biopsy bleeding, defined by a decrease in hemoglobin of 2 g dL−1 or more. In a study of patients with ‘decompensated’ liver disease Boks et al.  concluded that variceal bleeding was not related to the impairment of an array of coagulation and fibrinolysis tests . Additional evidence against the clinical relevance of abnormal hemostasis tests in chronic liver disease stems from the natural history of orthotopic liver transplantation. In the past this operation was known to be accompanied by massive intraoperative bleeding that required a huge use of transfusional blood products . During liver transplantation further changes in the already disturbed hemostatic system do indeed occur , but it is unlikely that these abnormalities play a key role in the determination of excessive bleeding, because the major improvements that have taken place in the last few years in surgical techniques have dramatically reduced blood losses and transfusion requirements, in spite of no significant change in the adoption of medical interventions and in particular of hemostatic drugs. Abnormal bleeding is no longer a frequent concern for surgeons and anesthesiologists, whereas postoperative thromboembolism has become a concern.
|Multiple coagulation factor deficiencies|
|Thrombocytopenia and thrombocytopathy|
|Hyperfibrinolysis (high tissue plasminogen activator and low thrombin activatable fibrinolysis inhibitor)|
|Deficiency of naturally occurring anticoagulants (antithrombin, thrombomodulin/protein C/protein S system, tissue factor pathway inhibitor)|
|Deficiency of profibrinolytic factors (plasminogen) and increase of the principal inhibitor of plasminogen activation (PAI-1)|
|High levels of the principal adhesive protein (von Willebrand factor)|
On the whole, the forementioned data indicate that hemostasis tests are not important predictors of the tendency of patients with chronic liver disease to bleed from esophageal varices or even after surgical procedures (including biopsies and transplantation). Why then there is such a solid belief that chronic liver disease is an important cause of abnormal hemostasis? The time honored dogma that hemostasis is abnormal in liver disease was based upon laboratory methods such as the prothrombin time and the activated partial thromboplastin time (and other tests designed to evaluate the capacity of liver to synthesize coagulation factors, i.e. the Normotest, HepatoQuick and Thrombotest). These tests only explore the early phase of the formation of thrombin, the key enzyme of the coagulation system. Thrombin formation is a dynamic process, in which the forming coagulation enzyme is continuously neutralized by such naturally occurring anticoagulant proteins as antithrombin, thrombomodulin, activated protein C and tissue factor pathway inhibitor. Standard coagulation tests do not measure global thrombin as it is formed and then neutralized, but only the small amounts that are needed to form a first visible clot. In patients with chronic liver disease the latter process is slowed by the presence of low plasma levels of coagulation factors, and hence standard coagulation tests are abnormally prolonged. Tripodi et al.  have recently showed that when blood coagulability is globally measured using a thrombin generation assay, modified by the addition of thrombomodulin and hence sensitive not only to the low plasma levels of coagulation factors but also to the reduced levels of naturally occurring inhibitors , patients with liver cirrhosis and abnormal standard coagulation tests form thrombin in amounts similar to those of healthy individuals taken as controls.
Another challenge to the concept of an abnormal hemostatic system in patients with chronic liver disease stems from the study of Lisman et al.  on fibrinolysis. In chronic liver disease, some of the components of the fibrinolytic system are altered in the direction of hyperfibrinolysis (high plasma levels of tissue plasminogen activator and low levels of α2-plasmin inhibitor), but others are altered in the direction of hypofibrinolysis (low plasminogen, high plasminogen activation inhibitor type 1), so that a balance does ultimately exist also for this hemostatic system (Table 1). Lisman et al.  have shown that the thrombin-activatable fibrinolysis inhibitor (TAFI) is reduced in patients with severe liver disease to a degree proportional to the severity of the disease. The decrease of this principal fibrinolysis inhibitor should in principle cause an excess of fibrinolysis. However, when fibrinolysis was explored with a global test sensitive to both activators and inhibitors, the results obtained in cirrhotics were not different from those obtained in healthy controls . Most importantly, an elegant in vitro experiment included in Lisman's study showed convincingly that the interplay of decreased activators and inhibitors eventually leads to normal fibrinolysis in severe liver disease . By artificially reducing the plasma levels of TAFI to half-normal, lysis of the clot was more accelerated than in a sample containing normal levels of the inhibitor. On the other hand, plasma with half-normal levels of antithrombin had a lysis time slower than normal, but plasma containing at the same time half-normal levels of both TAFI and antithrombin had a lysis time that was identical to that of normal plasma, indicating that the two decreases are eventually balanced .
Does an abnormal primary hemostasis because of thrombocytopenia and thrombocytopathy still stand as a cause of abnormal bleeding in liver disease, as purported by Boberg et al.? Hardly so. Using a method based on the measurement under flow conditions of the adhesion to the subendothelium of platelets from reconstituted blood with standardized platelet and red blood cell counts, Bongers et al.  showed that in liver cirrhosis high plasma levels of the principal adhesive protein von Willebrand factor compensate for any reduction of platelet count or any defect of platelet function.
In the past, and even in more recent years, there have been attempts to evaluate the role of hemostatic agents in the management of the most frequent problem of patients with severe liver disease, that is, bleeding form esophageal varices. Several types of hemostatic agents have been evaluated, spanning from antifibrinolytic amino acids such as ε-aminocaproic acid or tranexamic acid to, more recently, potent procoagulant agents such as recombinant activated factor VII . The results of these trials have been in general negative or inconclusive , and there are perhaps two possible reasons for this. The first is that the weapons currently used in the prevention and treatment of variceal bleeding (β-blockers, vasoactive agents, banding and sclerotherapy) are so effective that little space is left for adjuvant hemostatic agents, that obviously must be evaluated on top of these effective treatments. Accordingly, very large series of patients must be included in clinical trials to look for a modest improvement. Second, for all the reasons mentioned above that cast serious doubts on the relevance of abnormal hemostasis in the causation of excessive bleeding in liver disease, it is at least uncertain that the evaluation of these hemostatic drugs, some of which are outrageously expensive, has a solid rationale.