Dingeman C. Rijken, Erasmus University Medical Center Rotterdam, Department of Hematology, Room Ee 1393, Dr. Molewaterplein 50, 3015 GE Rotterdam, the Netherlands. Tel.: + 31 10 70 44 723; fax: + 31 10 70 44 745. E-mail: firstname.lastname@example.org
Summary. Background and objectives: It has been known for a long time that cirrhosis is associated with hyperfibrinolysis, which might contribute to an increased risk and severity of bleeding. However, recent papers have questioned the presence of a hyperfibrinolytic state in cirrhotic patients and postulated a rebalanced system owing to concomitant changes in both pro- and anti-fibrinolytic factors. Therefore we re-investigated the fibrinolytic state of cirrhotic patients using two different overall tests including a recently developed test for global fibrinolytic capacity (GFC) using whole blood.
Patients and methods: Blood was collected from 30 healthy controls and 75 patients with cirrhosis of varying severity (34 Child–Pugh A, 28 Child–Pugh B and 13 Child–Pugh C). The plasma clot lysis time (CLT), which is inversely correlated with fibrinolysis, was determined as well as the GFC.
Results: The mean CLT was 74.5 min in the controls and decreased significantly to 66.9 min in Child–Pugh class A patients, 59.3 min in class B patients and 61.0 min in class C patients, and hyperfibrinolysis existed in 40% of the patients. The median GFC was 1.7 μg mL−1 in the controls and increased significantly to 4.0 μg mL−1 in Child–Pugh class A patients, 11.1 μg mL−1 in class B patients and 22.5 μg mL−1 in class C patients, and hyperfibrinolysis existed in 43% of the patients. Taken together, 60% of the patients showed hyperfibrinolysis in at least one of the two global assays.
Conclusion: A rebalanced fibrinolytic system may occur, but hyperfibrinolysis is found in the majority of patients with cirrhosis.
Liver cirrhosis is associated with an increased bleeding tendency. One of the underlying mechanisms is an impaired hemostatic system, because the liver is responsible for the synthesis as well as for the clearance of several hemostatic factors. An important component of hemostasis is the fibrinolytic system . It has been known for a long time that this system may be disturbed in cirrhosis leading to the enhancement of fibrinolysis  by a shift in balance between pro-fibrinolytic and anti-fibrinolytic factors . This shift in balance is observed at the level of plasminogen activation, for example the balance between tissue-type plasminogen activator (tPA) and plasminogen activator inhibitor-1 (PAI-1), as well as at the level of plasmin, for example the balance between plasmin(ogen) and α2-antiplasmin. A more recently discovered major regulator of the fibrinolytic system is thrombin-activatable fibrinolysis inhibitor (TAFI) which is also synthesized in the liver . Lisman et al.  found that this inhibitor is significantly decreased in cirrhosis. However, global fibrinolysis, as observed in a plasma clot lysis system, appeared not to be increased in their study. This was explained by strongly reduced antithrombin levels associated with a relatively high thrombin generation in the assay and a concomitant higher proportion of TAFI activation, which might compensate for the reduced TAFI levels. Because the plasma clot lysis assay is considered to be superior to the older fibrinolysis assays used for the detection of accelerated fibrinolysis, the existence of hyperfibrinolysis in cirrhosis was questioned  and it was postulated that the fibrinolytic system is rebalanced in patients with cirrhosis by concomitant changes in both pro- and antifibrinolytic factors . Colucci et al.  confirmed the reduced TAFI levels in cirrhotic patients, but these authors did find increased fibrinolysis in a similar plasma clot lysis assay. The background to this controversy is still unclear. New global tests are needed to establish whether or not a true hyperfibrinolytic state exists in cirrhosis. [8,9].
The status of the fibrinolytic system cannot be satisfactorily assessed by measuring levels of individual components, because the system, as indicated above, strongly depends on the balance of various pro- and anti-fibrinolytic factors. The system also depends on the balance of various pro- and anti-coagulant factors that regulate amongst others the activation of TAFI. We recently developed a new test for the global fibrinolytic capacity in undiluted whole blood . In this test, in contrast to the plasma clot lysis assay, no exogenous tPA is added, and platelets as well as other blood cells are present to mimic better the in vivo situation in circulating blood. To address the question of whether hyperfibrinolysis is present in cirrhotic patients, we used this test, as well as the plasma clot lysis assay, to investigate the fibrinolytic system in 75 patients with cirrhosis of varying severity and in 30 healthy controls. This study was performed within the framework of our studies on the hemostatic and thrombotic complications of liver diseases [11–13].
Patients and methods
We studied 75 consecutive ambulatory and hospitalized patients with different types of etiology of cirrhosis and 30 healthy controls. The diagnosis of cirrhosis was histologically proven or clinically proven by a radiological examination of the liver. Patients were classified into three groups of increasing severity (A–C) according to the Child–Pugh’s score . In primary sclerosing cholangitis (PSC) and primary biliary cirrhosis (PBC), the bilirubin references were changed to reflect the fact that these diseases feature high conjugated bilirubin levels. The upper limit for 1 point was 70 μm and the upper limit for 2 points was 170 μm in the calculation of the Child–Pugh scores of these patients . The severity of the disease was also assessed by calculating the model for end-stage liver disease (MELD) score according to the formula 9.57 × ln(creatinine) + 3.78 × ln(bilirubin) + 11.2 × ln International Normalized Ratio (INR) + 6.43 .
In order to further characterise the patients with respect to bleeding, esophageal varices were classified retrospectively as small sized (n = 12 patients), moderate sized (n = 9), large sized (n = 11) and varices with red signs (n = 1). A number of patients (n = 18) had no varices and information of 24 patients was not available. Four patients developed gastrointestinal bleeding, as defined by Violi et al. , in a follow-up period of 1 year.
Exclusion criteria were malignancy, anticoagulant treatment or the use of antiplatelet drugs, pre-existing bleeding disorders, a history of venous thrombosis, pregnancy, recent platelet substitution or plasma substitution (< 1 week prior to inclusion), or recent bleeding episodes (< 1 week prior to inclusion). The control group consisted of 30 healthy subjects with no clinical evidence of any history of liver disease. They were recruited from friends, neighbors or partners of the patients.
This study was performed in accordance with the guidelines for Good Clinical Practice/ICH, the principles of the Declaration of Helsinki 1964, as modified by the 52nd WMA General Assembly, Edinburgh, Scotland, October 2000, and the local national laws governing the conduct of clinical research studies. The Medical Ethical Committee of the Erasmus MC approved the study protocol. All participants gave written informed consent.
Blood was collected by venipuncture after minimal stasis in the antecubital vein in standard citrated tubes (Vacutainer; Becton Dickinson, Franklin Lakes, NJ, USA) or Stabilyte tubes (0.45 m citrate pH 4.3; Trinity Biotech, Bray, Ireland) and centrifuged at 2500 g for 15 min at 4 °C. Plasma was collected and stored in aliquots at −80 °C. Blood for the global fibrinolytic capacity test  was collected in 3-mL Vacutainer tubes containing 1.4 NIH units of thrombin (Stat chemistry tubes; Becton Dickinson) supplemented with or without 60 μL 10 000 KIU mL−1 aprotinin (Bayer, Leverkusen, Germany) before blood collection.
Plasma clot lysis assay
The plasma clot lysis assay was performed essentially as described previously . Citrated plasma was diluted 1.7 times in assay buffer (25 mm Hepes, 137 mm NaCl, 3.5 mm KCl, 1% [w/v] bovine serum albumin, pH 7.4). The diluted plasma (85 μL) was added to wells of a microtiter plate containing 15 μL of a reaction mixture. The reaction mixture contained the following components with final concentrations in the clot: tissue factor (Innovin, 1000 times diluted; Dade Behring, Liederbach, Germany), CaCl2 (17 mm), tPA (25 ng mL−1, Actilyse; Boehringer, Ingelheim, Germany) and phospholipid vesicles (10 μm). After mixing the diluted plasma with the reaction mixture, each well was covered with 50 μL of paraffin oil and the microtiter plate was placed into the preheated chamber of the microtiter plate reader. The optical density at 405 nm was measured every minute for 300 min at 37 °C. The clot lysis time (CLT) was the time from the midpoint of clear to maximum turbidity, which represents clot formation, to the midpoint of maximum turbidity to clear transition, which represents clot lysis. The assay was performed in duplicate. The results of three patients (one of each Child–Pugh class) were missing.
Global fibrinolytic capacity test
The test of the global fibrinolytic capacity (GFC) was performed essentially as described previously . The thrombin-containing blood collection tubes were immediately incubated for 3 h at 37 °C. After this incubation period the clots were released from the tube wall with a plastic spatula. The tubes were centrifuged at 4 °C at 2500 g for 15 min. An aliquot of 500 μL serum was collected and mixed with 10 μL of aprotinin (10 000 KIU mL−1) to block plasmin activity. Serum samples were stored at −80 °C. The fibrin degradation products (FnDPs) were measured using a latex agglutination assay (Auto Dimer, Biopool) on a Sysmex CA-1500 Analyzer. The GFC was calculated by subtracting the FnDP concentration in the thrombin-containing Vacutainer tubes with aprotinin from the FnDP concentration in the thrombin-containing Vacutainer tubes without aprotinin. The assay was performed in duplicate and the results were expressed in μg mL−1.
Other hemostasis assays
The fibrinogen concentration was determined with a clotting rate assay according to Clauss  using thrombin from Dade Behring (Dade Thrombin Reagent). The plasminogen concentration was determined by a chromogenic substrate method using streptokinase (Kabikinase; Pharmacia, Woerden, the Netherlands) and the substrate S-2251 (Chromogenix, Milano, Italy). The α2-antiplasmin concentration was determined by a chromogenic substrate method using the COAMATIC Plasmin Inhibitor kit (Chromogenix). The antithrombin concentration was determined by a chromogenic substrate method using the COAMATIC Antithrombin kit (Chromogenix). The results of the last three assays were expressed in U mL−1 (pooled normal plasma contained 1 U ml−1). Levels of tPA antigen in plasma were assayed by a slightly modified commercially available enzyme-linked immunosorbent assay (t-PA Antigen Elisa Reagent Kit, Technoclone, Vienna, Austria). The functional assay of TAFI was performed using a clot lysis assay as described previously and expressed in minutes . The activity levels of t-PA and PAI-1 were determined in Stabilyte plasma using the bioimmunoassays Chromolize t-PA activity and Chromolize PAI-1 activity (Trinity Biotech). All other assays were performed in citrated plasma.
Statistical analysis was performed using the graphpad prism version 4 software package (GraphPad Software, San Diego, CA, USA). Parametric analysis was performed using a one-way anova with Bonferroni’s post-test and non-parametric analysis using the Kruskall–Wallis anova test with Dunn’s post-test. Correlations were determined by calculation using Spearman’s rho. P-values of < 0.05 were considered statistically significant.
The existence of hyperfibrinolysis was studied in a group of 75 cirrhotic patients and 30 healthy controls. As shown in Table 1, the patients had different types of etiology and were divided into Child–Pugh class A (N = 34), class B (N = 28) and class C (N = 13). The MELD score increased in accordance with the Child–Pugh classification. Table 2 summarizes the results of a series of hemostatic factor assays performed in plasma samples. In agreement with the literature, levels of fibrinogen, antithrombin and plasminogen decreased in cirrhosis, whereas tPA activity and tPA antigen increased. In addition, levels of α2-antiplasmin activity and TAFI activity decreased, whereas PAI-1 activity did not change significantly. All changes correlated with the severity of the disease.
Table 1. Demographic characteristics of the study population. The cirrhotic patients are grouped according to Child–Pugh class A, B or C
PSC, primary sclerosing cholangitis; PBC, primary biliary cirrhosis; MELD, model for end-stage liver disease.
Age in years (mean ± SD)
50 ± 13
50 ± 13
51 ± 11
55 ± 10
MELD score (median and range)
Table 2. Hemostatic parameters of the study population. The cirrhotic patients are grouped according to Child–Pugh class A, B or C
Controls (N = 30)
Class A (N = 34)
Class B (N = 28)
Class C (N = 13)
Data are presented as mean ± SD (normal distributions) or median with 25th and 75th percentile (non-normal distributions). The P-values were calculated using one-way anova or by Kruskal-Wallis test. The P-values for the comparisons with controls were calculated using Bonferroni’s multiple comparison test or with Dunn’s multiple comparison test, respectively. *P < 0.05; **P < 0.01; ***P < 0.001.
Table 2 and Fig. 1 show the results of the plasma clot lysis assay in our study population. The clot lysis time (CLT) decreases as fibrinolysis increases. The CLT (mean ± SD) was 74.5 ± 9.6 min in the controls and decreased significantly to 66.9 ± 8.3 min in Child–Pugh class A patients, 59.3 ± 10.6 min in class B patients and 61.0 ± 14.9 min in class C patients. The CLT also correlated significantly with the MELD score of the patients (ρ−0.40, P = 0.001). Hyperfibrinolysis, defined as a fibrinolytic state with a CLT below the lowest CLT value of the 30 controls (i.e. below 61.0 min) existed in 21% of class A patients, in 52% of class B patients and in 67% of class C patients (Fig. 1). The present results showed that hyperfibrinolysis, as measured with a plasma clot lysis assay, existed in 40% of patients with cirrhosis.
The existence of hyperfibrinolysis in cirrhosis was also studied using a new global fibrinolysis assay in whole blood (Table 2 and Fig. 2). The global fibrinolytic capacity (GFC) increases with intensity of fibrinolysis. The GFC (median with 25th and 75th percentile) was 1.7 (0.8–3.5) μg mL−1 in the controls and increased gradually to 4.0 (1.6–13.0) μg mL−1 in Child–Pugh class A patients, 11.1 (2.3–31.9) μg mL−1 in class B patients and 22.5 (3.1–89.6) μg mL−1 in class C patients. The GFC also correlated significantly with the MELD score of the patients (ρ 0.41, P < 0.001). Hyperfibrinolysis, defined as a fibrinolytic state with a GFC above the highest GFC value of the 30 controls (i.e. above 11.4 μg mL−1) existed in 29% of class A patients, in 50% of class B patients and in 62% of class C patients (Fig. 2). These results show that hyperfibrinolysis, as measured with the new global fibrinolysis assay, existed in 43% of patients with cirrhosis.
Male and female controls did not significantly differ in CLT and GFC (P = 0.35 and 0.14, respectively, Mann–Whitney test). Fibrinolytic differences between patients and controls could therefore not be ascribed to the overrepresentation of male subjects in the patient groups (Table 1).
Table 3 show correlations between the global fibrinolytic capacity (GFC) values and other hemostatic parameters in plasma of both controls and cirrhotic patients. In controls, the GFC was positively correlated with tPA activity, which is in line with our previous work . The correlation coefficients were negative for the fibrinolysis inhibitors a2-antiplasmin activity, PAI-1 activity and TAFI activity, but these correlations were not statistically significant. The correlation between the GFC and tPA activity in patients was even stronger than in the controls suggesting a causal role of increased tPA activity in hyperfibrinolysis. The GFC in patients was also strongly correlated, in a negative manner, with PAI-1 activity. This relationship might also be causal, because PAI-1 was not dependent on cirrhosis (Table 2).
Table 3. Correlations between the global fibrinolytic capacity (GFC) values and other hemostatic parameters in controls and cirrhotic patients. Spearman’s rho is given along with the P-value
The results of the two global fibrinolysis assays did not correlate in the controls (Table 3). They did, however, moderately correlate in the patients (r = −0.46, P < 0.001). The scatter diagrams in Fig. 3 further demonstrate this correlation. This figure also illustrates that in the patient group 29/72 (40%) showed no hyperfibrinolysis in any of the two global assays, 11/72 (15%) showed hyperfibrinolysis only in the plasma clot lysis assay, 14/72 (19%) showed hyperfibrinolysis only in the global fibrinolysis assay in whole blood and 18/72 (25%) showed hyperfibrinolysis in both assays. This implies that 43/72 (60%) of the patients exhibited hyperfibrinolysis in at least one of the two global assays. The group of 29 patients not showing hyperfibrinolysis in any of the global assays, even included three severely ill patients with Child–Pugh C. The absence of hyperfibrinolysis in two of them (etiology: cryptogenic and hepatitis C, respectively), could be explained by their strongly increased PAI-1 levels (36 and 46 IU mL−1, respectively) and correspondingly decreased tPA activities (0.19 and 0.25 IU mL−1, respectively). As mentioned previously, high PAI-1 levels are not typical of cirrhosis, but they may occasionally occur (Table 2).
Only a limited number of studies have so far been performed on the relationship between abnormalities in fibrinolysis and bleeding in cirrhosis. Violi et al.  observed that high D-dimer levels and high t-PA activity were associated with an increased risk of gastro-intestinal bleeding. However the association between hyperfibrinolysis and bleeding remains controversial. There is, in general, a relatively poor correlation between bleeding in cirrhosis and the outcome of conventional hemostasis tests [8,20]. One reason is that these tests, including the commonly used fibrinolysis tests, might not sufficiently reflect the complex balance between hemostatic factors that exists in vivo [8,21,22]. Although hyperfibrinolysis in cirrhosis seems to be a well-established phenomenon, serious concern exists about the reliability of the fibrinolysis tests used . In this study we confirmed, however, the existence of hyperfibrinolysis in a significant proportion (60%) of patients with cirrhosis using both the plasma clot lysis assay and a new global fibrinolysis assay in undiluted whole blood .
In order to obtain a deeper insight into the hemostatic characteristics of the patients in our study and to compare them with other populations described in the literature, we tested the most relevant fibrinolytic factors in plasma. In agreement with the literature, both pro-fibrinolytic (plasminogen) and anti-fibrinolytic (α2-antiplasmin, TAFI) factors synthesized by the liver, strongly decreased in cirrhosis. The activity of PAI-1, synthesized by a variety of hepatic and non-hepatic cell types, did not change, whereas tPA activity and tPA antigen, primarily synthesized by the vascular endothelium, significantly increased. The increase in tPA levels is usually ascribed to decreased liver clearance . Clearance of tPA involves both liver endothelial cells and liver parenchymal cells after recognition by the mannose receptor and low-density lipoprotein receptor-related protein, respectively [24–26]. Urokinase-type plasminogen activator, not specifically measured in this study but potentially involved in the global tests, might also be increased in cirrhosis .
To measure the overall effect of the changes in plasma levels of known and possibly unknown fibrinolysis factors we applied the global plasma clot lysis assay. In contrast to Lisman et al. , but in agreement with Colucci et al. , a significant decrease in clot lysis time was observed pointing to an increase in plasma fibrinolytic potential in cirrhotic patients. The antithrombin levels were significantly reduced in our cirrhotic patients, which might have resulted in more thrombin and more TAFI activation during the assay. However, this anti-fibrinolytic effect was apparently overruled by the pro-fibrinolytic effects in cirrhosis. In addition, more thrombin might also promote fibrinolysis by the so-called coagulation-associated enhancement of fibrinolysis based on thrombin-dependent inactivation of PAI-1 . As Colucci et al.  also noted, the possibility cannot be excluded that minor methodological differences in the plasma clot lysis assay and/or differences in the selection of patients may account for the discrepancy with the results of Lisman et al. . Our results are in line with those obtained in a small group of 14 cirrhotic patients tested using a new commercially available global fibrinolytic capacity test in citrated plasma .
One disadvantage of the plasma clot lysis assay is that a relatively high concentration (25 ng mL−1) of exogenous tPA is added to induce clot lysis. This makes the assay less sensitive to the balance of endogenous levels of tPA and PAI-1 in the circulation. We agree with Lisman et al.  that the plasma clot lysis assay might reflect local conditions where vessel wall injury leads to massive release of tPA from the endothelial cells. However, it is presently unclear whether systemic or local fibrinolytic conditions prevail during the regulation of the hemostatic system in cirrhosis. Because systemic tPA levels are strongly increased in the cirrhotic patients, it seems plausible that they do affect local hemostasis and that a global assay sensitive to endogenous levels of tPA is required to reflect local conditions. We therefore applied the global fibrinolytic capacity test in whole blood, which does depend fully on the endogenous levels of tPA and PAI-1 and should be considered as a different test, complementary to the plasma clot lysis assay. Using this assay, 32/72 patients showed hyperfibrinolysis; 18 of these patients also showed hyperfibrinolysis in the plasma clot lysis assay and 14 patients did not. The latter patients illustrate that the global fibrinolytic capacity test in whole blood is able to identify a subgroup of cirrhotic patients with hyperfibrinolysis that is not identified by the plasma clot lysis assay.
In conclusion, the hypothesis of a rebalanced fibrinolytic system in cirrhosis with no increased fibrinolysis  seems to be true in at most 40% of our patients. Hyperfibrinolysis in at least one of the two global fibrinolysis assays is detectable in the majority of the patients (60%). These findings remain an in vitro observation without direct implications for the clinical practise. However, they should be taken into account in current discussions on the coagulopathy of chronic liver disease [9,30]. In addition, the new global fibrinolytic tests may make it possible in future clinical studies to investigate whether hyperfibrinolysis is associated with bleeding problems in cirrhosis. The number of bleedings documented in the present study (n = 4) was too small for such an investigation. Larger studies are required to establish if and how these fibrinolysis assays could be used in the clinic to estimate the risk of bleeding.
We wish to thank R. van de Graaf and S. Petronia for their assistance in the collection of clinical data.
Disclosure of Conflict of Interest
The authors state that they have no conflict of interest.