An unbalance between von Willebrand factor and ADAMTS13 in acute liver failure: Implications for hemostasis and clinical outcome


  • Greg C. G. Hugenholtz,

    1. Surgical Research Laboratory, Department of Surgery, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
    Search for more papers by this author
  • Jelle Adelmeijer,

    1. Surgical Research Laboratory, Department of Surgery, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
    Search for more papers by this author
  • Joost C. M. Meijers,

    1. Section of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
    2. Vascular Medicine, Academic Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Robert J. Porte,

    1. Departments of Experimental Vascular Medicine, Amsterdam, The Netherlands
    Search for more papers by this author
  • R. Todd Stravitz,

    1. Section of Hepatology and Hume-Lee Transplant Center, Virginia Commonwealth University, Richmond, VA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Ton Lisman

    Corresponding author
    1. Surgical Research Laboratory, Department of Surgery, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands
    2. Departments of Experimental Vascular Medicine, Amsterdam, The Netherlands
    • Address reprint requests to: Ton Lisman, Ph.D., Department of Surgery, BA44, University Medical Center Groningen, P.O. Box 30.001, 9700 RB Groningen, the Netherlands. E-mail:; fax: +31−50-363-2796.

    Search for more papers by this author
    • These authors contributed equally to this work.

  • Potential conflict of interest: Nothing to report.

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

  • This work was an ancillary study of the Acute Liver Failure Study Group (National Institute of Diabetes and Digestive and Kidney Diseases (grant U01 DK58369; William M. Lee, M.D., Principal Investigator).


Emerging evidence supports the concept of a rebalanced hemostatic state in liver disease as a result of a commensurate decline in prohemostatic and antihemostatic drivers. In the present study, we assessed levels and functionality of the platelet-adhesive protein von Willebrand factor (VWF) and its cleaving protease ADAMTS13 in the plasma of patients with acute liver injury and acute liver failure (ALI/ALF). Furthermore, we explored possible associations between VWF, ADAMTS13, and disease outcome. We analyzed the plasma of 50 patients taken on the day of admission for ALI/ALF. The plasma of 40 healthy volunteers served as controls. VWF antigen levels were highly elevated in patients with ALI/ALF. In contrast, the collagen-binding activity and the ratio of the VWF ristocetin cofactor activity and VWF antigen was significantly decreased when compared with healthy controls. Also, the proportion of high molecular weight VWF multimers was reduced, despite severely decreased ADAMTS13 levels. In spite of these functional defects, platelet adhesion and aggregation were better supported by plasma of patients with ALI/ALF when compared with control plasma. Low ADAMTS13 activity, but not high VWF antigen, was associated with poor outcome in patients with ALI/ALF as evidenced by higher grades of encephalopathy, higher transplantation rates, and lower survival. VWF or ADAMTS13 levels were not associated with bleeding or thrombotic complications. Conclusion: Highly elevated levels of VWF in plasma of patients with ALI/ALF support platelet adhesion, despite a relative loss of function of the molecule. Furthermore, low ADAMTS13 activity is associated with progressive liver failure in the patient cohort, which might be attributed to platelet-induced microthrombus formation in the diseased liver resulting from a substantially unbalanced VWF/ADAMTS13 ratio. (Hepatology 2013;58:752–761)


acute liver failure


acute liver injury




enzyme-linked immunosorbent assay


high molecular weight


international normalized ratio


low molecular weight


Model for End-Stage Liver Disease




ultralarge VWF multimers


von Willebrand factor


von Willebrand factor antigen


von Willebrand factor ristocetin cofactor activity

Concepts of the clinical consequences of the hemostatic disorders in patients with liver failure have changed considerably over the last decade. It is now well established that patients with chronic liver failure and abnormal routine coagulation tests do not necessarily have an increased bleeding tendency and that thrombotic complications may occur in these patients.[1, 2] Moreover, recent studies of the coagulopathy of liver failure suggest a link between intrahepatic thrombosis and the progression of liver failure.[3, 4]

In patients with cirrhosis, the net effect of all hemostatic changes is a rebalanced yet precarious system that may easily tip toward a bleeding diathesis or a thrombotic tendency.[1] In acute liver failure, the net effect of all hemostatic changes is not clear, partly because the changes in the hemostatic system in these patients have been less well defined compared with those in patients with cirrhosis.

In an effort to elucidate this issue, we are systematically studying consequences of hemostatic defects in patients with acute liver failure. We recently demonstrated an intact thrombin generating capacity in plasma from patients with acute liver injury and acute liver failure (ALI/ALF) despite severely reduced plasma levels of coagulation factors and abnormal routine diagnostic tests of coagulation, such as the prothrombin time.[5] This intact thrombin generation has been ascribed to a concomitant decrease in both procoagulant and anticoagulant factors. In vivo, however, thrombin generation is not only a function of procoagulant and anticoagulant factors, but also of platelets.[6] The platelet surface provides a scaffold for the assembly of coagulation factor complexes, and this assembly is an essential step in the thrombin generation pathway. Primary and secondary hemostasis, therefore, are integrated physiologically to facilitate thrombin generation and fibrin formation.

In view of the physiological importance of platelets in supporting coagulation, we now aim to better define changes in the primary hemostatic system of patients with ALI/ALF and their net effect on bleeding, thrombosis, and disease progression. Our group initially studied parameters reflecting platelet function by thromboelastography using whole blood of patients with ALI/ALF.[7] We found evidence of normal to increased platelet activity in whole blood of patients with ALI/ALF when compared with normal controls despite reduced platelet numbers in a proportion of patients. The exact mechanisms underlying the observed increase in parameters reflecting platelet function and adhesion are unknown, but it may be attributed to increased levels of the adhesive protein von Willebrand factor (VWF). Indeed, we have demonstrated that elevated levels of VWF may (over)compensate for abnormalities in platelet number and function in patients with cirrhosis.[8] These high VWF plasma levels result from disease-related overactivation of the reticulo-endothelial system in endothelial cells.[9, 10] VWF is a large, multimeric protein, and its interaction with platelet glycoprotein Ib is essential for platelet adhesion under conditions of flow, as evidenced by the bleeding tendency associated with qualitative or quantitative defects in VWF in von Willebrand disease. The functional capacity of VWF is normally strictly regulated in the blood by the VWF-cleaving protease, ADAMTS13, as VWF reactivity towards platelets is directly proportional to its multimeric size.[11] The importance of this regulation is apparent from patients with congenital or acquired ADAMTS13 deficiency who have severe thrombotic episodes in the microvasculature (i.e., thrombotic thrombocytopenic purpura).[12]

We hypothesized that VWF also compensates for qualitative or quantitative platelet abnormalities in patients with ALI/ALF. To test this hypothesis, we analyzed qualitative and quantitative parameters of VWF and ADAMTS13 in a group of 50 patients with ALI/ALF in samples taken on admission to a single tertiary referral center. In addition, we used plasma of these patients in a model of primary hemostasis to examine the ability of VWF to support platelet adhesion under physiological flow conditions. Finally, given that the liver is the major source of ADAMTS13 synthesis, we anticipated reduced ADAMTS13 plasma levels with a consequent substantial unbalance between ADAMTS13 and VWF in these patients. A VWF/ADAMTS13 unbalance is a potential high-risk state for unintentional platelet (micro)thrombus formation.[13] Emerging evidence from epidemiological, clinical, and animal studies indicates that intrahepatic activation of hemostasis and formation of microthrombi contributes to liver failure progression,[3, 4, 14] and ADAMTS13 and VWF have even been proposed as new predictors for outcome in patients with liver failure.[15-17] Therefore, we also explored possible relationships between VWF, ADAMTS13, and the outcome of patients with ALI/ALF in the present study.

Patients and Methods


Fifty consecutive patients were prospectively studied after admission for acute liver injury/acute liver failure (ALI/ALF) to Virginia Commonwealth University Medical Center between March 2009 and May 2011. Patients' details have been provided.[7] Informed consent was obtained from either the patient or their next-of-kin, depending on the patient's level of altered mental status (hepatic encephalopathy), as part of entry into the US Acute Liver Failure Study Group Registry. Patients with acute liver injury were defined as those with (1) an international normalized ratio (INR) of ≥ 1.5; (2) absence of a history of liver disease; and (3) illness of ≤26 weeks duration. Patients with ALF were defined as those with ALI and hepatic encephalopathy. Patients who received procoagulant treatment other than vitamin K prior to enrollment were excluded. The INR was assayed using the Innovin reagent (Siemens Healthcare Diagnostics, Marburg, Germany), which has an international sensitivity index of 0.9. We calculated Model for End-Stage Liver Disease (MELD) scores according to the equation: [0.957 × loge (creatinine) + 0.378 × loge (bilirubin) + 1.12 × loge (INR) + 0.643] × 10, and determined whether patients fulfilled the King's College criteria for acute liver failure as described.[18]

Plasma samples from 40 healthy volunteers were used to establish reference values for the various tests performed in this study. Blood sample retrieval and processing have been described.[7] Samples were centrifuged within 2 hours of withdrawal and immediately placed at −80°C and stored until assayed. The study was conducted in accordance with local Institutional Review Board regulations.

Definition of Complications and Final Outcomes of ALI/ALF

We studied the following complications of ALI/ALF: hepatic encephalopathy, infection, systemic inflammatory response, renal failure, thrombosis, and bleeding. These complications were defined as follows:

Hepatic encephalopathy was defined and graded according to West Haven criteria.[19] Infection was defined as a positive urine culture, presence of a pulmonary infiltrate on chest X-ray consistent with infectious etiology, or a positive blood culture not felt to be a contaminant with a skin organism. More than one positive blood culture was required for bacteremia with commensal organisms. Systematic inflammatory response syndrome was defined according to established criteria[20]: white blood cell count >12 or <4 × 109 cells/L, temperature <36°C or >38°C, respiratory rate >20/minutes, and pulse >90 beats per minute. Renal failure was defined as persistent azotemia or oliguria despite rehydration requiring continuous veno-venous hemofiltration. Thrombosis was defined as occlusion of a native blood vessel or indwelling dialysis catheter. When occlusion of a native blood vessel was suspected on clinical grounds, these were confirmed by ultrasound or CT scanning. Bleeding was defined as the presence of blood per naso-gastric tube, blood per rectum or endotracheal tube, or bleeding at the site of invasive procedure. Final outcomes of ALI/ALF were transplant-free survival, orthotopic liver transplantation, or death.

VWF and ADAMTS13 Assays

VWF antigen (VWF:Ag) levels were determined with an in-house enzyme-linked immunosorbent assay (ELISA) assay using commercially available polyclonal antibodies against VWF (DAKO, Glostrup, Denmark).

VWF ristocetin cofactor activity (VWF:RCo) was determined using the BC VWF-reagent (Siemens Healthcare Diagnostics) on a Behring Coagulation System (Siemens Healthcare Diagnostics). VWF:Ag and VWF:RCo levels of pooled normal plasma were set at 100% and the values obtained in patient samples were expressed as a percentage of pooled normal plasma.

VWF collagen binding activity was determined with an in-house ELISA assay as described.[8] The collagen-binding activity of pooled normal plasma was set at 100% and the activity measured in patient samples was expressed as a percentage of pooled normal plasma.

VWF multimer analysis was performed by sodium dodecyl sulfate agarose gel electrophoresis followed by western blotting. The blots were incubated with rabbit anti-VWF antibody (DAKO) and goat anti-rabbit IRDye 800 CW (LI-COR Biosciences, Lincoln, NE). The first five bands were considered as low-molecular weight multimers, whereas other bands were considered as high molecular weight (HMW) multimers. The blots were scanned by the Odyssey Imager (Westburg, Leusden, The Netherlands) and were quantified by morphometric analysis using the ImageScope software package (Aperio, Vista, CA). After shading correction and interactive thresholding, the selected positive pixels were measured. The positive area was the sum of the area of positive pixels of low-molecular weight and HMW bands. Data was expressed as the percentage of HMW multimers per total VWF multimers, which equals the percentage of positive pixels in the HMW band area per total positive pixel area.

ADAMTS13 activity was measured in plasma of patients with ALI/ALF and pooled plasma of healthy volunteers which was pretreated with bilirubin oxidase (10U/mL; Sigma-Aldrich, Zwijndrecht, The Netherlands) to avoid interference of bilirubin with the assay. Activity was assessed using the FRETS-VWF73 assay (Peptanova, Sandhausen, Germany) based on the method described by Kokame et al.[21] The activity of ADAMTS13 in normal pooled plasma was set at 100%, and values obtained in test plasmas were expressed as a percentage of pooled normal plasma.

ADAMTS13 antigen levels were measured using a commercially available ELISA according to the manufacturer's instructions (Sekisui Diagnostics, Stamford, CT).

Platelet Adhesion Assay

The ability of VWF from patients with ALI/ALF to support platelet adhesion was studied under flow conditions in a reconstituted blood model. Red blood cells and platelets were isolated from whole blood of healthy volunteers who had blood group O as described.[22] Cells were mixed with patient plasma or plasma from healthy volunteers to obtain reconstituted blood with a hematocrit of 40% and a platelet count of 250,000/μL. VWF-dependent platelet adhesion in reconstituted blood samples was assessed using a cone and plate viscometer (Diamed Impact R, Turnhout, Belgium). Uncoated Diamed wells were perfused at shear rate of 1,800/second for 2 minutes according to the instructions of the manufacturer. Platelet adhesion was quantified using May-Grünwald staining followed by software-assisted morphometric analysis using the Diamed apparatus and software delivered by the manufacturer.

Statistical Analysis

Statistical analysis was performed with the Graphpad InStat (San Diego, CA) software package. Continuous variables are expressed as the mean ± SD or median and range. Continuous data were tested for normality and analyzed by t test or Mann-Whitney U test as appropriate. Categorical data are expressed as numbers and percentage. P < 0.05 was considered statistically significant.


Demographic, Laboratory, and Clinical Characteristics of Patients With ALI/ALF

Patient demographics, vital signs, and laboratory test results at the time of admission for ALI/ALF, as well as clinical outcome data, are presented in Table 1. The mean age of the cohort was 43 years, 64% of the patients were female, 58% of the patients were Caucasian, and the mean body mass index was 28 kg/m2. The etiologies of ALI/ALF in this cohort were acetaminophen (APAP) overdose in 50%; hepatitis B virus infection in 14%; idiosyncratic drug reactions in 12%; autoimmune hepatitis in 10%; indeterminate in 6%; and heat stroke, Amanita mushroom poisoning, malignant infiltration, and hepatic ischemia in 2% each. All patients with APAP-induced ALI/ALF (50%) were treated with N-acetylcysteine (NAC), as were 19 patients (38%) with non–APAP-induced liver injury. Laboratory results included a median aspartate aminotransferase and alanine aminotransferase level of 4,990 and 3,578 IU/L, respectively; a median creatinine, lactate, and bilirubin level of 1.0, 3.5, and 5.0 mg/dL, respectively; a mean pH of 7.35, bicarbonate level of 19.7 mg/dL, phosphate level of 3.3 mg/dL, and INR of 3.4. On admission, 39 (78%) patients with ALI had already developed hepatic encephalopathy, and 24 (48%) progressed to high-grade encephalopathy (grade 3 or 4) at some point over the first week of admission. Complications of the study population other than encephalopathy included infection in 13 (26%) patients, systemic inflammatory response syndrome in 28 (56%), renal failure in 18 (36%), and thrombosis and bleeding in 9 (18%) patients each. The thrombotic complications included bowel ischemia due to thrombosis detected by contrast tomography and ultrasound (n = 1), limb ischemia due to both arterial and venous thromboses detected by Doppler ultrasound (n = 1), portal vein thrombosis detected by Doppler ultrasound (n = 1), and thrombosed continuous veno-venous hemofiltration catheters (n = 6). Twenty-eight (56%) patients recovered spontaneously, 7 (14%) patients underwent liver transplantation, and 15 (30%) patients died.

Table 1. Demographic, Laboratory, and Clinical Characteristics of the Study Cohort
  1. Abbreviations: AIH, autoimmune hepatitis; ALT, alanine aminotransferase; aPTT, activated partial thromboplastin time; AST, aspartate aminotransferase; BMI, body mass index; HBV, hepatitis B virus; MAP, mean arterial pressure; SIRS, systematic inflammatory response syndrome; WBC, white blood cell count.

Age, years, mean (SD)43.1 (13.5)
Female sex, n (%)32 (64)
Caucasian race, n (%)29 (58)
BMI, kg/m2, mean (SD)28.2 (6.8)
MAP on admission, mm Hg, mean (SD)84.8 (14.5)
Etiology, n (%) 
APAP25 (50)
HBV7 (14)
Idiosyncratic drug6 (12)
AIH5 (10)
Indeterminate3 (6)
Heat stroke1 (2)
Malignancy1 (2)
Amanita1 (2)
Ischemia1 (2)
NAC treatment, n (%)44 (88)
Admission laboratory tests 
AST, IU/L, median (range)4,990 (138-19,060)
ALT, IU/L, median (range)3,578 (197-12,729)
Creatinine, mg/dL, median (range)1.0 (0.4-7.5)
Total bilirubin, mg/dL, median (range)5.0 (0.3-44.2)
INR, mean (SD)3.4 (1.8)
aPTT, seconds, mean (SD)47.4 (14.7)
Platelet count, ×109/L, mean (SD)193 (99)
Hemoglobin, g/L, mean (SD)12.0 (2.2)
Lactate, mg/dL, median (range)3.5 (0.4-20)
Venous ammonia, μmol/L, mean (SD)80.3 (45.3)
Phosphate, mg/dL, mean (SD)3.3 (2.2)
pH, mean (SD)7.35 (0.13)
HCO3, mg/dL, mean (SD)19.7 (7.7)
Hepatic encephalopathy, n (%) 
Admission39 (78)
Maximal grade 3 or 424 (48)
Infection, n (%)13 (26)
SIRS (2-4 components present), n (%)28 (56)
Admission SIRS 
WBC, ×109/L, mean (SD)10.7 (5.7)
Heart rate, beats per minute, mean (SD)99 (22)
Respiratory rate, breaths per minute, mean (SD)20 (6)
Temperature, °C, mean (SD)36.7 (1.0)
Renal failure requiring renal replacement therapy, n (%)18 (36)
Thrombosis, n (%)9 (18)
Bleeding, n (%)9 (18)
Final outcome 
Transplant-free survival, n (%)28 (56)
Liver transplantation, n (%)7 (14)
Death, n (%)15 (30)

Highly Elevated VWF:Ag Levels in Patients With ALI/ALF

As shown in Fig. 1A, VWF:Ag levels were substantially elevated in patients with ALI/ALF (547% [242%-1,420%])) when compared with the reference group in which the median VWF:Ag level was 107% (38%-180%) (P < 0.01). Interestingly, VWF:Ag levels were not different between patients with blood group O compared with those with non-O blood groups (583% [267%-1,027%] versus 558% [243%-1,429%], respectively) (P = 0.977).

Figure 1.

(A) VWF:Ag levels, (B) VWF:RCo levels, (C) VWF:RCo/VWF:Ag ratio, and (D) VWF collagen-binding activity in patients with ALI/ALF and in healthy controls. VWF:Ag levels and VWF:RCo levels are expressed as a percentage of pooled normal plasma. The collagen-binding activity of pooled normal plasma was set at 100%. The collagen-binding activity was measured at equal plasma levels of VWF:Ag. Horizontal bars represent medians.

Low VWF:Rco/VWF:Ag Ratio in Patients With ALI/ALF

In Fig. 1B, it is shown that VWF:Rco activity was substantially elevated in patients with ALI/ALF (278% [11%-684%]) compared with the healthy control group in which the median activity was 105% (33%-222%) (P < 0.01). However, the VWF:RCo levels, which reflect the ability of VWF to bind the platelet receptor glycoprotein Ib, are not elevated to the same extent as the VWF:Ag levels. In other words, although VWF is substantially elevated in patients with ALI/ALF, its binding to glycoprotein Ib is inferior in these patients. This is demonstrated in Fig. 1C by a significantly depressed VWF:RCo/VWF:Ag ratio in patients with ALI/ALF (0.55 [0.01-1.06] versus healthy controls (0.96 [0.67-1.54]) (P < 0.01).

Reduced Collagen Binding Activity of VWF in Patients With ALI/ALF

In Fig. 1D, it is demonstrated that the collagen binding activity of VWF is slightly but significantly decreased in patients with ALI/ALF when compared with healthy controls (97% [93%-115%] versus 105% [72%-133%]) (P < 0.01).

Reduced Proportion of HMW-VWF Multimers in Plasma of Patients with ALI/ALF, Despite Severely Decreased ADAMTS13 Activity and Antigen Levels

As shown in Fig. 2A, ADAMTS13 activity was severely reduced in patients with ALI/ALF (28% [0%-106%]) when compared with the healthy control group in which the median activity was 92% [61%-135%] (P < 0.01). ADAMTS13 activity in patients with APAP-induced ALI/ALF were substantially higher compared with levels in patients with other etiologies (38% [0%-106%] in patients with APAP-induced ALF versus 21% [0%-69%] in with ALF from other etiologies (P < 0.05). Of note, in the plasma of six patients, ADAMTS13 activity was <1%; two of these patients had APAP-induced ALF, and four had ALF from other etiologies. As shown in Fig. 2B, ADAMTS13 antigen levels were reduced in patients with ALI/ALF (466 ng/mL [204-1,335 ng/mL] versus 655 ng/mL [359-956 ng/mL] in healthy individuals (P < 0.01), but the decrease in antigen levels was not as severe as the decrease in activity levels. This is demonstrated in Fig. 2C by a significantly depressed ADAMTS13 activity/antigen ratio in patients with ALI/ALF (0.06 [0-0.19] versus healthy controls (0.14 [0.11-0.22]) (P < 0.01), indicating that the specific activity of ADAMTS13 was reduced by >50% in patients with ALI/ALF. Despite the observation that the activity of ADAMTS13 was decreased and even below 1% in some patients, the proportion of high-molecular weight VWF multimers was reduced in patients with ALI/ALF (13.5% [4.6-24.9] versus 20.3% [11-24.6] in the plasma of healthy individuals) (P < 0.01).

Figure 2.

(A) ADAMTS13 activity, (B) ADAMTS13 antigen levels, and (C) ADAMTS13 activity/antigen ratio in patients with ALI/ALF and in healthy controls. (D) Proportion of HMW-VWF multimers in plasma from patients with ALI/ALF and in plasma from healthy controls. ADAMTS13 activity was calibrated to pooled normal plasma in which the ADAMTS13 activity was set at 100%. ADAMTS13 antigen levels are expressed in nanograms per milliliters plasma. HMW multimers were quantified by software-assisted morphometric analysis after western blotting and antigen detection. HMW multimers are expressed as a percentage of total VWF multimers in plasma. Horizontal bars represent medians.

Elevated VWF-Dependent Platelet Adhesion and Aggregation in Plasma of Patients With ALI/ALF

We studied the ability of VWF in a given plasma sample to support adhesion and aggregation of platelets isolated from healthy individuals under conditions of high shear in a reconstituted blood model. Representative images from platelets adhered in the presence of plasma from healthy volunteers or plasma from patients with ALF are shown in Fig. 3A,B. On these images, a higher surface coverage was observed when platelets were sheared in the presence of ALI/ALF plasma. Also, the morphological appearance of the platelet thrombi indicated larger aggregates when compared with thrombi generated in plasma of healthy controls. Indeed, as shown in Fig. 3C, the surface covered with platelet thrombi was significantly larger when using patient versus control plasma (11% [4.7-23] versus 9.5% [4.9-14], respectively) (P < 0.01). Concomitantly, platelet aggregation was increased when using plasma of patients with ALI/ALF compared with controls, with an average aggregate size of 29 μm2 (18-77 μm2) versus 23 μm2 (18-35 μm2), respectively (P < 0.01).

Figure 3.

Representative images of platelet adhesion and aggregation using plasma of (A) healthy controls and (B) patients with ALI/ALF in a close to physiological model of primary hemostasis. Plasma from individual patients with ALI/ALF or from healthy controls was mixed with red blood cells and platelets isolated from healthy volunteers and perfused at a shear rate of 1,800/second for 2 minutes. Platelets were visualized by May-Grünwald staining. Images were taken using an Impact R camera (original magnification ×400). (C) Surface coverage and (D) platelet aggregate size from individual reconstituted blood samples were determined by software-assisted morphometric analysis. Horizontal bars represent medians.

Low ADAMTS-13 Activity, but Not High VWF:Ag, Is Associated With Poor Outcome in Patients With ALI/ALF

We explored possible relationships between VWF:Ag levels, ADAMTS13 activity, and the outcome of patients with ALI/ALF. In Fig. 4A, it is shown that patients who received a liver transplant or those who died had a substantially lower ADAMTS13 activity of 17% [0%-69%] on admission when compared with transplant-free survivors (40% [0%-106%]) (P < 0.01). Similarly, as shown in Fig. 4B, patients with encephalopathy grade 4 on admission had profoundly lower ADAMTS13 activity (11% [0%-40%]) compared with patients without encephalopathy on admission (55% [0%-106%]) (P < 0.01). No differences in VWF:Ag levels between patients who spontaneously survived and those who did not were detected, nor were VWF:Ag levels different between the different grades of encephalopathy (Fig. 4C,D). ADAMTS13 activity did not differ between patients that did or did not fulfill the King's College criteria (20% [0%-106%] versus 29% [0%-88%], respectively) (P = 0.42), and was not correlated with MELD scores (r = 0.19, P = 0.21). Table 2 shows that other complications of ALI/ALF were not associated with admission ADAMTS13 activity or VWF:Ag levels in this cohort. Notably, ADAMTS13 activity or VWF antigen levels were not associated with bleeding or thrombosis.

Table 2. ADAMTS13 Activity and VWF:Ag Levels in Patients With and Without Specific Complications of ALI/ALF
Complications of ALI/ALFADAMTS13 Activity, %VWF:Ag, %
  1. Data are expressed as the median (range).

  2. a

    P < 0.05 (hepatic encephalopathy grade 0 versus grade 4). **P < 0.01 (liver transplantation or death versus transplant-free survival).

Hepatic encephalopathy grade  
055 (0-106)523 (299-961)
128 (14-45)568 (429-854)
219 (0-62)505 (243-1,027)
321 (11-51)563 (319-1,420)
411 (0-40)*570 (348-662)
Infection20 (1-48)625 (243-954)
No infection30 (0-106)501 (267-1,420)
SIRS (2-4 components)21 (0-98)610 (243-1,420)
No SIRS45 (10-106)543 (277-845)
Renal failure17 (6-98)625 (470-1,027)
No renal failure28 (10-106)538 (319-1,420)
Thrombosis17 (0-88)625 (267-1,420)
No thrombosis31 (0-106)534 (243-1,027)
Bleeding36 (0-69)535 (243-1,027)
No bleeding25 (0-106)552 (267-1,420)
Final outcome  
Liver transplantation or death17 (0-69)601 (243-1,420)
Transplant-free survival40 (0-106)**512 (267-961)
Figure 4.

Outcome of patients with ALI/ALF according to ADAMTS13 activity and VWF:Ag levels on admission to the hospital. (A) ADAMT13 activity in transplant-free survivors and in patients who received a transplant or died. (B) ADAMTS13 activity according to the degree of hepatic encephalopathy on admission. (C) VWF:Ag in transplant-free survivors and in patients who received a transplant or died. (D) VWF:Ag levels according to the degree of hepatic encephalopathy on admission. Horizontal bars represent medians.


This study shows a remarkable elevation of VWF levels in plasma of patients with ALI/ALF, comparable to the high VWF levels we reported in patients with chronic liver disease.[8] In addition, associated with these high levels of VWF, plasma from patients with ALI/ALF better supported platelet adhesion and aggregation under shear conditions compared with plasma from healthy individuals, consistent with prior observations in cirrhotic plasma.[8] The enhanced platelet adhesion and aggregation occur despite a loss of function of VWF in ALI/ALF as evidenced by a reduced VWF:RCo/VWF:Ag ratio and a reduced collagen-binding activity. Apparently, the decrease in function of VWF is more than compensated for by the quantitative increase in concentration of the molecule.

We are systematically studying consequences of hemostatic defects in patients with ALF. First, we have shown that parameters reflecting primary and secondary hemostasis were normal when assessed by thromboelastography.[7] Secondly, we demonstrated an intact capacity of the secondary hemostatic system to support thrombin generation despite abnormal laboratory test of coagulation such as the PT/INR.[5] The data presented in this study suggest that the primary hemostatic system remains functional, perhaps even overcompensated, as the net effect of alterations in components of the system. We believe that the message of our combined investigations suggests that the hemostatic system of patients with ALF is in fact rebalanced with a normal function, similar to the hemostatic rebalance observed in patients with cirrhosis.[1] This observation may have important clinical consequences. The presence of physiological compensatory mechanisms, such as high VWF levels, sustaining appropriate hemostasis, suggests that the routine correction of abnormal tests of hemostatic function may be unnecessary in patients with ALI/ALF. Indeed, prohemostatic replacement therapy is often initiated based on the assumption that a prolonged PT/INR, but also a decreased platelet count and function, indicates a bleeding risk.[23] Our data suggest that the prophylactic administration of platelet concentrates in ALI/ALF patients with thrombocytopenia or platelet function defects may not be indicated, and may even result in an increased risk of thrombotic complications.

In addition to high VWF levels, we also observed a severe decrease of levels of the VWF cleaving protease, ADAMTS13, in the present study. Furthermore, we observed a low activity-to-antigen ratio, reflecting a loss of function of the ADAMTS13 molecule, which may be due to proteolytic inactivation by, for example, thrombin or plasmin.[24] The decrease in both antigen and activity levels may be explained by a reduced or defective synthesis of the protein in the failing liver, but possibly also by an accelerated turnover of ADAMTS13 molecules driven by ongoing VWF release, similar to the decrease in ADAMTS13 levels in individuals receiving 1-deamino-8-d-arginine vasopressin.[25] Despite defective ADAMTS13 activity, we observed a reduced rather than an increased proportion of HMW-VWF multimers in patients compared with controls, suggesting that other proteases, such as plasmin, elastase, or cathepsin G, may be responsible for the processing of the freshly released VWF.[26] A complementary explanation for the reduced percentage of HMW-VWF multimers may be that the vast majority of the patients in this cohort were treated with NAC, which was recently shown to effectively reduce the size of VWF multimers in human plasma.[27] In the present study, all blood samples were taken after administration of NAC, and we are thus not able to ascertain whether the reduced proportion of VWF multimers observed in our patients are due to proteolysis by proteases other than ADAMTS13 or by the effect of NAC on VWF. In future studies, comparisons between samples taken prior to and after administration of NAC will be required to investigate to what extent NAC contributes to VWF proteolysis in patients with ALF.

The elevated VWF levels combined with a substantial decrease of ADAMTS13 activity may have adverse clinical consequences for the patient with ALI/ALF. An unbalanced ADAMTS13/VWF ratio has been shown to be a risk factor for arterial thrombosis[28] and may lead to the local formation of platelet-rich thrombi resulting in organ dysfunction in several pathologies, including thrombotic thrombocytopenic purpura, severe sepsis, malaria, and Dengue fever.[12, 29-31] In the present study, we demonstrated for the first time that low ADAMTS13 activity was associated with a poor outcome of patients with ALI/ALF. This association appeared independent of established predictors of poor outcome such as the King's College criteria or the MELD score, which may indicate that further research into the prognostic value of ADAMTS13 is warranted. A potential drawback of ADAMTS13 as a prognostic indicator is that the laboratory test is currently only available in specialized hemostasis laboratories.

Interestingly, low ADAMTS13 activity did not appear to be related to systemic thrombotic complications. The occurrence of massive systemic thrombosis is a characteristic feature in patients with thrombotic thrombocytopenic purpura in consequence of an isolated ADAMTS13 deficiency.[12] The absence of such a phenotype in patients with ALI/ALF likely reflects adequate processing of ultralarge VWF multimers (ULVWF) in ALI/ALF, at least in the systemic circulation. The combination of ADAMTS13 deficiency and local endothelial cell activation within the failing liver may still have led to locally elevated ULVWF levels. We hypothesize that ULVWF-induced formation of platelet thrombi within the hepatic microvasculature led to tissue ischemia resulting in the progression of the disease course in patients with low ADAMTS13 activity. Intrahepatic thrombosis has been shown to promote the progression of chronic liver failure in several epidemiological studies and animals studies[14, 32, 33] and recently in a clinical study that assessed the efficacy of low molecular weight heparin in preventing portal vein thrombosis.[4] In addition, animal studies have also shown that intrahepatic formation of fibrin clots contributes to the progression of ALF,[34] and we speculate that intrahepatic formation of platelet-rich thrombi produces similar effects.

In conclusion, highly elevated levels of VWF in patients with ALI/ALF supported a (supra) normal primary hemostatic function, despite a loss of function of the molecule. Furthermore, low ADAMTS13 activity was associated with progressive liver failure in the patient cohort, which might be attributed to platelet-induced microthrombus formation in the diseased liver resulting from a locally unbalanced VWF/ADAMTS13 ratio.