Coagulopathy in liver disease

What every physician needs to know about coagulopathy in liver disease

The liver plays a key role in blood coagulation, being involved in both primary and secondary hemostasis. It is the site of synthesis of all coagulation factors and their inhibitors, except for von Willebrand factor (vWF). Liver damage is commonly associated with impairment of coagulation. The hemostatic system is in a delicate balance between prothrombotic and antithrombotic processes, aiming to prevent excessive blood loss from injured vessels and to prevent spontaneous thrombosis. Liver failure is accompanied by multiple changes in the haemostatic system, because of reduced plasma levels of procoagulant and anticoagulant factors synthesised by the intact liver. Vitamin K deficiency may coexist, resulting in defective carboxylation of clotting factors and inhibitors.

During liver failure, there is a reduced capacity to clear activated haemostatic proteins and protein inhibitor complexes from the circulation. Portal hypertension with collateral circulation and secondary splenomegaly causes thrombocytopenia due to splenic sequestration. Thrombocytopenia may also be caused by decreased hepatic thrombopoietin synthesis. There may also be impaired platelet function. Complications of cirrhosis such as variceal bleeding or infection/sepsis may lead to the onset of bleeding.

A baseline consumptive coagulopathy, other than that secondary to sepsis or other predisposing causes, is disputed. There is evidence for either normal or increased thrombin formation in patients with liver cirrhosis.

Continue Reading

Therapy for abnormal coagulation in liver disease is needed only during bleeding episodes or before invasive procedures.

When end stage liver disease occurs, liver transplantation is the only treatment available. This can restore normal haemostasis, and can correct genetic defects such as hemophilia or factor V Leiden mutation. During liver transplantation, hemorrhage may occur due to the preexisting hypocoagulable state, from the collateral circulation (caused by portal hypertension), or because of the increased fibrinolysis which occurs during this surgery. (Table I)

Table 1.n

What features of the presentation will guide me toward possible causes and next treatment steps:

Clinical presentation of coagulation disturbances in patients with liver cirrhosis

Patients with advanced liver disease rarely develop spontaneous bleeding. Thrombotic complications can, paradoxically, occur in cirrhotic patients clinically regarded as having a hemorrhagic tendency. Despite abnormally prolonged coagulation tests, these patients cannot be considered “anti-coagulated”. Portal vein thrombosis complicates liver cirrhosis in 0.6 to 15% of cases, leading to worsening of liver function and sometimes to mesenteric vein obstruction/infarction. Anticoagulation is indicated to prevent extension of the thrombus, and allow the splanchnic veins to recanalize (in about 50% of patients).

Patients with liver cirrhosis are at greater risk of developing deep vein thrombosis. Patients with cholestatic disease, such as primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), often exhibit a procoagulant state demonstrated by thromboelastography (TEG), and may be prone to thrombosis, but this is not documented.

What laboratory studies should you order to help make the diagnosis and how should you interpret the results?

The PT (prothrombin time) and INR (international normalized ratio) are poorly correlated with the risk of bleeding during interventional procedures in patients with liver cirrhosis. If platelets are adjusted to correspond to whole-blood counts, patients with cirrhosis generate less thrombin than controls for platelet counts of less than 80,000/uL. The correlation between thrombin generation and bleeding or thrombotic events is currently unknown.

In liver disease, minor bleeding may be present, such as from gums or nose, but major bleeding is less frequent. The role of haemostatic abnormalities in the risk of variceal bleeding is not clear.

Impaired laboratory studies of haemostasis, which can indicate defective hepatic protein synthesis, are more a surrogate of severe disease than a risk factor for hemorrhage. PT and INR have been demonstrated to be independently associated with risk of re-bleeding in patients with varices and severe liver dysfunction, but are not good prognostic indicators for a first bleeding episode.

Hyperfibrinolysis has been shown to be correlated with an increased risk of variceal bleeding in a cohort of 61 cirrhotics. Higher levels of fibrinogen degradation products were associated with a greater risk of variceal bleeding compared to patients without them (odds ratio of 8), but Child-Pugh score and endoscopic characteristics of varices remained the most important prognostic factors.

What conditions can underlie coagulopathy in liver disease:

Pathophysiology of coagulation abnormalities in patients with liver disease

The liver is the site of synthesis of fibrinogen and factors II, V, VII, IX, X, XI, and XII. VWF is synthetised predominantly by endothelial cells. Factor VIII is synthetised mainly by hepatic sinusoidal endothelial cells, but also by endothelial and non-parenchymal cells in the kidney, spleen, lungs, and brain. Thus, the plasma concentration of factor VIII is not decreased with liver disease, and may even be increased, because many chronic liver diseases are associated with inflammation. These entities are associated with increased endothelial synthesis of factor VIII and/or reduced clearance of factor VIII (clearance normally occurs via a low-density lipoprotein receptor-related protein, which is synthesized by the liver).

Vitamin K is an essential cofactor for the production of biologically active forms of the coagulation factors II, VII, IX, and X, and the anticoagulant proteins, C and S. Vitamin K promotes hepatic post-ribosomial conversion of glutamic acid residues in protein precursors, to gamma-carboxyglutamic acid (Gla) forms that chelate calcium and allow effective hemostatic function. When gamma-carboxylation is impaired due to deficiency or antagonism of vitamin K, inert precursers are synthesized, (known as proteins induced by vitamin K absence [PIVKA]) and released into the blood.

In cholestasis, reduction of vitamin K absorption from the small intestine due to decreased bile salt production can be treated by parenteral administration of vitamin K (10mg daily for 24 to 48 hours). In parenchymal liver disease, decreased levels of coagulation and inhibitor proteins are the result of decreased synthesis, and so there is often no improvement with vitamin K administration. However, 25% of patients with acute liver injury have a sub-clinical deficit of vitamin K, and may respond to parenteral administration with improvement of the INR.

In acute liver failure, the first factors whose plasma concentrations fall are those with the shortest half-lives, factor V and VII (12 and 4 to 6 hours, respectively). Factors II, IX, and X subsequently also decrease. Factor VIII, together with vWF, is usually elevated. The synthesis of vWF, an acute phase reactant, is increased in tissue injury, and in endotoxemia associated with endothelial dysfunction.

In chronic liver disease, increased shear stress of endothelial cells related to portal hypertension may also contribute to the high plasma levels of vWF.

Plasma fibrinogen is also an acute-phase reactant, but usually remains normal or increased in patients with liver disease. Lower levels of fibrinogen due to decreased synthesis (although still usually greater than 100mg/dL), occur only in severe liver disease.

The high fibrinogen concentrations in patients with chronic hepatitis, cholestatic jaundice, and hepatocellular carcinoma do not result in increased clot formation, as the synthetised fibrinogen’s functions poorly. This causes a prolonged thrombin time (TT), despite almost normal PT and PTT (partial thromboplastin time) values, with a normal concentration of antigenic fibrinogen. This acquired dysfibrinogenemia is reversed following recovery of liver function.

About one third of patients with chronic liver disease develop thrombocytopenia that is usually mild to moderate (70,000 to 90,000 x 109/L), and worsens with increased hypersplenism/additional platelet sequestration. Thrombocytopenia has not been associated with an increased risk of bleeding from esophageal varices or other sites (although there are only few studies evaluating this), but does correlate with blood loss during surgery.

A higher spleen diameter/platelet count ratio has been shown to have a predictive value for the presence of esophageal varices in patients with liver cirrhosis. The issue of sequestration versus other causes of thrombocytopenia in cirrhosis has been evaluated recently by comparing platelet number in extrahepatic portal hypertension, to that of cirrhotics with a similar size spleen. There was less severe thrombocytopenia in the non-cirrhotic patients.

The liver produces thrombopoietin (TPO), which regulates platelet production in the bone marrow. Less TPO mRNA (messenger RNA) and TPO are produced in thrombocytopenic patients with severe chronic liver diseases, than in those with a normal liver; platelet production is therefore reduced in the bone marrow.

Acute infection with hepatitis C virus (HCV), alcohol abuse, and folate deficiency can all contribute to myelosuppression, further lowering platelet counts. DIC (disseminated intravascular coagulation) does not commonly influence platelet counts in cirrhosis.

Thrombocytopenia may sometimes also result from immune-mediated mechanisms. Increased production from B cells of antibodies binding platelet surface antigens GPIIb-IIIa and GPIb has been demonstrated in hepatitis B and C virus-related cirrhosis, as well as in the cholestatic liver disorders (primary biliary cirrhosis [PBC] and primary sclerosing cholangitis [PSC]). Among 368 patients with virus-induced cirrhosis, elevated titers of platelet-associated IgG (immunoglobulin G) were observed in 88% with HCV and 47% with HBV (hepatitis B virus).

A platelet count greater than 60 x109/L in platelet-rich plasma from patients with liver cirrhosis is sufficient to preserve normal thrombin generation in vitro.

Platelet function, as well as decreased platelet number, may be impaired in patients with liver disease. Platelet aggregation in response to ADP (adenosine diphosphate), arachidonic acid, collagen, and thrombin may be subnormal, probably due to a defective signal transduction pathway.

Synthesis of anticoagulant factors is decreased in patients with liver disease. Antithrombin III (ATIII) is a non-vitamin K-dependent glycoprotein synthesized by the liver and endothelium that is reduced in plasma concentration in patients with liver disease, probably predominantly due to decreased synthesis. Usually the ATIII decrease is mild, and thrombosis as a complication is rare.

Proteins C and S are vitamin K-dependent glycoproteins synthesized mainly by hepatocytes. During acute or chronic liver disease, their concentrations can decrease concomitantly with other coagulation factors (but usually not less than 20% of normal values). Genetic deficiency of protein C is rare in the general population, but found in 20% patients with Budd-Chiari syndrome (BCS). In patients with liver disease who also have genetic deficiency, plasma protein C concentrations may be less than 20%.

When there is severe liver disease, it can be difficult to exclude coexistent genetic deficiency, because levels may be very low due to decreased synthesis. In this situation, a concomitant finding of a normal level of factor II and a protein C/factor VII ratio can confirm a coexistent genetic deficit. In acquired deficiency of vitamin K, a defective protein C lacking gamma-carboxyl groups (PIVKA) is produced. Protein C deficiency is not associated with extrahepatic portal vein thrombosis. Genetic deficiency of protein S is extremely rare, yet accounts up to 7% of patients with BCS or portal vein thrombosis (PVT) in Asians.

All the proteins involved in fibrinolysis, except for tPA (tissue plasminogen activator) and PAI-1 (plasminogen activator inhibitor-1), are synthesized in the liver. Reduced plasma levels of plasminogen, alpha 2 (a2)-antiplasmin, histidine-rich glycoprotein (HRG), factor XIII, and thrombin-activable fibrinolysis inhibitor (TAFI, a carboxypeptidase) occur in patients with liver cirrhosis. Conversely, tPA levels are increased in liver disease due to decreased clearance, and the tPA inhibitor, PAI-1, is normal or only slightly increased in plasma. The inhibitor concentrations are insufficient to counteract the increase in tPA, accounting for a net increase in fibrinolysis. In contrast, in patients with acute liver failure, there are high levels of the acute phase reactant PAI-1 and hypofibrinolysis.

Hyperfibrinolysis is correlated with the severity of liver cirrhosis, as assessed by Child-Pugh score. Ascitic fluid also has increased fibrinolytic activity. Increased levels of D-dimers, prothrombin fragments 1+2 (F1+2), fibrin degradation products, and plasmin-a2-antiplasmin complexes are found. Many studies using different methodologies demonstrate hyperfibrinolysis (thromboelastography, diluted whole blood clot lysis assay, and euglobulin clot lysis time. Recently thrombin-activatable fibrinolysis inhibitor (TAFI) has been found to be decreased by 26% in cirrhosis and by 50% in acute liver failure. Other studies have not confirmed hyperfibrinolytic activity in cirrhosis.

There may be a correlation between hyperfibrinolysis and an increased risk of variceal bleeding. In addition, increased fibrinolytic activity during liver transplantation and hepatic resection correlates with blood loss.

Patients with cholestatic liver disease have higher PAI-1 concentrations, and this offsets their increased tPA activity. The clinical result is less hyperfibrinolysis in the reperfusion phase during liver transplantation, and antifibrinolytic therapy is not usually administered.

Disseminated intravascular coagulation (DIC) is characterized by intravascular fibrin deposition due to activation of the clotting cascade that overwhelms anticoagulation mechanisms. Secondary to the activation of the coagulation pathway, consumption of coagulation factors and platelets, associated with secondary fibrinolysis, occurs and causes an increased bleeding tendency. Low grade DIC and hemostatic abnormalities present in cirrhotic patients that share common laboratory features (prolonged PT and PTT, low fibrinogen level, elevated fibrin-degradation products and D-dimer, and thrombocytopenia), and can be confused with each other.

Early reports linked chronic liver disease to low grade DIC with accelerated fibrinolysis. However, the presence of DIC in liver cirrhosis is currently disputed. Although DIC-like laboratory abnormalities (“pseudo-DIC”) are observed, autopsy studies in cirrhotics have shown little fibrin deposition, and clinical manifestations of DIC are rare.

More sensitive tests quantifying proteolytic cleavage products of coagulation reactions (fibrinopeptide A, F1+2) and fibrinolysis (D-dimer, high molecular weight fibrin/fibrinogen complexes, soluble fibrin) demonstrate an abnormal profile called “accelerated intravascular coagulation and fibrinolysis” (AICF). The studies to date demonstrate AICF in about 30% of cirrhotics, depending on the severity of liver disease. In two studies, AICF correlated with a decrease in levels of ATIII, protein C, and protein S.

In contrast to these results, Ben Ari, et al. (see references below) analyzed 52 patients with stable liver disease for F1+2, thrombin-antithrombin III complexes (TAT) and D-dimer levels, and found no differences from controls. TEG studies were, however, able to detect hyperfibrinolysis. AICF may be important in the portal venous system, as the phenomenon is more pronounced there, than in systemic blood. This observation could be related to endotoxemia in portal blood, which can trigger release of IL-6 and TNF tumor necrosis factor)-alfa, and activate intravascular coagulation.

Assessment of risk of bleeding for invasive procedures

Historically, PT and platelet count have been used to assess the risk of bleeding prior to invasive procedures. Patients with cirrhosis have increased mortality and morbidity during surgery, mainly due to increased bleeding in 60% of cases. Early studies linked PT to this surgical risk (PT prolongation greater than 1.5 and greater than 2.5 seconds associated with 47% and 87% mortality, respectively). A platelet count less than 50,000/mm3 and PT increased by more than 3 seconds have been considered relative contraindications to elective surgery or invasive procedures, and have been used as thresholds to require prophylactic transfusion of platelets and clotting factors. However, the most commonly observed risk factor for bleeding during surgery may be the severity of portal hypertension and the presence of collateral veins.

A study performed in patients undergoing laparoscopic liver biopsy failed to demonstrate any correlation between bleeding at the hepatic puncture site and coagulation tests prior to the procedure, which would indicate that the extent of liver esection may be the most important risk factor for bleeding during this procedure.

Liver biopsy is widely used diagnostically to grade the severity of liver disease or fibrosis. Moreover, it is an essential tool after liver transplantation to diagnose rejection and other causes of graft dysfunction. Bleeding complications occur in 0.35 to 0.5% of patients, leading to mortality in 0.1%. Direct ultrasound guidance is often used to mark the optimal biopsy site, and may reduce the risk of complications (recommended in some guidelines).

Despite the evidence that there were no threshold abnormalities of clotting tests associated with risk of bleeding during laparoscopic liver biopsy, thresholds for INR and platelet count are still often used to determine the risk of bleeding during percutaneous liver biopsy. An audit from the British Society of Gastroenterology (BSG) performed in 1991 showed a doubling of the risk of bleeding in patients with INR greater than 1.5; however, only 7.1% of the bleeding occurred with INR greater than 1.5, and 90% occurred with an INR less than 1.3. Thus, having a normal INR no more excludes the risk of bleeding than an elevated INR suggests it.

A cut-off for platelet count is difficult to justify from the literature. Most textbooks in the UK and the BSG guidelines require a platelet count above 80,000/mm3, whereas a survey form the Mayo Clinic suggested 50,000 mm3 as the cut-off. Current recommendations state that a percutaneous liver biopsy can be done safely without platelet transfusion support if platelet counts are above 60,000/mm3.

Burroughs, et al. (see references below) advocated evaluating the use of bleeding time to assess the risk of bleeding for percutaneous liver biopsy, but this has not become widespread clinical practice. If clotting parameters are outside stipulated ranges, a transjugular liver biopsy can be performed more safely (without plasma or platelet therapy), despite the recommendation for percutaneous liver biopsy in some guidelines. A plugged liver biopsy may also be safer than standard percutaneous biopsy, except that it may cause a greater risk of bleeding in hypocoagulable patients.

There are no firm guidelines for clotting test result requirements in order to perform thoracentesis, paracentesis, lumbar puncture, or dental extraction in patients with liver disease without the use of blood products. The largest review on 608 patients with cirrhosis who underwent paracentesis or thoracentesis with mild coagulation abnormalities reported 0.2% excessive bleeding requiring transfusion and 0.02% mortality. There was no correlation with PT, PTT, or platelet count and the risk of bleeding. In another study which evaluated paracentesis in 200 cirrhotics with an INR greater than 3 and platelet count of less than 19,000/mm3, no complications were seen, regardless of baseline INR and platelet count.

Central venous catheterization can be safely performed in patients with abnormalities of haemostatis without blood product infusion. The internal jugular route should be used. Platelet transfusions for platelet counts less than 50,000/mm3 is sometimes used, although evidence for this is meager. If hyperfibrinolysis is present, its correction reduces the risk of bleeding during central venous access.

A contraindication to invasive procedures is clinically evident DIC or fibrinolysis.

A prospective multicenter study on the prognostic role of coagulation testing in predicting gastrointestinal bleeding after variceal band ligation failed to find any correlation.

Currently, there are no evidence-based reports to establish “safe” coagulation test results for patients to undergo various procedures. However, it is important to limit invasive procedures to those where the perceived benefit is clearly greater than the risk of bleeding, because the latter cannot be accurately estimated. The recent findings that many patients with cirrhosis may not be hypocoagulable may well influence recommendations for use of plasma products and platelets in the future.

When do you need to get more aggressive tests:


What imaging studies (if any) will be helpful?


What therapies should you initiate immediately and under what circumstances – even if root cause is unidentified?

During acute variceal bleeding and massive blood transfusion, in order to avoid dilutional decrease of clotting factors, for every 2 units of blood, 1 unit of fresh-frozen plasma (FFP) is typically given.

A report described initially effective hemostasis after infusion of recombinant activated factor VII (rFVIIa) in 10 patients with variceal bleeding, but six experienced early re-bleeding and all of them died. In a cohort of eight patients with acute variceal bleeding uncontrolled with endoscopic and medical therapy, rFVIIa administration achieved hemostasis in 25% of patients after a single dose, but all re-bled. The safety of rFVIIa, especially regarding possible prothrombotic effect or triggering of DIC, remains to be assessed in large studies of patients with liver disease.

Two randomized controlled trials which evaluated the role of rFVIII in patients with variceal bleeding did not show a decrease in morbidity and mortality.

What other therapies are helpful for reducing complications?

Therapy for hemostatic abnormalities of liver disease is needed only during variceal bleeding, surgery, or before invasive procedures.

Intravenous vitamin K injection of 10mg daily for 24 to 48 hours can reverse vitamin K deficiency. FFP contains all clotting factors and can often correct an elevated PT, depending on the volume infused and the baseline PT abnormality. Whether or not this correction of the PT results in improved hemostasis is not known. In addition, correction is brief (24 to 48 hours), and especially dependent on the half-life of factor VII.

A common indication for FFP infusion is the presence of persistent bleeding in patients with INR greater than or equal to 2 or PT prolongation greater than 4 seconds. In surgical or invasive procedures, attaining 50% of the normal PT (that is, INR of 2) is a target for replacement therapy, and for neurological procedures such as intracranial pressure monitoring during liver failure, 80% of normal PT (that is, an INR of about 1.2 to 1.3).

To increase the activity of a clotting factor by 1 to 2%, a dose of 1mL FFP/kg of body weight is necessary.

Despite the possible improvement in standard laboratory assessment of coagulation after administration of FFP in cirrhotic patients, thrombin generation is not increased after addition of FFP in vitro to the plasma of cirrhotic patients. Because of the high volume required, adequate replacement is difficult both in cirrhotic patients (intravascular plasma volume is already expanded and ascites may be present), and in acute liver failure (increasing plasma volume can lead to increases in intracerebral pressure). Moreover, the short half-life of factor VII requires FFP infusion every 6 to 12 hours in patients with INR greater than 1.5.

FFP is given (12 to 15mL/kg) before liver biopsy, but there is no evidence supporting this. Transjugular biopsy should be used in patients with coagulopathy not sufficiently corrected with FFP. Platelet transfusion, 1 unit every 10kg, is typically administered. The platelet count should be checked 1 hour after the infusion; however, no correlation among amelioration of bleeding time, increase in platelet count, and enhanced hemostasis has been shown.

Cryoprecipitate contains factors VIII, vWF, fibrinogen, fibronectin, and factor XIII. Because of the small volumes (30 to 50mL/U/10kg) required, it can be useful in liver cirrhosis and acute liver failure, but it lacks some coagulation factors and may worsen the imbalance already present.

Desmopressin (1-deamino-8-D-arginine vasopressin [DDAVP]), analogue of a diuretic hormone, increases the plasma level of factor VIII and vWF by increasing the release of vWF from endothelial storage sites. It can improve bleeding time and enhance primary hemostasis at the dose of 0.3μg/kg in patients with liver failure. In a recent randomized trial, DDVAP failed to decrease blood loss during hepatic resection, despite increased factor VIII and vWF.

ATIII infusion is not routinely recommended.

RFVIIa was first developed for the treatment of patients with hemophilia A and B who developed inhibitors. It may also have a promising role in the treatment of the coagulation disorders in liver disease, as recombinant factor VIIa corrects prolonged PT in a dose dependent manner in non-bleeding cirrhotic patients. However, a randomized study using rFVIIa in 71 patients undergoing laparoscopic liver biopsy found no differences in liver bleeding. Two complications occurred in the rFVIIa group (one DIC and one peripheral venous thrombosis).

In acute liver failure, rFVIIa may be useful in normalizing PT in the setting of intracranial pressure monitoring, as only a small volume of infusion is required. In a randomized trial of variceal bleeding, a modest reduction of early re-bleeding was observed in a subgroup of “Child B” and “Child C” patients after rFVII infusion, although no difference in control of bleeding or transfusion requirements was shown overall.

A recent meta-analysis of four randomized clinical trials on the prophylactic use of rFVII in cirrhotic patients undergoing liver resection or liver transplantation showed no significant difference between rFVII and placebo in terms of mortality, red blood cell units transfused, or adverse events.

What should you tell the patient and the family about prognosis?

What are the important prognostic indicators?

In cirrhosis, plasma levels of coagulation factors are indicators of hepatic synthesis, and thus of liver function. A prolonged PT, which is not corrected by intravenous vitamin K administration (10mg daily for 2 days) helps differentiate vitamin K deficiency from parenchymal liver diseases. PT is part of the Child-Pugh score, which is the most commonly used prognostic assessment of the severity of liver disease. Recently the MELD (Model for End-Stage Liver Disease) score, which incorporates INR, has been used to allocate priority for liver transplantation in the United States of America based on estimated probability of death within 3 months.

Determination of individual coagulation factors adds little prognostic information to the PT or INR. Determination of factors V, VII, and VIII are useful indicators of DIC and prognostic indicators in acute liver failure. The Clichy criteria indicate poor prognosis and need for liver transplantation in acute liver failure when factor V is below 20% in patients aged 30 or older, and above or below 30% in patients below age 30.

Factor V has less prognostic value in acetaminophen-induced fulminant hepatic failure. In the King’s College criteria for acetaminophen induced liver failure, PT greater than or equal to 100 seconds is a prognostic indicator on its own for liver transplantation, independently of the grade of coma.

In patients with non-acetaminophen induced acute liver failure, PT greater than or equal to 50 seconds, together with two of the following criteria, age less than 10 and more than 40 years, drug toxicity, interval between jaundice and encephalopathy onset after 7 days, and serum bilirubin greater than 300μmol/L are indications of poor prognosis and the need for liver transplantation.

A multivariate analysis of prognostic factors in cirrhotic patients showed that the level of factor VII was an independent predictor of survival, that is, factor VII less than 34% was predictive of a 93% mortality.

Plasma concentration of D-dimer greater than 300ng/mL has been shown to be correlated with adverse outcome in patients with liver cirrhosis, but has only a 21% positive predictive value with a relative risk of 7.5. Recently, a vWF serum level greater than 260U/dL has been correlated with complications of portal hypertension in patients with end stage liver disease.

"What if" scenarios

Which coagulation abnormalities would be expected when the cirrhotic patient is infected?

The overall cumulative incidence of infection in patients with cirrhosis is estimated to be at least 30%, and is possibly associated with increased risk of variceal bleeding. Infection is associated with early variceal re-bleeding and increased mortality. Prophylactic antibiotic therapy led to less early re-bleeding and better control of bleeding in two randomized controlled trials.

Using TEG, 20 cirrhotic patients who experienced early re-bleeding were found to have worsening TEG parameters the day before the episode. Moreover, patients with a bacterial infection have worse TEG parameters, which can be corrected in vitro by heparinase cleavage of heparin-like substances.

The presence of heparin-like substances is associated in some, with increased anti-Xa activity. Heparin-like substances have been detected hours after variceal bleeding in cirrhotic patients. Based on these observations, the hypothesis is that endotoxins and inflammation due to infection can release heparinoids from the endothelium and mast cells. One study showed increased heparan sulphate concentrations in patients with variceal bleeding, compared to patients without. Moreover, sepsis can cause decreased platelet number and aggregability.

Cytokines IL(interleukin)-6 and TNF-alfa are released during infection and can trigger DIC with hyperfibrinolysis. One study showed a strong association between both fragment F1+2 and D-dimer with endotoxemia. These markers returned to normal after antibiotic therapy. Another recent report showed decreased platelet counts and levels of factor VII, X, V, and II in cirrhotic patients with severe sepsis, suggesting consumptive coagulopathy. Another study found decreased activity of protein C, which is associated with increased risk of thrombosis.

Prevention of infection during variceal bleeding is a well-defined treatment which has been proven to reduce re-bleeding and mortality. The effect may also be associated with preventing a worsening of portal hemodynamics. This has been shown indirectly in a single study by Del Arbol in which hepatic venous pressure gradient was higher following spontaneous bacterial peritonitis in some patients, even after resolution of infection. This was confirmed in another study reporting that hepatic venous pressure gradients decreased after intestinal decontamination with rifaximin in patients with alcohol-related decompensated cirrhosis, possibly because plasma endotoxin levels were reduced.

Derangement of coagulation due to infection has been shown using TEG, hours after the bleeding, and this could be another reason why antibiotics are effective.

Infusion of protamine sulphate has been used with the intent of reverting the heparin-like effect seen after reperfusion during liver transplantation in cirrhotic patients. However, reduction in bleeding and transfusion requirement was not demonstrated. Use of heparinase (Neutralase) to treat coagulopathy due to the release of exogenous heparin in patients undergoing coronary artery bypass surgery worsened the outcome, compared to standard protamine treatment. Theoretically, the ability of heparinase to cleave endogenous heparinoids more specifically than protamine could offer an advantage in patients with demonstrable heparin-like-effect, due to endogenous glycosaminoglycans in cirrhotics with infection or bleeding; however, its clinical utility, if any, is unknown.


See “What conditions can underlie coagulopathy in liver disease”

What other clinical manifestations may help me to diagnose coagulopathy in liver disease?


What other additional laboratory studies may be ordered?

Recently, the role of infection and endogenous heparin-like substances demonstrated by TEG has been evaluated in variceal bleeding. Infection may be a trigger for bleeding, and both infection and heparin-like substances may be responsible for the persistence of bleeding in some patients. TEG, which is a quick and reliable method to assess clot formation and lysis, also allows detection of heparin-like substances. Studies from our group have shown worsening coagulation during infection, due to low molecular weight heparin-like substances in the blood (detected by TEG).

What’s the evidence?

Lisman, T, Caldwell, SH, Burroughs, A. “Hemostasis and thrombosis in patients with liver disease: the ups and downs”. J Hepatol. vol. 53. 2010. pp. 362-71. [This consensus paper is derived from the annual convention of coagulopathy in liver disease held by a panel of experts. The review focused on the new concepts of rebalanced haemostasis in patients with liver disease, analyzing all possible clinical manifestations.]

Tripodi, A, Salerno, F, Chantarangkul, V. “Evidence of normal thrombin generation in cirrhosis despite abnormal conventional coagulation tests”. Hepatology. vol. 41. 2005. pp. 553-8. [The authors evaluated haemostatic balance in patients with liver cirrhosis beyond conventional laboratory tests by quantifying thrombin generation. The addition of thrombomodulin to the test activated the protein C anticoagulant pathway and demonstrated that patients with liver cirrhosis are characterized by a similar thrombin generation compared to controls, despite prolonged clotting time using standard laboratory tests.]

Amitrano, L, Brancaccio, V, Guardascione, MA. “Inherited coagulation disorders in cirrhotic patients with portal vein thrombosis”. Hepatology. vol. 31. 2000. pp. 345-8. [In this manuscript the authors, for the first time, clearly demonstrated the association of genetic thrombophilic defects with the occurrence of portal vein thrombosis in patients with cirrhosis. A thrombophilic genotype was detected in 69.5% of the patients with PVT, being in most of the cases a mutation of prothrombin gene.]

Violi, F, Ferro, D, Basili, S. “Hyperfibrinolysis increases the risk of gastrointestinal hemorrhage in patients with advanced cirrhosis”. Hepatology. vol. 15. 1992. pp. 672-6. [The authors, for the first time in a prospective study with a follow-up of 3 years, demonstrated a correlation between hyperfibrinolysis and the risk of variceal bleeding in cirrhotics. In particular, patients with positive fibrinogen degradation products or class C had a higher risk of gastrointestinal bleeding than patients with negative fibrinogen degradation products (odds ratio = 8 [for fibrinogen]) or class B (odds ratio = 3.5 [for class C against class B]). This remains the only consistent correlation found between coagulation index and risk for first bleeding in patients with end-stage liver disease.]

Ben, Ari, Panagou, M, Patch, D. “Hypercoagulability in patients with primary biliary cirrhosis and primary sclerosing cholangitis evaluated by thrombelastography”. J Hepatol. vol. 26. 1997. pp. 554-9. [Patients with primary biliary cirrhosis and primary sclerosing cholangitis survive variceal bleeding better than patients with alcoholic cirrhosis, and have less bleeding at liver transplantation. The authors demonstrated by thrombelastography, in 13 of 47 (28%) primary biliary cirrhosis and in nine of 21 (43%) primary sclerosing cholangitis, a hypercoagulable state.]

Ben, Ari, Osman, E, Hutton, RA, Burroughs, AK. “Disseminated intravascular coagulation in liver cirrhosis: fact or fiction?”. Am J Gastroenterol. vol. 94. 1999. pp. 2977-82. [In this manuscript the authors demonstrated, by the Clot Lysis Index, a mild hyperfibrinolysis in patients with cirrhosis compared to controls, whereas no difference in thrombin generation or degradation was found. This is in contrast with some other papers, which showed an accelerated intravascular coagulation and fibrin degradation in patients with chronic liver disease.]

Piscaglia, F, Siringo, S, Hermida, RC. “Diurnal changes of fibrinolysis in patients with liver cirrhosis and esophageal varices”. Hepatology. vol. 31. 2000. pp. 349-57. [Variceal bleeding, whose triggering mechanisms are largely unknown, occurs with a circadian rhythmicity with two peaks, one greater in the evening, and one smaller in the early morning. tPA activity showed a circadian rhythm in cirrhosis, with a peak of 2.85 times the trough value, calculated at 18:42. Total fibrinolytic activity showed a similar pattern, which was statistically significant also in controls. tPA and PAI antigens also showed significant circadian rhythm both in controls and cirrhotics, with higher values in the morning. The existence of a circadian rhythm of fibrinolysis in cirrhosis, whose temporal distribution might suggest a role of fibrinolysis in variceal hemorrhage on the basis of the comparison to the known chronorisk of variceal bleeding.]

Goulis, J, Armonis, A, Patch, D. “Bacterial infection is independently associated with failure to control bleeding in cirrhotic patients with gastrointestinal hemorrhage”. Hepatology. vol. 27. 1998. pp. 1207-12. [Bacterial infection is frequently diagnosed in cirrhotic patients with variceal hemorrhage. In this prospective study, multivariate analysis showed that proven bacterial infection (P less than .0001) or antibiotic use (P less than .003), as well as active bleeding at endoscopy (P less than .001) and Child-Pugh score (P less than .02) were independent prognostic factors of failure to control bleeding. This is an important message which changed the care of cirrhotics with variceal haemorrhage.]

Bernard, B, Cadranel, JF, Valla, D. “Prognostic significance of bacterial infection in bleeding cirrhotic patients: a prospective study”. Gastroenterology. vol. 108. 1995. pp. 1828-34. [In this study, the authors demonstrated that infections are common in patients with cirrhosis and recent variceal bleeding, and that bacterial infections are associated with re-bleeding and mortality.]

Steib, A, Freys, G, Lehmann, C. “Intraoperative blood losses and transfusion requirements during adult liver transplantation remain difficult to predict”. Can J Anaesth. vol. 48. 2001. pp. 1075-9. [In this study, the authors failed to demonstrate any correlation between standard laboratory coagulation tests and intraoperative blood losses in patients with liver cirrhosis undergoing liver transplantation. This is in keeping with other reports which evaluated standard and extended tests of haemostasis in the prediction of blood requirements during liver transplantation.]

Tripodi, A, Mannucci, PM. “The coagulopathy of chronic liver disease”. N Engl J Med. vol. 365. 2011. pp. 147-56. [This is an extensive review on the haemostatic changes in patients with chronic liver diseases. It is focused on the pathogenesis of haemostatic alterations, with few considerations of clinical implications and clinical management of these patients.]

Giannini, E, Botta, F, Borro, P. “Platelet count/spleen diameter ratio: proposal and validation of a non-invasive parameter to predict the presence of oesophageal varices in patients with liver cirrhosis”. Gut. vol. 52. 2003. pp. 1200-5. [The authors demonstrated that screening cirrhotic patients for oesophageal varices (OV) according to spleen diameter and platelet count/spleen ratio was more effective than diffuse scoping. In particular, a platelet count/spleen diameter ratio cut off value of 909 had 100% negative predictive value for a diagnosis of OV.]

Senzolo, M, Burra, P, Cholongitas, E, Burroughs, AK. “New insights into the coagulopathy of liver disease and liver transplantation”. World J Gastroenterol. vol. 12. 2006. pp. 7725-36. [In this review, the authors focused on pathogenesis of coagulation abnormalities in liver disease and liver transplantation, with special attention to clinical consequences of coagulation derangements during pre- and intra-operative care of patients undergoing liver transplantation.]

Senzolo, M, Burroughs, AK, Ginès, P, Kamath, PS, Arroyo, V. “Correction of Abnormalities of Haemostasis in Chronic Liver Disease”. 2011. pp. 453-476. [In this chapter, the authors reviewed all the available treatments to correct coagulation abnormalities in patients with liver disease before invasive procedures or surgery, including liver transplantation. This review is evidence based and gives readers a clear guideline for clinical practice.]

Senzolo, M, Burroughs, AK, Kitchens, G, Alving, B, Kessler, G. “Hemostasis alterations in liver disease and liver transplantation”. vol. 2007. pp. 647-659. [In this chapter, the authors reviewed all the evidence on the pathogenesis of haemostatic alterations in patients with chronic liver disease of different aetiologies and the changes during and after liver transplantation.]