Patients With Budd–Chiari Syndrome Have Variable Coagulation Status on Thromboelastography at Diagnosis

Abstract

Background & aims

Budd–Chiari Syndrome (BCS) is considered a thrombophilic state, and most patients with BCS have thrombophilic disorder. Liver dysfunction–related coagulopathy makes coagulation function unpredictable in BCS. Thromboelastography (TEG) assesses the dynamics, strength, and stability of clot formation. We conducted a pilot study using TEG to evaluate coagulation status in patients with BCS.

Methods

Fifty-one patients with newly diagnosed BCS (age 32.3 [10.7] years; 23 men) underwent TEG (TEG®5000 Hemostasis Analyzer®, USA), and its components were analyzed and correlated with clinical profile and thrombophilic disorders. Patients who had received anticoagulation, antiplatelet drugs, or radiological intervention were excluded.

Results

Twenty-nine patients had normal TEG, 11 had procoagulant TEG, and 11 had hypocoagulant TEG. Among patients with hypocoagulant TEG, Coagulation Index (CI) was < −3 in 11 patients, R was >8 min in 6 patients, K was >3 min in 9 patients, alpha <55 in 9 patients, and MA <51 in 7 patients; among those with hypercoagulant TEG, CI was >3 in 3 patients, R < 2 min in 2 patients, K <1 min in 2 patients, alpha >78 in none, and MA >69 mm in 7 patients. TEG findings were similar in patients with and without thrombophilic disorder. The mean platelet count (1.75, 2.22, and 1.79 × 10⁵/mm³; P = 0.13) and international normalized ratio (1.27, 1.34, and 1.28, P = 0.69) were similar in those with procoagulant, normal, and hypocoagulant TEG. Two patients in Rotterdam class-III had abnormal LY30. Other clinical parameters did not correlate with TEG findings.

Conclusion

Patients with BCS are heterogeneous with respect to coagulation status, with one-fifth of patients are hypocoagulant on TEG. Patients with advanced disease may have accelerated fibrinolysis.

Budd–Chiari Syndrome (BCS) is associated with presence of inherited and acquired thrombophilic disorders, which are detected in 84% of patients.^{1, 2, 3, 4} Up to 25% of patients may have more than one thrombophilic state.⁶ BCS is considered to be a prothrombotic state, and anticoagulation is recommended in all patients as first-line therapy.^{3, 5} A combination of thrombophilic disorders and liver dysfunction affects the balance of coagulation status and thus may further affect the results of baseline International Normalized Ratio (INR) and the INR value after anticoagulation.

INR is affected by the decrease in hepatic synthesis of procoagulant factors but is not a good predictor of bleeding in patients with cirrhosis.⁷ Presence of endothelial dysfunction, a reduction in anticoagulant factors, and increase in von Willebrand factor and factor VIII levels in patients with cirrhosis complicate the coagulation status.⁷ INR does not account for the impact of these changes on the status of coagulation balance in patients with cirrhosis.^8,9 It is also a poor predictor of bleeding episodes in patients with liver disease.⁷

Thromboelastography (TEG) assesses clot formation in whole blood, including plasma and cellular components.¹⁰ It demonstrates the net effect of all the hemostatic components including coagulation factors, platelets, and other cellular elements through all phases, from clot initiation to clot lysis. In patients with cirrhosis, TEG has been used in peritransplant setting and to guide requirement of blood products prior to invasive procedures.¹¹ It provides an overall assessment of hemostasis and thus can be utilized to detect thrombophilic potential as well as hyperfibrinolysis to predict risk of bleeding .^{10, 12, 13}

We therefore did a pilot study to evaluate the coagulation status using TEG in patients with BCS.

Methods

Patient Characteristics

Fifty-four consecutive patients with BCS diagnosed on basis of imaging (magnetic resonance imaging/computed tomography/Doppler) were evaluated. Out of these, three were excluded because of incomplete data. Clinical details and demographics of the patients were recorded in a proforma. Patients underwent complete hemogram, liver biochemistry (serum bilirubin, aspartate transaminase, alanine transaminase, alkaline phosphatase, and albumin) and prothrombin time with INR. Rotterdam score was calculated to determine the severity of BCS and stratify patients into class I, II, and III.¹⁴

Thrombophilic workup included determination of genetic defects in the factor V Leiden and G20210A prothrombin gene, protein C, protein S, antithrombin III, Methylene Tetrahydrofolate Reductase (MTHFR) gene mutation, determination of antiphospholipid antibodies and lupus anticoagulant, and flow cytometry testing for paroxysmal nocturnal hemoglobinuria and homocysteine. For underlying myeloproliferative disorder, Janus Kinase-2 (JAK-2) V617F mutation analysis was done. The result of JAK 2 V617F mutation was recorded as present or absent. Results of factor V Leiden, prothrombin gene, and MTHFR gene mutation were recorded as homozygous mutation, heterozygous mutation, or normal (no mutations). The presence of antiphospholipid syndrome was diagnosed as per the Sydney consensus.¹⁵ Protein S and protein C were measured by enzyme-linked immunosorbent assays. Reference range was defined as the mean (standard deviation) of this distribution and was approximately 70–140%. Antithrombin III activity was measured. Homocysteine value more than 16 ng/ml was considered significant. Absence or reduced expression of both CD59 and CD55 on red blood cells detected by flow cytometry was diagnostic of paroxysmal nocturnal hemoglobinuria.

Patients with active sepsis, renal dysfunction, recent trauma, or blood transfusion (over last 3 months) and those having an additional etiology for liver disease, e.g. chronic viral hepatitis and alcohol and known bleeding disorder were excluded.

Thromboelastography

TEG was performed, using TEG® 5000 (Hemostasis Analyzer®, USA), in all patients after stopping drugs affecting coagulation status. Heparin was stopped 48 h before testing, and warfarin was stopped 4 weeks before testing, and antiplatelet drugs were stopped 14 days before testing. Upon discontinuation of warfarin, patients were started on heparin, which was stopped 48 h prior to TEG. Native unmodified venous blood (3.5 ml) was collected using a clean venepuncture at the antecubital fossa and assayed within 4 min of drawing the sample, in accordance to manufacturer recommendations. Following are the TEG parameters^{16, 17, 18} that were recorded.

The R time indicates the latency period till initial fibrin formation. It is measured from start of tracing to 1 mm divergence. The K time indicates the speed to reach 20 mm amplitude in the tracing. The alpha angle measures the rapidity of fibrin build-up and cross-linking. It represents the fibrinogen level. It is measured between the midline of the tracing and a line from 1-mm point tangential to the curve. The MA is a direct function of the maximum dynamic properties of fibrin and platelet bonding via GPIIb/IIIa. It represents maximum platelet function. It is the width of the curve at the widest part. The Coagulation Index (CI) is a linear combination of R, K, alpha, and MA. CI is given by the formula −0.1227 × R + 0.0092 × K + 0.1655 MA - 0.0241 × A - 5.022. LY30 is the measure of the rate of amplitude reduction 30 min after MA.

Presence of any one of the following: CI >3, R < 2 min, K < 1 min, alpha > 78, or MA > 69 mm were considered as hypercoagulant.^18,19 CI < −3, R > 8 min, K > 3 min, alpha < 55, or MA < 51 mm were considered as hypocoagulant. LY30 > 8% was considered as hyperfibrinolysis.^{16, 17, 18}

TEG parameters were compared with platelet count and INR, Rotterdam class, and presence or absence of thrombophilic disorder.

Informed consent in writing was obtained from each patient, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the institutional review board.

Statistical Analysis

Baseline characteristics of patients were studied using descriptive statistics. Dichotomous variables were compared using Chi-square test or Fischer exact test, and continuous variables were compared using nonparametric tests. Pearson's correlation was used to compare TEG results with platelet and INR. Results of TEG were compared to results of thrombophilia workup and with severity of their underlying disease as assessed by Rotterdam score. Analysis was performed by SPSS, v 16 (SPSS Inc, Chicago, IL, USA).

Results

The baseline characteristics of the patients are outlined in Table 1.

Table 1.

Baseline Characteristics of Patients.

Age (mean [SD])	32.3 [10.7] years
Sex	23 men; 28 women
Duration of symptoms (median [range])	5 months [1 month to 13 years]
Clinical features (n [%])
Ascites	30 (58.8)
Gastrointestinal bleeding	8 (15.6)
Hepatomegaly	31 (60.7)
Splenomegaly	18 (35.2)
Hepatic encephalopathy	0
Site of obstruction (n [%])
1) HV	40 (78.4)
2) IVC	3 (5.8)
3) HV + IVC	8 (15.6)
Type of presentation (n [%])
Acute	14 (27.4)
Subacute	16 (36.5)
Chronic	21 (41.1)
Rotterdam class (n [%])
Class I	22 (43.1)
Class II	27 (52.9)
Class III	2 (3.9)
Lab parameters (mean [SD])
Bilirubin (mg/dL)	1.64 (1.56)
AST (IU/L)	49.5 (22.3)
ALT (IU/L)	32.3 (20.2)
Albumin (g/dL)	3.7 (0.57)
INR	1.31 (0.28)

IVC, Inferior Vena Cava; INR: International Normalized Ratio; HV, Hepatic vein; AST, Aspartate aminotransferase; ALT, Alanine aminotransferase.

Thrombophilic states were seen in 21 patients, with 7 having two coexisting conditions; these included MTHFR mutation (n = 9), antiphospholipid syndrome (n = 7), factor V Leiden mutation (n = 5), myeloproliferative disease (n = 3), hyperhomocysteinemia (n = 2), and protein C deficiency (n = 2).

Eleven patients had procoagulant TEG, and 11 (21%) had a hypocoagulant TEG with CI < −3, and 29(57%) patients showed a normocoagulant TEG with −3 < CI < 3. Figure 1 shows TEG tracings in patients with normocoagulant, hypercoagulant, and hypocoagulant states.

Figure 1.

Typical TEG tracings. (A) Normocoagulant TEG. (B) Hypercoagulant TEG. (C) Hypocoagulant TEG. CI, Coagulation Index; TEG, Thromboelastography.

Among the 11 patients with at least one procoagulant TEG parameter, MA > 69 mm was present in 7 patients (63.6%), CI > 3 in 3/11 patients (27.3%), R < 2 min in 2 patients (18.2%), K < 1 min in 2 patients (18.2%), and alpha angle > 78 in none.

Among those with hypocoagulant TEG, abnormal parameters were CI < −3 in 11 patients (100%), R >8 min in 6 patients (54%), K > 3 min in 9 patients (81%), alpha < 55° in 9 patients (81%), and MA < 51 mm in 7 patients (63%).

Out of 11 patients with positive TEG, 7 had no associated thrombophilic condition while 4 had one or more associated thrombophilic states—(1) MTHFR heterozygous mutation, (2) hyperhomocysteinemia, (3) MTHFR heterozygous with factor V Leiden, and (4) MTHFR heterozygous with anti-phospholipid antibody syndrome (APLA). Thrombophilic state was not assessed in one patient with procoagulant TEG.

The platelet count and INR were similar in patients with procoagulant, normal, and hypocoagulant TEG (Table 2).

Table 2.

TEG Results in Patients With or Without Thrombophilia.

TEG classification	Thrombophilic condition identified (n)	No thrombophilic condition identified (n)	Total	Mean (SD) platelet count(105/mm3)	Mean (SD) INR
Hypercoagulant TEG	4	7	11	1.75 ± 0.78	1.27 ± 0.14
Normal TEG	12	17	29	2.22 ± 0.83	1.34 ± 0.33
Hypocoagulant TEG	7	4	11	1.79 ± 0.69	1.28 ± 0.24
Total	23	28	51

INR: International Normalized Ratio; TEG, Thromboelastography; SD, Standard Deviation.

There was no correlation between the various TEG parameters and platelet count and INR (Table 3).

Table 3.

Correlation Between Conventional Clotting Parameters and TEG.

TEG classification	R	K	Angle	MA	EPL	CI	LY30
Hypercoagulant TEG
Platelet count	0.21	0.01	−0.06	0.33	0.29	−0.05	0.22
INR	−0.08	0.03	−0.01	−0.07	−0.54	0.02	−0.52
Normocoagulant TEG
Platelet count	−0.12	0.08	−0.01	−0.06	−0.13	0.03	−0.12
INR	−0.08	−0.17	0.35	−0.01	0.29	0.14	−0.04
Hypocoagulant TEG
Platelet count	−0.77∗	−0.52	0.25	0.03	0.46	0.60∗	0.46
INR	0.18	0.02	0.16	−0.23	0.05	−0.12	0.04
Overall
Platelet Count	−0.07	−0.03	0.04	0.01	−0.04	0.04	−0.04
INR	−0.01	−0.04	0.20	−0.04	−0.13	0.02	−0.15

CI, Coagulation Index; INR, International Normalized Ratio; TEG, Thromboelastography; EPL, Estimated percent lysis.

The values in Table 3 are Pearson's coefficient of correlation.

^∗P < 0.05.

There was no difference in TEG parameters between Rotterdam class I and Rotterdam class II and III combined (Table 4, Figure 2). Mean LY30 was higher in Rotterdam class III (41.0) as compared to class I (2.93) and II (3.8).

Table 4.

Comparison of TEG Parameters and Rotterdam Class I Vs Class II and III.

TEG classification	R time	K time	Alpha	MA	CI	EPL	LY30
Rotterdam class-I (n = 22)	7.5 ± 5.5	3.3 ± 4.4	54.7 ± 16.1	58.6 ± 14.7	−2.3 ± 7	3.4 ± 13	3.8 ± 13
Rotterdam Class-II, III (n = 29)	6.5 ± 3.2	2.9 ± 1.9	52.8 ± 16.3	55.6 ± 16.1	−1.6 ± 4	4.9 ± 16.1	5.6 ± 16.1
P	0.81	0.65	0.67	0.48	0.64	0.71	0.65
Thrombophilic disorder (n = 23)	7.31 ± 5.4	3.5 ± 4.2	52.2 ± 15.2	56.3 ± 16.7	−2.54 ± 7	5.23 ± 17.2	5.46 ± 16.3
No thrombophilic disorder (=28)	6.5 ± 3.4	2.73 ± 2	53.4 ± 17.7	57.2 ± 14.9	−1.35 ± 3.9	3.44 ± 12.9	4.36 ± 13.9
P	0.52	0.4	0.96	0.83	0.45	0.67	0.79

SD, Standard Deviation.

Values are as mean ± SD.

Figure 2.

Scatter plot of TEG parameter with Rotterdam class. TEG, Thromboelastography.

There was no difference in TEG parameters between patients with underlying thrombophilic conditions and those without thrombophilic conditions (Table 4). There was no correlation between TEG parameters and Child–Turcotte–Pugh and model for end-stage liver disease status.

The sensitivity, specificity, positive predictive value, and negative predictive value of TEG to predict thrombophilia were 17.3%, 75%, 36.3%, and 52.5%, respectively.

Discussion

Our study shows that more than half the patients with BCS have normocoagulant and one-fifth have hypocoagulant TEG parameters. INR, platelet count, thrombophilic disorders, and disease severity scores were similar in patients with normal, hypercoagulant, and hypocoagulant TEG.

BCS is considered to be a prothrombotic state. The most frequent cause of BCS is thrombophilic disorders, which is detected in up to 84% of patients.^{1, 2, 3, 4, 5, 19} Amongst these, myeloproliferative diseases and factor V Leiden mutations are the most common.^{3, 20} In over 25% of cases, more than one thrombophilic state may be present.⁶ In this study, a thrombophilic state was found only in 21 (41%) of the patients which is lesser than reported in other series.^{1, 2, 3, 4, 5, 6}

The coagulation status in a patient with BCS is affected by multiple factors. The liver dysfunction due to BCS and presence of thrombophilic disorder complicates the balance of procoagulant and anticoagulant factors. TEG is used to assess comprehensive global coagulation in the whole blood and provides a better overview of hemostasis than conventional coagulation tests. Therefore, we used TEG to assess coagulation status in patients with BCS in this study.

In patients with other liver diseases, conventional tests such as platelet count, INR, and Partial Thromboplastin Time (PTT) have been found to be unreliable in predicting coagulation status and risk of bleeding. Defects in platelet number and function, procoagulant function, and regulation of fibrinolysis appear balanced to some extent by additional changes in the hemostatic system. Tests like INR and PTT do not represent the balance between the procoagulant and anticoagulant proteins because they are not sensitive to detect deficiencies of anticoagulants.⁸ Tripodi et al. were the first to demonstrate that thrombin generation in patients with cirrhosis is not different from that in healthy volunteers, provided thrombomodulin was added to the test mixture.⁹ The defects in platelet number and function are compensated by substantially elevated levels of the platelet adhesive protein VWF. The fibrinolytic system may also be in balance in patients with cirrhosis due to the concomitant decrease of antifibrinolytics (antiplasmin, thrombin activatable fibrinolysis inhibitor) and plasminogen.⁸ Similarly, in our study, the INR and platelet count did not correlate with the findings of TEG.

TEG has also been assessed to detect thrombophilic conditions in nonliver-related conditions. In a study of 87 patients with personal or family history of thrombosis, only 17 out of 39 patients with hypercoagulable TEG had a demonstrable thrombophilic condition. Conversely, of 30 patients with proven thrombophilia, 17 also had a hypercoagulable TEG, and it failed to identify 43% of underlying thrombophilic conditions.²¹ TEG has been used to detect thrombophilic condition in pregnancy,²² predicting postoperative thrombosis, postoperative myocardial infarction,²³ and prediction of unfavorable outcome of ischemic stroke.²⁴ An established indication of TEG in liver disease is to guide factor repletion and fibrinolytic therapy during liver transplantation. The use of TEG to detect thrombophilic conditions has not been attempted previously in BCS. In our study, TEG findings were similar in patients with and without identified thrombophilic state. The use of TEG to guide antithrombotic therapy has been reported in BCS in a case report.²⁵ TEG is also used to guide blood product replacement in cirrhotic patients and during transplantation. Correction of TEG parameters include 2 units of plasma for an r-time greater than 15 min, 10 units of platelets for a maximum amplitude less than 40 mm, and 6 units of cryoprecipitate for an α-angle less than 40–45° .^26,27 Pietri et al demonstrated that use of TEG led to 83% reduction in requirement of blood requirements prior to invasive procedures without any significant adverse events post procedure.¹¹

TEG is useful to predict bleeding episodes in cirrhosis. Bleeding episodes are frequent complications in BCS with reported frequency of 22.8 per 100 patient-years.²⁸ This is higher than bleeding rates in patients with atrial fibrillation on anticoagulation (2.45 [1.64–3.65] per 100 patient-years).²⁹ In BCS, excess anticoagulation was identified in only 27% as a cause of bleeding.³⁰ A hypocoagulable TEG has been shown to be superior to INR, PTT, and platelet count for estimating the risk of variceal rebleeding.³⁰ Presence of hypocoagulant status at baseline in a subgroup of patients (as demonstrated in our study) could be one of the contributing reasons for high incidence of bleeding in patients with BCS despite normal INR and platelet count.³¹

We excluded patients who had sepsis as TEG parameters are affected by presence of sepsis. Patients with decompensated cirrhosis and infection have been found to have a more hypocoagulable TEG, driven by endogenous heparin-like substances. But conventional tests of coagulation did not differ between the infected and noninfected in patients with cirrhosis.^{32, 33} TEG parameters may normalize on treatment of infections. This may reflect better sensitivity of TEG for overall hemostasis.

Our study suggests on the basis of TEG that two-thirds of the patients with BCS have a normal CI. Hypercoagulable TEG was detected in 22% of patients in our study. We could not find any study using TEG to evaluate coagulation in a setting of prothrombotic disorder with liver disease. However, Ben-Ari et al. demonstrated hypercoagulability in 28% of primary biliary cirrhosis, 43% of primary sclerosing cholangitis, 5% of noncholestatic cirrhosis, and none in the healthy controls using TEG.³⁴ In a cohort of 273 patients with stable cirrhosis, TEG parameters were within normal limits, although the maximum amplitude was decreased in proportion to the severity of thrombocytopenia.¹⁶ A normal TEG in a majority of cirrhotics supports the concept of rebalanced hemostasis,³⁰ and our study provides evidence of the same in BCS.

In our study, patients with hypocoagulant, hypercoagulant, and normocoagulant TEG had similar platelet count and INR. Thus, these tests are unreliable in predicting the underlying coagulation status in BCS.

Hyperfibrinolysis was seen in 1 patient with Rotterdam class III severity. This finding may demand a more cautious approach to anticoagulation in these patients. Petri et al did not find hyperfibrinolysis in any of the 30 patients with cirrhosis undergoing TEG.¹¹ However, due to a small sample size of Rotterdam class III (n = 2) in our study, an alpha error maybe present.

No other differences in TEG parameters were observed across different Rotterdam class. In another study, decompensated cirrhotics had a lower mean MA (51.5 ± 10.4 vs 45.0 ± 9.9) and alpha angle (62.6 ± 9.3 vs 58.1 ± 10.8) as compared to compensated cirrhotics.¹⁶

In a systematic review of use of TEG to predict postoperative thrombotic events, the sensitivity and specificity ranged from 0% to 100% and 62%–92%, respectively, and area under the curve of the receiver operating characteristic (AUROC) ranged from 0.57 to 0.91.³⁵ Our study shows that TEG lacks sensitivity and specificity to predict thrombophilic conditions in BCS.

In our study, four TEG parameters were measured as a representation for hemostasis: R (reaction time); time for initial fibrin threads formation, α (alpha angle); rate of fibrin formation, K time; speed of fibrin formation, MA (maximum amplitude); the maximum strength of the clot and a reflection of the structural interactions and fibrinogen, interlaced with fibrin polymers as well as platelet function, and CI was calculated. Coagulation index is a mathematical formula given by the manufacturer which takes into account the relative contribution of the components of TEG to the coagulation. The procoagulant status has been defined variably in literature and among all the parameters including CI, MA is shown to be most predictive of postoperative thrombotic event.^{35, 36} Our study also found that the MA was the most common procoagulant TEG parameter abnormality.

Since, there is no data available for BCS which is a unique disease due to a combination of thrombophilic disorders and liver dysfunction, we defined hypercoagulability as presence of one or more abnormal TEG parameters.

This is the first study to describe the TEG parameters in patients with BCS. It is a proof of concept study that TEG depicts the complex balance of procoagulants and anticoagulants due to liver dysfunction and thrombophilic disorder in BCS. Our study also demonstrates that a proportion of patients with BCS have a hypocoagulant status as found on the basis of TEG, contrary to the popular belief that all patients with BCS are hypercoagulant.

Conclusions

This is the pilot study of assessment of coagulation status using TEG in patients of BCS. Patients of BCS have varied coagulation status with one-fifth having hypocoagulant TEG. Patients with advanced disease may have accelerated fibrinolysis. Further studies are needed to assess the significance of TEG in predicting prognosis and complications of therapy as well as its utility for long-term monitoring of patient with BCS.

Conflicts of interest

The authors have none to declare.