NEJM —The Coagulopathy of Chronic Liver Disease

Armando Tripodi, Ph.D., and Pier Mannuccio Mannucci, M.D.

N Engl J Med 2011; 365:147-156July 14, 2011DOI: 10.1056/NEJMra1011170

Article

Chronic liver disease,
particularly in the end stage, is characterized by clinical bleeding and
decreased levels of most procoagulant factors, with the notable
exceptions of factor VIII and von Willebrand factor, which are elevated.1
Decreased levels of the procoagulants are, however, accompanied by
decreases in levels of such naturally occurring anticoagulants as
antithrombin and protein C.1 In physiologic conditions, the coagulation system is balanced by these two opposing drivers (Figure 1Figure
1Coagulation with Thrombin Generation and Inhibition.), but the
mechanistic significance of the parallel decrease of both procoagulants
and anticoagulants in patients with chronic liver disease escaped
attention for many years. As a consequence, chronic liver disease is
still considered the epitome of acquired bleeding disorders and is
featured as such in most hematology textbooks. The basic laboratory
tests of coagulation (i.e., measurement of the prothrombin time and
activated partial-thromboplastin time) have been used to assess the risk
of bleeding.

However, their results are poorly correlated with
the onset and duration of bleeding after liver biopsy or other
potentially hemorrhagic procedures.2-7
These test results are also poorly correlated with the occurrence of
gastrointestinal bleeding, the prototype of hemorrhagic events in
patients with end-stage liver disease.8,9
Additional evidence that argues against the clinical relevance of the
coagulation defects as detected by conventional laboratory tests in
determining the bleeding tendency in these patients can be drawn from
the natural history of liver transplantation. In the past, this major
surgical procedure required massive transfusions of plasma and other
blood products to correct the marked abnormalities on tests of
hemostasis (assessments of coagulation, platelets, and fibrinolysis)
observed both preoperatively and perioperatively. The need for
transfusions, however, has declined considerably over time — not because
of any substantial change in medication, but rather because of improved
surgical procedures.10
Finally and most important, randomized clinical trials involving
patients with chronic liver disease have shown that powerful
procoagulant agents, such as recombinant activated factor VII, fail to
control bleeding from the upper intestinal tract11,12 or bleeding during liver transplantation,13,14 even though the postinfusion prothrombin time is considerably shortened.11
In this review, we consider the evidence regarding the balance in the
hemostatic system (involving coagulation, platelets, and fibrinolysis).

The
aforementioned observations question the validity of the
prothrombin-time test and related tests for assessing the risk of
hemorrhage and guiding the transfusion of fresh-frozen plasma or use of
procoagulant agents in patients with chronic liver disease. An old dogma
is being dispelled in favor of the newly emerging concept that blood
coagulation in such patients is rebalanced, owing to the parallel
reduction of procoagulant and anticoagulant factors (Table 1Table
1Patterns of Prohemostatic and Antihemostatic Drivers in the Different
Phases of Hemostasis in Patients with Chronic Liver Disease.). Indeed,
studies show that plasma from patients with cirrhosis generates as much
thrombin (the final enzyme of coagulation) as plasma from healthy
subjects, provided that thrombin is measured by methods that reflect the
action of both procoagulants and anticoagulants.23,24
Thrombin generation in vivo and in vitro is down-regulated by
thrombomodulin, a transmembrane protein situated on vascular endothelial
cells that acts as the main physiologic activator of protein C (Figure 2Figure
2Protein C Activation by Thrombin on the Membrane of Endothelial Cells,
and the Balance of Antihemostatic and Prohemostatic Drivers in the
Different Phases of Hemostasis.).25
Plasma and reagents that are used to measure the prothrombin time do
not contain thrombomodulin. Accordingly, this test measures the amount
of thrombin generated in plasma as a function of the procoagulant
drivers, but not the thrombin inhibited by the anticoagulant drivers,
especially protein C, which is not fully activated in the absence of
thrombomodulin. This might explain why the prothrombin-time test and
related tests do not truly represent the balance of coagulation in vivo
and are inadequate for assessing the risk of hemorrhage in those
acquired conditions, such as the coagulopathies of liver disease and
neonatal coagulopathies, in which there is a restored balance due to the
concomitant decrease of procoagulants and anticoagulants.26

As
for end-stage liver disease, another problem is that the prothrombin
time expressed as the international normalized ratio (INR) is widely
used as a prognostic index to calculate the patient's Model for
End-Stage Liver Disease (MELD) score, which is used to prioritize
candidates for liver transplantation. However, the INR was devised and
validated to standardize across laboratories the prothrombin times in
patients receiving anticoagulation therapy with vitamin K antagonists
such as warfarin and its congeners. The INR cannot be used for patients
with chronic liver disease unless an alternative system of
standardization specifically developed for them is adopted.27
This alternative system involves using a different calibration based on
plasma from patients with chronic liver disease rather than plasma from
patients receiving vitamin K antagonists.

Together, the above
observations indicate that the bleeding tendency frequently observed in
patients with end-stage liver disease should be explained by mechanisms
other than hypocoagulability, such as those triggered by underlying
conditions that favor hemorrhage (i.e., hemodynamic alterations
subsequent to portal hypertension, endothelial dysfunction, bacterial
infections, and renal failure20,28-31) (Table 2Table
2Underlying Conditions That Explain the Bleeding Tendency in Patients
with Decompensated Chronic Liver Disease.). It should also be understood
that although rebalanced, the coagulation system in patients with
chronic liver disease is not as stable as that in healthy persons, who
have an excess of both procoagulants and anticoagulants. Therefore, the
relative deficiency of both coagulation-system drivers makes the balance
fragile in patients with liver disease and may tip it toward hemorrhage
or thrombosis, depending on the prevailing circumstantial risk factors (Figure 2C).

Under
normal conditions, platelets have a dual function. They adhere to
damaged vessel walls through an interaction with the multimeric adhesive
protein von Willebrand factor, thus promoting aggregation and
ultimately the formation of the primary hemostatic plug. Platelets also
support thrombin generation by assembling activated coagulation factors
on their surfaces. Thrombocytopenia, a typical feature of chronic liver
disease,17 may therefore be another cause of bleeding (Table 1).
However, very high levels of von Willebrand factor, a common finding in
patients with chronic liver disease, may restore platelet adhesion to
the subendothelium at sites of vascular injury (Table 1), as shown by in vitro experiments carried out under flow conditions mimicking those that occur in vivo.15
Levels of ADAMTS 13, a naturally occurring plasma metalloprotease that
limits in vivo the functions of von Willebrand factor on platelets, are
reduced in patients with cirrhosis16; this may further contribute to the restoration of platelet function (Table 1). Finally, a platelet count as low as 60×10⁹
per liter in platelet-rich plasma from patients with cirrhosis is
usually sufficient to preserve thrombin generation at a level equivalent
to the lower limit of the normal range in healthy subjects.24

Fibrinolysis
is a highly regulated mechanism that, on deposition of fibrin within
the vascular system, converts the proenzyme plasminogen into the active
enzyme plasmin, which in turn degrades fibrin (Figure 3Figure
3Fibrinolysis Activation and Inhibition.). Under normal conditions,
plasminogen-to-plasmin conversion is regulated by such activators as
tissue plasminogen activator (t-PA), urokinase plasminogen activator,
and activated factor XII. These activators (profibrinolytic drivers) are
opposed by such antiactivators as t-PA inhibitors (mainly, plasminogen
activator inhibitor [PAI]), plasmin inhibitor, and thrombin-activatable
fibrinolysis inhibitor (TAFI), which cumulatively act as
antifibrinolytic drivers. Any perturbation of this balance may result in
hyperfibrinolysis, which increases the risk of hemorrhage, or
hypofibrinolysis, which increases the risk of thrombosis.

Plasma
hyperfibrinolysis has been reported in patients with chronic liver
disease, but its mechanistic role in bleeding is still debated.20
Uncertainty rests mainly on the lack of appropriate laboratory tests
for its evaluation, because most observations are based on the
measurement of the individual components of the system rather than on
the overall activity stemming from the action of both profibrinolytic
and antifibrinolytic drivers. Cirrhosis has been variably associated
with laboratory changes favoring hyperfibrinolysis, such as increased
levels of t-PA and reduced levels of plasmin inhibitor and TAFI, but
also with changes favoring hypofibrinolysis, such as reduced levels of
plasminogen and increased levels of PAI (Table 1).
Hence, although contrasting results have been reported, the balance of
fibrinolysis is probably restored in patients with liver disease by the
parallel changes in profibrinolytic and antifibrinolytic drivers.21,22

Overall,
the aforementioned observations suggest that patients with chronic
liver disease are not naturally “autoanticoagulated,” as previously
believed. This concept is reinforced by clinical evidence indicating
that they are not protected from32,33 and may even be at increased risk for34 thrombosis, particularly but not exclusively in the portal venous system,35,36 and especially in the presence of inherited prothrombotic mutations.37

Laboratory signs of a procoagulant imbalance, which was not evident in the previous studies,23,24 have been reported in association with chronic liver disease.18,19 As noted above, thrombin generation in vivo and in vitro is down-regulated by thrombomodulin (Figure 2),25
which effectively quenches thrombin generation when added to plasma
from healthy subjects but is much less effective when added to plasma
from patients with chronic liver disease.18
This indicates that in such patients, the plasma is partially resistant
to anticoagulation mediated by thrombomodulin. This resistance is
evident only when the results of thrombin-generation tests are expressed
as the ratio of thrombin activity in the presence of thrombomodulin to
thrombin activity in its absence. The resistance is probably the result
of two alterations typically found in patients with chronic liver
disease18,19: markedly increased plasma levels of factor VIII (one of the most potent drivers of thrombin generation38)
and the concomitant decrease in levels of protein C (one of the most
potent anticoagulant drivers in quenching thrombin generation25).
Although protein C is reduced owing to the impaired synthetic capacity
of the liver, the increased levels of factor VIII are likely to be
explained by decreased clearance of this moiety from plasma,39
mediated by two mechanisms, one involving von Willebrand factor, and
the other the low-density lipoprotein receptor–related protein.39 Von Willebrand factor binds factor VIII in vivo and protects it from cleavage by plasma proteases and from premature clearance.40 High plasma levels of von Willebrand factor in patients with cirrhosis15
may be mechanistically involved in maintaining high plasma levels of
factor VIII through the stabilization of its procoagulant activity. The
low-density lipoprotein receptor–related protein, a multifunctional
ligand that mediates the cellular uptake and subsequent degradation of
factor VIII,41 is inadequately expressed in patients with cirrhosis39 and, in conjunction with high levels of von Willebrand factor, may help sustain the high plasma levels of factor VIII.

The
procoagulant imbalance associated with chronic liver disease can be
detected by measuring thrombin generation in plasma in the presence and
absence of thrombomodulin.18 An alternative method uses a snake-venom extract (Protac, Pentapharm)19
that acts as a surrogate activator of protein C in a manner similar to
that of thrombomodulin. Whereas the results of the first test are
expressed as the ratio of the thrombin concentration generated in the
presence of thrombomodulin to the concentration generated in its
absence,18
the results of the second test are expressed as the percentage of
extract-induced coagulation inhibition, measured as the amount of
thrombin generated in the presence versus the absence of the venom
extract.19
By definition, the higher the ratio or the lower the percentage of
extract-induced coagulation inhibition, the greater the degree of
procoagulant imbalance. As detected by these assays in the context of
chronic liver disease, the procoagulant imbalance is negatively
correlated with levels of plasma protein C and positively correlated
with levels of factor VIII.18,19 Furthermore, the degree of imbalance increases with the severity of cirrhosis as assessed by the Child–Pugh score.18,19
Whether the procoagulant imbalance detected in the laboratory as
thrombomodulin resistance is a risk factor for thrombosis in patients
with chronic liver disease remains to be established by prospective
studies. It must be recognized that although thrombin-generation tests
mimic the conditions operating in vivo much more closely than do
conventional tests, they remain artificial because they use
platelet-free plasma and the amount of thrombomodulin added in vitro is
chosen arbitrarily, not on the basis of the density of the protein on
endothelial cells.

The in vitro procoagulant imbalance associated with chronic liver disease18,19
may have clinical implications. First, it calls into question the
unrestricted use of plasma infusion to correct the results of
conventional coagulation tests in patients undergoing invasive
procedures. This is still a common practice, despite a lack of evidence
from controlled, randomized trials and the recent guidelines of the
American Association for the Study of Liver Diseases, which warn against
the indiscriminate use of plasma therapy before liver biopsy.42 Second, the procoagulant imbalance18,19
may help explain mechanistically why these patients are not protected
from clinical events such as peripheral-vein thrombosis, portal-vein
thrombosis, atherothrombosis, and the progression of liver fibrosis. In
the next sections, these potential clinical implications are discussed.

Retrospective
studies showed that patients with chronic liver disease are not
protected from venous thromboembolism (deep-vein thrombosis and
pulmonary embolism).32,33 Recently, a nationwide, population-based case–control study34
involving 99,444 patients with venous thromboembolism and 496,872
controls showed that patients with liver disease had an increased
relative risk of venous thromboembolism, with the risk being greater for
deep-vein thrombosis than for pulmonary embolism and for cirrhosis than
for noncirrhosis liver disease. However, other studies have shown a low
prevalence of venous thromboembolism among patients with chronic liver
disease.43,44
The retrospective design of all these studies makes it difficult to
assess the true risk of venous thromboembolism among such patients. It
is clear, however, that patients with chronic liver disease are not
autoanticoagulated and may eventually have clinical manifestations of
thromboembolism, even though the abnormal results of conventional
coagulation tests would suggest the opposite.

Thrombosis in
patients with chronic liver disease might become an emerging issue owing
to their increasing life expectancy and changing lifestyle, which
expose them much more than in the past to such circumstantial risk
factors as tumors, surgery, obesity, prolonged hospitalization, and
inadequate physical activity. Thus, the logical consequence is that
patients with chronic liver disease who have peripheral-vein thrombosis
should be treated with anticoagulants just as any other patient would;
it is important to note that the long-term safety of this approach has
not been studied. Furthermore, the in vitro procoagulant imbalance
associated with chronic liver disease, confirmed by many independent
studies,45-47
suggests that these patients are eligible for antithrombotic
prophylaxis when exposed to such risky situations as major surgery and
prolonged immobilization. This notion contradicts current clinical
practice, whereby patients with cirrhosis often receive no or suboptimal
prophylaxis because of the perceived risk of bleeding.48 Clinical studies are needed to determine the appropriate care of these patients.

Even
though it is not firmly established that patients with chronic liver
disease have an increased risk of arterial thrombosis (i.e., coronary
artery disease and stroke), they are not free from these and other
clinical manifestations of atherothrombosis.49
Furthermore, the occurrence of hepatic-artery occlusion after liver
transplantation worsens the prognosis for these patients. Therefore,
early detection of this complication is important.50
Whether aspirin or other antiplatelet agents are indicated in the
primary prophylaxis of this complication warrants evaluation in clinical
trials.

The prevalence of portal-vein thrombosis
in patients with cirrhosis increases with the severity of the disease:
approximately 1% among patients with compensated cirrhosis35 but 8 to 25% among those who are candidates for liver transplantation.36 Because not only reduced flow velocity51 but also procoagulant imbalance and vessel-wall abnormalities (Virchow's triad)52
are mechanistic factors in this complication, antithrombotic therapy
(low-molecular-weight heparin or vitamin K antagonists) is commonly
used.53,54 This approach is relatively safe,36 but varices may need to be treated (with vasoactive drugs or endoscopic ligation)28
before patients start taking anticoagulants. Portal-vein thrombosis
worsens the post-transplantation prognosis, so primary prevention with
low-molecular-weight heparin or vitamin K antagonists should be
considered in patients awaiting liver transplantation. Randomized
clinical trials to test the efficacy of these drugs are under way.55
However, because of the mechanistic role played by low levels of
protein C in the balance of coagulation in patients with chronic liver
disease,18,19
vitamin K antagonists are perhaps not the ideal drugs. Protein C is a
vitamin K–dependent protein, and treatment with vitamin K antagonists
might therefore further reduce levels of this naturally occurring
anticoagulant in patients with end-stage liver disease, increasing the
risk of thrombosis.

The newer direct thrombin inhibitors and inhibitors of activated factor X56
(e.g., dabigatran, rivaroxaban, and apixaban) may be attractive
alternatives to vitamin K antagonists because they do not reduce protein
C levels. Moreover, they do not require regular laboratory monitoring
to adjust the dosage, whereas vitamin K antagonists require monitoring
with the use of the INR, the validity of which has been questioned in
patients with chronic liver disease.57
Other potential advantages of these new drugs over low-molecular-weight
heparin are their oral route of administration and their mechanism of
action, which is independent of antithrombin (low in these patients).
However, specially designed clinical trials are needed because patients
with chronic liver disease are usually excluded from the randomized
clinical trials of these drugs.58-60

Another
consequence of procoagulant imbalance in chronic liver disease pertains
to liver fibrosis and its progression. Two hypotheses are currently
considered for the pathogenesis of this condition. Both involve
coagulation, and they might be synergistic. One hypothesis centers on
the role of microemboli. Obliterative lesions in the portal and hepatic
veins frequently occur in patients with cirrhosis, owing to the
formation of microthrombi that lead to tissue ischemia, cell death, and
fibrosis through parenchymal extinction.61

Another
hypothesis suggests that coagulation activation within the liver's
vascular system may play a role in the development and progression of
the fibrotic process. Thrombin, besides being a potent procoagulant, has
many cellular effects that are mediated by a family of widely expressed
G-protein–coupled receptors called protease-activated receptors (PARs).62
Thrombin signaling through PARs expressed on hepatic stellate cells,
which are responsible for tissue repair, might therefore play a crucial
role in the mechanisms and progression of fibrosis.63
The degree of thrombin-receptor expression is associated with the
severity of liver disease, and it has also been observed that humans64 and mice65
with hypercoagulability due to a gain-of-function mutation in the
factor V gene (factor V Leiden) have an accelerated progression of liver
fibrosis. PAR1 antagonists can provide protection against experimental
liver fibrosis in rodents,63 and anticoagulant drugs slow fibrosis progression in mice.65
Furthermore, low-molecular-weight heparin prevents hepatic fibrogenesis
caused by the injection of carbon tetrachloride in rodents.66 These observations are consistent with the hypothesis that thrombin generation and fibrosis are directly associated.65
Accordingly, a controlled, randomized clinical trial is being carried
out to investigate whether vitamin K antagonists can influence the
progression of fibrosis in patients with hepatitis C (ClinicalTrials.gov
number, NCT00180674).

Undoubtedly, patients
with end-stage liver disease have prominent bleeding symptoms,
particularly in the gastrointestinal tract. Yet evaluation of this
bleeding tendency solely on the basis of abnormal levels of the
conventional coagulation biomarkers should be reconsidered. When
patients are assessed by means of global tests such as the
thrombin-generation test, the results do not show hypocoagulability.18,19,45-47
Thus, the main culprits for the bleeding tendency observed in patients
with end-stage liver disease should be sought among underlying
conditions that favor hemorrhage, such as portal hypertension,
endothelial dysfunction, bacterial infection, and renal failure20,28-31 (Table 2).

On
the other hand, the restored balance of hemostasis afforded by the
concomitant reduction of procoagulant and anticoagulant factors,
together with increased levels of factor VIII (Table 1),
might explain why patients with chronic liver disease are not protected
from arterial and venous thrombosis. This apparent clinical paradox may
be explained by the findings that these patients have a procoagulant
imbalance in vitro owing to resistance to thrombomodulin18,19,45-47
and that their thrombocytopenia is compensated for by increased plasma
levels of the adhesive protein von Willebrand factor. Another dogma is
being challenged by the finding that platelet activation plays a crucial
role in the immune-mediated progression of liver disease in an animal
model of viral hepatitis.67

In
conclusion, the reassessment of hemostasis in patients with chronic
liver disease challenges the dogma that the major coagulopathy in these
patients leads consistently to bleeding. Other changes that accompany
chronic liver disease may restore the balance of anticoagulant and
procoagulant effects (Figure 2C).
In certain circumstances, the risk of thrombotic events may be greater
than the risk of hemorrhage. We speculate that drugs that are often
regarded as contraindicated in patients with chronic liver disease may
instead prove beneficial and should be tested in appropriate clinical
trials.