SEARCH

SEARCH BY CITATION

Keywords:

  • acetaminophen;
  • acute liver failure;
  • acute liver injury;
  • coagulation;
  • fibrinolysis;
  • thrombin

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Summary.  Background:  It has been well established that hemostatic potential in patients with chronic liver disease is in a rebalanced status due to a concomitant decrease in pro- and antihemostatic drivers. The hemostatic changes in patients with acute liver injury/failure (ALI/ALF) are similar but not identical to the changes in patients with chronic liver disease and have not been studied in great detail.

Objective:  To assess thrombin generation and fibrinolytic potential in patients with ALI/ALF.

Methods:  We performed thrombin generation tests and clot lysis assays in platelet-poor plasma from 50 patients with ALI/ALF. Results were compared with values obtained in plasma from 40 healthy volunteers.

Results and conclusion:  The thrombin generation capacity of plasma from patients with ALI/ALF sampled on the day of admission to hospital was indistinguishable from that of healthy controls, provided thrombomodulin was added to the test mixture. Fibrinolytic capacity was profoundly impaired in patients with ALI/ALF on admission (no lysis in 73.5% of patients, compared with 2.5% of the healthy controls), which was associated with decreased levels of the plasminogen and increased levels of plasminogen activator inhibitor type 1. The intact thrombin generating capacity and the hypofibrinolytic status persisted during the first week of admission. Patients with ALI/ALF have a normal thrombin generating capacity and a decreased capacity to remove fibrin clots. These results contrast with routine laboratory tests such as the PT/INR, which are by definition prolonged in patients with ALI/ALF and suggest a bleeding tendency.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

The liver plays a key role in the hemostatic system as it synthesizes most of the proteins involved in coagulation and fibrinolysis. Consequently, patients with liver disease may have profound changes in their hemostatic system [1]. It has now been well established that the hemostatic alterations in patients with chronic liver disease result in a ′rebalanced′ hemostatic status due to a concomitant decrease in both pro- and anticoagulant mechanisms [2]. Importantly, the anticoagulant systems are not reflected in the outcome of routine laboratory tests such as the prothrombin time (PT) or activated partial thromboplastin time (APTT). Laboratory evidence for a rebalanced coagulation system in chronic liver disease has been provided by studies in which thrombin generation was tested in the presence of activators of the anticoagulant protein C system [3,4]. Furthermore, global laboratory tests of the fibrinolytic system also have suggested rebalanced fibrinolysis in chronic liver disease [5], although there are also reports of hyperfibrinolysis in chronic liver disease [6].

The hemostatic alterations in chronic liver disease have been much more extensively studied as compared with the hemostatic changes in acute liver injury (ALI) or acute liver failure (ALF). Acute liver injury and acute liver failure (ALI/ALF) are syndromes of diverse etiology in which patients without previously recognized liver disease sustain a liver injury that results in rapid loss of hepatic function [7]. ALI/ALF is defined by a ‘coagulopathy’, as an international normalized ratio (INR) of > 1.5 is required for the diagnosis. ALF represents a more severe liver injury than ALI, resulting in hepatic encephalopathy. As patients with ALI/ALF have by definition an elevated INR, it is evident that hemostatic alterations affect all patients with acute liver failure [8].

The hemostatic changes in acute liver failure are similar to those observed in chronic liver disease [8,9]. There are, however, notable differences. Patients with acute liver failure have a more severe decrease in levels of pro- and anticoagulant proteins as compared with patients with chronic liver disease. This is reflected by the substantial proportion of patients with acute liver failure with substantially elevated INRs. One study showed that 14% of patients with acute liver failure had an INR between 5 and 10 upon admission, and 5% of patients even had an INR > 10 [10]. This degree of coagulopathy is exceptional in patients with chronic liver disease.

Recently, we have provided evidence for a rebalanced hemostatic status in patients with ALI/ALF [11]. Using kaolin-activated whole blood thromboelastography, we demonstrated that two-thirds of patients with ALI/ALF had thromboelastographic traces that were within the normal range, whereas 8% of patients even had evidence of hypercoagulability, despite severely reduced levels of coagulation factor levels. Surprisingly, we observed normal clot formation despite the absence of an activator of the anticoagulant protein C system in the test. Interestingly, in this large series of patients, the number of thrombotic events was higher as compared with the number of bleeding events, suggesting that these patients do not have a bleeding tendency as suggested by abnormal routine laboratory tests (PT/INR and APTT).

Given the limitations of kaolin-induced thromboelastography and the absence of a precise molecular mechanism behind the normal thromboelastographic tracings in patients with ALI/ALF despite abnormal levels of coagulation proteins and profoundly abnormal routine tests of coagulation (PT/INR, APTT and platelet count), we set out to perform a more in-depth analysis of the hemostatic system in these patients. Specifically, we performed overall tests of coagulation and fibrinolysis, and in addition analyzed individual levels of crucial fibrinolytic proteins in samples taken from a relatively large cohort of patients with ALI/ALF on admission to hospital. Finally, we compared hemostatic potential in samples taken on the day of admission with samples taken at later time-points during the first week of hospitalization.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patients

The study population consisted of 50 patients admitted consecutively for acute liver injury/acute liver failure (ALI/ALF) to Virginia Commonwealth University Medical Center between March 2009 and May 2011. Forty of the 50 patients have previously been described in detail [11]. Informed consent was obtained from either the patient or their next-of-kin, depending on the patient’s level of hepatic encephalopathy, and the study was approved by the local Institutional Review Board. Inclusion criteria included: (i) an INR of ≥ 1.5; (ii) absence of a previous history of liver disease; and (iii) illness ≤ 26 weeks duration. Patients with ALF were also required to have altered mental status ascribed to their liver injury (hepatic encephalopathy). Patients who received procoagulant treatments other than vitamin K prior to enrollment were excluded. Individual plasma samples from 40 healthy volunteers from our laboratory were used to establish reference values for the various tests performed in this study. For patients, blood samples were collected using an 18G butterfly needle or were drawn directly from an arterial line into a 5-mL citrated Vacutainer® , and a 21G needle was used for the controls. Samples were centrifuged within 2 h of withdrawal at 1700×g at room temperature for 15 min, followed by a centrifugation step of 15 min at 3000×g (controls) or just once at 3000×g at room temperature for 15 min (patients). Plasma was immediately placed at −80 °C and stored until assayed. Definitions of outcomes and complications (bleeding and thrombosis) have been provided previously [11].

Thrombin generation testing

The Calibrated Automated Thrombogram was used to determine the generation of thrombin in clotting plasma as described previously [12]. Coagulation was triggered by recalcification in the presence of 1 pm recombinant human tissue factor (Innovin; Siemens Healthcare Diagnostics, Marburg, Germany; 1:10 000 dilution), 4 μm phospholipids and 417 μm fluorogenic substrate Z-Gly-Gly-Arg-AMC (Bachem, Bubendorf, Switzerland) in the presence or absence of rabbit lung thrombomodulin (2 nm; American Diagnostica, Greenwich, CT, USA). The assay was run at 37 °C, and the plate was incubated in the reader for 10 min prior to addition of the fluorogenic substrate and calcium. The endogenous thrombin potential (ETP), peak, time-to-peak, lag time and velocity index were calculated using the Thrombinoscope software (version 4.0, Thrombinoscope BV, Maastricht, the Netherlands). The peak and velocity index were normalized to pooled normal plasma from > 200 healthy volunteers as described previously [12]. A normalized ratio of the ETP in the presence or absence of thrombomodulin (TM-SR) was determined by the following equation: inline image. A TM-SR > 1 reflects a decreased anticoagulant response to thrombomodulin in comparison to pooled normal plasma.

Overall fibrinolytic potential

Overall plasma fibrinolytic potential was determined using an in-house assay as previously described [13]. Briefly, 50 μL plasma was mixed with 50 μL of a solution containing phospholipid vesicles (40%l-α-dioleoylphosphatidylcholine, 20%l-α-dioleoylphosphatidylserine and 40%l-α-dioleoylphosphatidylethanolamine, final concentration 10 μm), t-PA (final concentration 56 ng mL−1), recombinant human tissue factor (Innovin; Siemens Healthcare Diagnostics, 10 pm, final dilution 1/1000) and CaCl2 (final concentration 17 mm) diluted in HEPES buffer (25 mm HEPES [N-2-hydroxytethylpiperazine-N′-2-ethanesulfonic acid], 137 mm NaCl, 3.5 mm KCl, 3 mm CaCl2, 0.1% bovine serum albumin, pH 7.4). The plate was incubated at 37°C in a Spectramax 340 kinetic microplate reader (Molecular Devices Corporation, Menlo Park, CA, USA), and the optical density at 405 nm was monitored every 20 s, resulting in a clot-lysis turbidity profile. The clot lysis time (CLT) was derived from this clot-lysis profile and defined as the time from the midpoint of the clear to maximum turbid transition, representing clot formation, to the midpoint of the maximum turbid to clear transition, representing the lysis of the clot.

Assays of individual proteins

Free tissue factor pathway inhibitor (TFPI) antigen levels were determined using an enzyme-linked immunosorbent assay (ELISA) according to the instructions of the manufacturer (TFPI Asserachrom; Diagnostica Stago, Asnieres sur Seine, France). Plasminogen and antiplasmin activity levels were determined by chromogenic kits from Siemens Healthcare Diagnostics and Chromogenix (Molndal, Sweden), respectively. Antigen levels of tissue-type plasminogen activator (tPA) were determined by ELISA kits according to the instructions of the manufacturer (American Diagnostica, Stamford, CT, USA), as were levels of plasminogen activator inhibitor-1 (PAI-1), which were determined using an ELISA kit purchased from Hyphen Biomed (Neuville-Sur-Oise, France). Plasma levels of thrombin activatable fibrinolysis inhibitor (TAFI) were determined using an in-house ELISA as described [14].

Statistical analyses

All calculations were performed using the GraphPad InStat software package (GraphPad, San Diego, CA, USA). P values < 0.05 were considered statistically significant. Values derived from the thrombin generation curves were compared using one-way analysis of variance (anova), using the Tukey or Kruskal–Wallis post-test as appropriate, with the exception of the TM-SR values, which were compared using the Mann–Whitney U-test. CLT values were dichotomized (lysis within 3 h vs. no lysis within 3 h) and analyzed using the Fishers exact test. Levels of individual proteins were analyzed using t-test or Mann–Whitney U-test as appropriate.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patient characteristics and clinical course

Baseline demographics, vital signs and laboratory test results on admission and outcomes of the study population are presented in Table 1. The mean age of study participants was 43 years, 64% were female, 58% Caucasian, and the mean BMI was 28 kg m−2. The etiologies of ALI/ALF were acetaminophen (APAP) in 50%, hepatitis B virus (HBV) in 14%, idiosyncratic drug reactions in 12%, autoimmune hepatitis (AIH) in 10%, indeterminate in 6%, and heat stroke, Amanita mushroom poisoning, malignant infiltration and hepatic ischemia in 2% each. Thirty-nine patients (78%) had hepatic encephalopathy on admission, and 24 (48%) progressed to high-grade encephalopathy (grade 3 or 4) at some point over the first week of admission. All patients with APAP-induced ALI/ALF were treated with N-acetylcysteine (NAC), as were 19 patients with non-APAP-induced liver injury. The systemic inflammatory response syndrome (SIRS, defined as the presence of 2–4 SIRS components) was present in 28 (56%) patients on admission to the study. Admission laboratory test results included a median creatinine and bilirubin of 1.0 and 5.0 mg dL−1, respectively, and mean pH of 7.35, bicarbonate 19.7 mg dL−1, lactate 5.9 mg dL−1, phosphate 3.3 mg dL−1 and INR 3.4 (assayed in this study using the Innovin reagent, which has an International Sensitivity Index of 0.9).

Table 1.   Clinical characteristics and outcome of study population. Data are expressed as number (N) and percentage, mean [SD] or median [range]
CharacteristicN (%) or mean [SD] or median [range]
Age (years)43.1 [13.5]
Gender (female)32 (64)
Race (Caucasian)29 (58)
BMI (kg m−2)28.2 [6.8]
Etiology
APAP25 (50)
HBV7 (14)
Idiosyncratic drug6 (12)
AIH5 (10)
Indeterminate3 (6)
Heat stroke1 (2)
Malignancy1 (2)
Amanita1 (2)
Ischemia1 (2)
Hepatic encephalopathy
Admission39 (78)
Maximal grade 3 or 424 (48)
Mean arterial pressure on admission (mmHg)84.8 [14.5]
NAC treatment44 (88)
Admission SIRS
WBC (× 109)10.7 [5.7]
Pulse (beats min−1)98.8 [22.4]
Respirations (per min)20.3 [6.3]
Temperature (°C)36.7 [1.0]
SIRS (2–4 components present)28 (56)
Admission laboratory tests
AST (IU L−1)4990.3 [4377.8]
ALT (IU L−1)3578.6 [2765.1]
Creatinine (mg dL−1)1.0 [0.4–7.5]
Total bilirubin (mg dL−1)5.0 [0.3–44.2]
INR3.4 [1.8]
aPTT (s)47.4 [14.7]
Platelet count (g L−1)193 [99]
Hemoglobin (g L−1)12.0 [2.2]
Lactate (mg dL−1)5.9 [5.6]
Venous ammonia (μm)80.3 [45.3]
Phosphate (mg dL−1)3.3 [2.2]
pH7.35 [0.13]
HCO3 (mg dL−1)19.7 [7.7]
Complications
Renal failure (requiring renal replacement therapy)18 (36)
Infection13 (26)
Intracranial hypertension3 (6)
Thrombosis9 (18)
Bleeding9 (18)
Outcome
Transplant-free survival28 (56)
Liver transplantation7 (14)
Death15 (30)

Complications of the study population included renal failure requiring renal replacement therapy in 18 (36%), infection in 13 (26%), intracranial hypertension in 3 (6%) and thrombosis or bleeding in 9 (18%) patients. Fifteen (30%) patients died, 7 (14%) underwent liver transplantation and 28 (56%) recovered without liver transplantation.

Intact thrombin generation in patients with ALI/ALF on admission

We performed thrombin generation tests in samples taken from patients with ALI/ALF on admission to the hospital. Total thrombin generation as assessed using the ETP was substantially lower in patients as compared with healthy controls in the absence of thrombomodulin, the physiological activator of the protein C system (Fig. 1A). However, when thrombomodulin was added to the test mixture, the ETP was not significantly different between patients and controls. Furthermore, whereas the ETP decreased substantially in the controls upon addition of thrombomodulin, patients were fully resistant to the action of thrombomodulin, because there was virtually no change in the ETP when thrombomodulin was added to patient plasma (Fig. 1B), which is compatible with the profound protein C deficiency in these samples (median 5% of normal) as described previously [11].

image

Figure 1.  Thrombin generation in ALI/ALF. (A) Box-and-whisker plots summarizing endogenous thrombin potential (ETP) values derived from thrombin generation curves in the presence or absence of thrombomodulin (TM) from plasma of 40 healthy controls and 50 patients with ALI/ALF sampled on the day of admission. (B) Box-and-whisker plots summarizing thrombomodulin sensitivity ratios (TM-SR) from the data in panel A. A TM-SR > 1 indicates a decreased anticoagulant response to TM in comparison to pooled normal plasma. ***P < 0.001.

Download figure to PowerPoint

A more detailed inspection of the thrombin generation curves revealed no differences in the lag time, which is the time from start of the assay until the first thrombin is being detected, between patients and controls (Fig. 2A). However, patients showed a decreased time to the peak of thrombin generation as compared with the controls, indicating that once thrombin generation has started, it progresses more rapidly in ALI/ALF patients as compared with controls (Fig. 2B). The increased velocity index in patients confirms a more rapid increase in thrombin generation over time in patients as compared with controls (Fig. 2C). The lag time and velocity index were not significantly affected by addition of thrombomodulin in both patients and controls, whereas the time-to-peak was reduced by addition of thrombomodulin only in the controls. The peak thrombin level was higher in controls as compared with patients in the absence of thrombomodulin, but was similar in patients and controls when thrombomodulin was present in the test mixture (Fig. 2D).

image

Figure 2.  Thrombin generation characteristics in ALI/ALF. Box-and-whisker plots summarizing parameters derived from thrombin generation curves generated in the presence or absence of thrombomodulin (TM) from plasma of 40 healthy controls and 50 patients with ALI/ALF sampled on the day of admission. *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant.

Download figure to PowerPoint

Previously, we have reported substantially decreased levels of both pro- and anticoagulant proteins in this patient cohort, with the exception of FVIII, which was substantially increased compared with healthy controls [11]. In the current study, we also assessed plasma levels of TFPI, which were not measured in the previous study. Plasma levels of TFPI were substantially increased in patients with ALI/ALF as compared with healthy controls (39 ng mL−1 [11–215] in patients vs. 16 ng mL−1 [5–48] in controls, median [range], P < 0.001).

The ETP in the presence of thrombomodulin was not significantly different between patients that had spontaneous recovery or those who died or received a liver transplant (736 nm IIa × min [343–1153] vs. 798 [493–1247], median [range], P = 0.24). Furthermore, the ETP in the presence of thrombomodulin was not different between patients with APAP-induced ALI/ALF and patients with other etiologies (671 nm IIa × min [484–846] vs. 846 nm IIa × min [343–1247], median [range], P = 0.11). Also, the ETP in the presence of thrombomodulin did not discriminate between patients with bleeding complications (no bleeding 770 nm IIa × min [507–1247] vs. patients with bleeding complication 678 [343–1014], median [range], P = 0.14) or thrombotic events (no thrombosis 772 [343–1247] vs. patients with thrombotic events 671 [484–888], median [range], P = 0.28). Routine diagnostic tests (INR and APTT) also did not discriminate between patients that did or did not have bleeding complications (data not shown).

A pronounced hypofibrinolytic status in patients with ALI/ALF on admission

Using a plasma-based global fibrinolytic assay, we assessed plasma fibrinolytic potential in samples from patients with ALI/ALF taken on admission to the hospital in comparison with healthy controls. Whereas the CLT, the time from halfway through the formation of the clot until halfway through the lysis of the clot, is on average around 1 h in controls, 73.5% of clots generated from plasma from patients had not reached the point of half maximal lysis after 3 h, which is the maximal time the assay was run (Fig. 3A). In the control group, only 2.5% of clots did not lyse within 3 h.

image

Figure 3.  Fibrinolysis in ALI/ALF. Overall fibrinolytic potential derived from plasma-based clot-lysis experiments (panel A) and box-and-whisker plots summarizing plasma levels of individual fibrinolytic factors (panels B–F) in healthy controls or patients with ALI/ALF sampled on the day of admission. ***P < 0.001.

Download figure to PowerPoint

The proportion of patients showing extreme hypofibrinolysis (i.e. no lysis within 3 h) was not significantly different between patients that had spontaneous recovery or those who died or received a liver transplant. Also, the proportion of patients with extreme hypofibrinolysis did not discriminate between patients with bleeding complications or thrombotic events, nor was it different between patients with APAP-induced ALI/ALF and patients with other etiologies (data not shown).

Plasma levels of fibrinolytic proteins in patients with ALI/ALF on admission

The plasma fibrinolytic potential is determined to a large extent by plasma levels of the profibrinolytic protein plasminogen, and the antifibrinolytic factors antiplasmin, PAI-1 and TAFI [15]. Plasminogen levels (Fig. 3B) were substantially and significantly decreased in patients compared with controls, as were levels of antiplasmin (Fig. 3C) and TAFI (Fig. 3D). Levels of tPA and PAI-1 were both significantly and substantially increased in patients as compared with controls (Fig. 3E,F). Previous studies showed that isolated plasminogen levels below 50% or elevated PAI-1 levels (> 125 ng mL−1) resulted in a lack of clot lysis within 3 h [5]. Indeed, in those plasma samples that did not lyse within 3 h, median plasma levels of plasminogen were lower and median PAI-1 levels were higher as compared with those samples that did show lysis within 3 h. Plasminogen levels in the samples that showed lysis within 3 h were 38% (17–60), and 21% (9–62) in those samples without lysis in 3 h (median [range], P = 0.03). PAI-1 levels were 38 ng mL−1 (18–149) in those samples that lysed within 3 h, compared with 83 ng mL−1 (18–376) in those that did not (median [range], P = 0.02).

Plasma levels of TAFI were substantially higher in patients with APAP-induced ALI/ALF as compared with patients with other etiologies (58 ± 25% for patients with APAP-induced ALF vs. 40 ± 22%, [mean ± SD], P = 0.014). Also, plasma levels of PAI-1 were substantially higher in patients with APAP-induced disease (106 ng mL−1 [28–1023] vs. 39 [18–385], median [range], P = 0.0015). Plasma levels of tPA, plasminogen and antiplasmin were not different between patients with APAP-induced ALF and patients with other etiologies (data not shown).

Persisting normal thrombin generation and hypofibrinolysis during the first 7 days after admittance

From a subset of patients who survived the first day of admission, plasma samples taken one or more days after admission were also studied. We compared total thrombin generation and fibrinolytic potential in those patients for whom two or more samples were available (n = 45). Total thrombin generation in the presence or absence of thrombomodulin was slightly but significantly higher in samples taken between day 2 and day 7 after admission, as compared with samples taken on day 1 (Fig. 4). Samples taken between day 2 and day 7 still showed full thrombomodulin resistance.

image

Figure 4.  Unchanged thrombin generation in ALI/ALF in the first week of admission. Box-and-whisker plots summarizing endogenous thrombin potential values in the absence or presence of thrombomodulin in patients with ALI/ALF sampled either on the day of admission, or one or more times between the 2nd and 7th day after admission. *P < 0.05.

Download figure to PowerPoint

The proportion of samples with extreme hypofibrinolysis (i.e. no lysis within the 3 h of the assay) was lower in samples taken between day 2 and day 7, as compared with samples taken on admission (50% vs. 73.5%, P < 0.001), but was still elevated as compared with the controls, for whom only 2.5% of samples showed extreme hypofibrinolysis (P < 0.001). In accordance with these findings, plasma levels of PAI-1, which are the most important determinant of CLT, were significantly lower in samples taken between days 2 and 7 as compared with samples taken on day 1 (median [range] 60 ng mL−1 [18–1023] vs. 31 [7–389] ng mL−1, P = 0.002), but still elevated compared with controls, for whom median PAI-1 levels were 5.1 ng mL−1 (1.4–55.4) (P < 0.001).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

In this study we performed an in-depth analysis of coagulation and fibrinolysis in patients with ALI/ALF to assess the hemostatic balance in these patients at the time of hospital admission and at later time-points during the first week of admission. On admission, total thrombin generation was similar in patients as compared with healthy controls when thrombomodulin, the physiological activator of the anticoagulant protein C system, was added to the test mixture. Plasma fibrinolytic potential was markedly impaired, with the majority of patients showing no fibrin clot lysis during the course of the experiment. Over time, total thrombin generation slightly increased, whereas hypofibrinolysis persisted. The intact thrombin generation combined with a defective capacity to remove fibrin clots suggests that patients with ALI/ALF are in a normal hemostatic status, or may even be hypercoagulable. The normo- or hypercoagulable status as assessed with modern global tests of coagulation and fibrinolysis is in sharp contrast to results of routine diagnostic tests, such as the PT and APTT, which suggest a bleeding tendency. Thrombin generation capacity at the day of admission as assessed by the ETP did not discriminate between patients with bleeding or thrombotic events or clinical outcome.

The results of this study are in line with hemostatic testing using thromboelastography in the same patients we reported on previously [11]. Clot formation assessed by kaolin-initiated thromboelastography was shown to be preserved despite profoundly abnormal levels of coagulation factors and inhibitors, and substantially prolonged routine tests of coagulation (PT and APTT). In this study we used tissue factor-induced coagulation in plasma to assess thrombin generation profiles in time. Calibrated automated thrombography allows detailed analyses of thrombin generation in plasma that is clotted by the physiological activator of coagulation. The decreased total thrombin generation in the absence of thrombomodulin is consistent with the decreased plasma levels of procoagulants. When thrombomodulin, the physiological activator of the anticoagulant protein C, was added to the test mixture, total thrombin generation in patients was indistinguishable from that in controls, which is likely to be attributable to a concomitant decrease in pro- and anticoagulant proteins similar to the rebalanced hemostatic status in cirrhosis [2]. Surprisingly, total thrombin generation in the presence of thrombomodulin was similar in ALI/ALF patients and healthy controls despite the substantially elevated levels of TFPI, an important inhibitor of initiation of coagulation, in the patients.

A more detailed inspection of the thrombin generation curves revealed that once thrombin generation has started, the velocity of thrombin generation is higher in ALI/ALF patients as compared with healthy controls. A higher velocity of thrombin generation has been shown to affect the characteristics of the fibrin clot [16]. Thinner fibers that form networks with smaller pores are formed at higher rates of thrombin generation, and these less permeable networks are less susceptible to fibrinolysis [17]. Part of the thrombin generation profile thus suggests hypercoagulability, which is in line with the clinical observation that thrombotic complications are common in patients with ALI/ALF [11]. Furthermore, besides systemic thrombotic complications, there is evidence of intrahepatic thrombus formation in patients with acute liver failure [18]. Animal models of acute liver failure have demonstrated that intrahepatic thrombus formation contributes to the progression of acute liver failure, and have demonstrated that anticoagulant treatment alleviates liver injury induced by acetaminophen or Fas ligand [19,20]. Thus, also this local intrahepatic thrombotic phenomenon may be (in part) related to a hypercoagulable status.

Patients with ALI/ALF show a profound hypofibrinolytic status, which is in sharp contrast to the normo- or hyperfibrinolytic status that has been reported in patients with cirrhosis [5,6]. Hypofibrinolysis as assessed with the assay used in this study has been shown by us and others to be associated with a substantially increased risk of a first venous or arterial thrombosis [15,21,22]. Dysregulated fibrinolysis may thus also explain the thrombotic episodes in some of the patients, and may contribute to progression of the disease as intrahepatic thrombi cannot be removed by endogenous fibrinolysis. The hypofibrinolytic state is likely to be attributed to the combination of very low levels of plasminogen and substantially elevated PAI-1 levels. It has to be noted that the PAI-1 levels assessed by ELISA in this study reflect both free PAI-1 and PAI-1 in complex with tPA. However, given the much higher PAI-1 antigen levels in comparison to the tPA antigen levels, the majority of the PAI-1 antigen in the patients will not be complexed with tPA. Previously, we have shown substantially decreased clot lysis in this assay at plasminogen levels below 50% and at PAI-1 levels above 125 ng mL−1 [5], both of which are frequent in the patients studied. The decreased levels of fibrinolytic proteins plasminogen, TAFI and antiplasmin may be explained by decreased synthesis and/or consumption, whereas the increased plasma levels of tPA and PAI-1 may be explained by endothelial cell activation as a result of systemic stresses induced by the disease.

The combined results of our laboratory investigations may have important clinical consequences. Whereas there appears to be consensus that prophylactic correction of the PT/INR to avoid spontaneous bleeding in patients with ALI/ALF is not indicated [9], it is still common practice in many centers to try to correct the coagulopathy in these patients by infusion of plasma concentrates or recombinant factor VIIa prior to invasive procedures [10,23]. Prohemostatic therapy is initiated based on the assumption that the prolonged PT/INR is indicative of a bleeding risk. Both clinical and laboratory evidence, including the data of the present study, suggest that hemostasis is preserved in patients with ALI/ALF due to the concomitant decrease of both pro- and antihemostatic drivers [8,11]. We believe that prohemostatic therapy should be cautiously used in patients with liver disease for a number of reasons. First, transfusion of blood products is associated with the risk of general transfusion reactions, some of which may even be higher in patients with liver disease [24]. Second, partial correction of the PT/INR requires transfusion of massive amounts of fresh frozen plasma, which carries a significant risk of volume overload, which may exacerbate intracranial hypertension. Third, administration of prohemostatic therapy might result in exacerbation of intrahepatic thrombus formation, which may result in a more rapid progression of the disease [19,20]. Given the profound hypofibrinolytic status, exacerbation of intrahepatic thrombosis with prohemostatic therapy appears to be an important concern. This may be particularly the case when rFVIIa is given as a prohemostatic drug, because there appears to be an increased thrombotic risk associated with the drug in patients with liver disease [25,26]. Fourth, whereas the INR is not useful in predicting bleeding in patients with ALI/ALF, it is a useful indicator of recovering or worsening liver function, and this indicator is obscured by administration of fresh frozen plasma or rFVIIa.

A potential limitation of our study is the fact that blood samples of the healthy controls that have been used to establish reference values for the various parameters studied have not been treated identically in terms of blood draw and sample processing. Nevertheless, even with this potential caveat in mind, the differences in the major parameters studied (TM-SR and CLT) are so large that they are unlikely to be explained by subtle differences in sample collection and processing.

In conclusion, patients with ALI/ALF are capable of generating similar amounts of thrombin as compared with controls, whereas the fibrinolytic system is profoundly inhibited. This intact hemostatic capacity as tested by modern tests is in sharp contrast to the results of routine diagnostic tests of hemostasis, such as the PT and APTT. The intact hemostatic potency is confirmed by clinical observations, and suggests that prohemostatic therapy should be used cautiously in these patients.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

This study was supported in part by a grant from the Stichting Tekke Huizinga Fonds.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References