Cause of Portal or Hepatic Venous Thrombosis in Adults: The Role of Multiple Concurrent Factors

Transcription

1 Cause of Portal or Hepatic Venous Thrombosis in Adults: The Role of Multiple Concurrent Factors MARIE-HÉLÈNE DENNINGER, 1 YASMINE CHAïT, 2 NICOLE CASADEVALL, 3 SOPHIE HILLAIRE, 4 MARIE-CLAUDE GUILLIN, 1 ANNIE BEZEAUD, 1 SERGE ERLINGER, 4 JEAN BRIERE, 2 AND DOMINIQUE VALLA 4 According to a recent hypothesis, venous thrombosis results from the concurrence of several factors. This hypothesis was assessed in patients with portal or hepatic venous thrombosis by simultaneously investigating most of the currently identified prothrombotic disorders, local precipitating factors, and other risk factors such as oral contraceptive use. Patients with a tumorous obstruction and patients with cirrhosis with portal vein thrombosis were excluded. The prothrombotic disorders that were investigated included classical and occult myeloproliferative disorders; antiphospholipid syndrome; protein C; protein S and antithrombin deficiency; factor V Leiden; factor II; and methylene-tetrahydrofolate-reductase gene mutations. We found 1 or several prothrombotic disorders and a local precipitating factor in 26 and 10 of the 36 patients with portal vein thrombosis, respectively; and in 28 and none of the 32 patients with hepatic vein thrombosis, respectively. We found a combination of prothrombotic disorders in 5 and 9 patients with portal and hepatic vein thrombosis, respectively, whereas such a combination is expected in less than 1% of asymptomatic subjects. Of the 10 patients with a local precipitating factor, 8 had a prothrombotic disorder. Of the 13 patients who use oral contraceptives, 10 had a prothrombotic disorder. We conclude that portal or hepatic venous thrombosis should be regarded as an index for 1 or several prothrombotic disorders, whether or not local precipitating factors or oral contraceptive use are found. Concurrence of prothrombotic disorders is more common than expected. Extensive investigation of prothrombotic disorders and anticoagulation should be considered in patients with portal or hepatic venous thrombosis. (HEPATOLOGY 2000;31: ) According to a recent concept, venous thrombosis would result from the convergence of an inherited predisposition, owing to a mutation in 1 or more genes, and an acquired Abbreviation: MTHFR, methylene-tetrahydrofolate reductase. From the 1 Laboratoire d Hématologie et d Immunologie, 2 Service d Hématologie Clinique, and 4 Service d Hépatologie and Laboratoire d Hémodynamique Splanchnique et de Biologie Vasculaire, Hôpital Beaujon, Clichy, France; and 3 Laboratoire d hématologie, Hôpital Raymond Poincarré, Garches, France. Received May 25, 1999; accepted December 10, Supported by Délégation à la recherche clinique de l Assistance publique, Hôpitaux de Paris (CRC ). Address reprint requests to: Dominique Valla, M.D., Service d Hépatologie, Hôpital Beaujon, Clichy, France. dominique.valla@bjn.ap-hop-paris.fr; fax: Copyright 2000 by the American Association for the Study of Liver Diseases /00/ $3.00/0 587 thrombogenic stimulus. 1 However, this concept has not been evaluated in the setting of hepatic or portal venous thrombosis. This concept, if validated, would be of major consequence for the investigation and treatment of patients. Portal or hepatic venous thromboses have been commonly found to be associated with acquired prothrombotic disorders 2-7 but inherited disorders have not been extensively investigated. Therefore, the study reported here was undertaken to ascertain the prevalence of newly recognized inherited prothrombotic disorders in patients with hepatic or portal vein thrombosis, 8-10 and also to assess to what extent inherited as well as acquired prothrombotic disorders, local precipitating factors, and other risk factors for thrombosis concur to the development of hepatic or portal venous thrombosis. PATIENTS AND METHODS Patients with portal or hepatic vein thrombosis diagnosed on usual criteria, 5,6 and seen between January 1994 and June 1997, were included when there was no evidence of cancer of the liver, pancreas, or bile ducts at imaging studies and follow-up. Some of the patients had been included in previously reported studies 2-5 in which only a few specific causes were investigated. In patients with portal vein thrombosis, cirrhosis was excluded either by normal results of imaging and laboratory studies, or by liver biopsy. Tumors, infectious or inflammatory disorders, as well as malformative lesions in direct connection with portal, mesenteric, splenic, or hepatic veins, were considered to be a possible precipitating local factor for thrombosis only when their acute stage coincided with recent portal venous thrombosis. Portal venous thrombosis was considered to be recent when there was abdominal pain of recent onset, no sign of chronic portal hypertension, and no evidence of portal cavernoma. We regarded primary myeloproliferative disorders, paroxysmal nocturnal hemoglobinuria, primary deficiencies in natural coagulation inhibitors, G1691A factor V gene (Leiden) mutation, G20210A prothrombin gene mutation, and C677T methylene-tetrahydrofolate reductase (MTHFR) gene mutations as prothombotic disorders. Oral contraceptive use and pregnancy were considered thrombotic risk factors but were not regarded as prothrombotic disorders. Prothrombotic disorders were looked for in every patient using previously reported methods and criteria as follows. Patients were studied at least 1 month after a thrombotic episode. Endogenous erythroid colonies and the total red cell mass were assessed as previously described. 2,3 Overt polycythemia vera was characterized by a hematocrit greater than 125% of calculated normal value and suggestive bone marrow findings 11,12 ; covert polycythemia vera was characterized by normal or low peripheral hematocrit value but a red cell mass greater than 125% of calculated normal value after correction of iron deficiency, and by suggestive bone marrow findings 11,12 ; the so-called forme fruste of primary myeloproliferative disorder was characterized by endogenous erythroid colonies, by normal or decreased total red cell volume with normal iron stores, and by the absence of other well-characterized myeloproliferative disorders. 13

2 588 DENNINGER ET AL. HEPATOLOGY March 2000 Evaluation of the CD55-CD59 cluster was performed to rule out paroxysmal nocturnal hemoglobinuria. 14 Antithrombin and protein C functional deficiencies and total and/or free protein S antigenic deficiencies were assessed using commercial kits (Diagnostica Stago, Asnieres, France) and only considered primary when isolated, that is, when the levels of other inhibitors or clotting factors were normal so that hepatocellular insufficiency could be ruled out as a cause for a secondary deficiency in these coagulation inhibitors. A commercial kit was used to assay antiphospholipid antibodies of immunoglobulin G (IgG) and IgM types (Biogenic, Montpellier, France). Lupus anticoagulant was investigated as previously described. 15 Genetic analyses were performed according to previously reported methods for G1691A factor V, 8 G20210A prothrombin, 9 and C677T (MTHFR) gene mutations. 10 All patients gave informed consent to the investigations that were performed as part of the routine treatment of portal or hepatic venous thrombosis. RESULTS Portal Vein Thrombosis. There were 58 patients with portal vein thrombosis, 36 of whom could be completely investigated because adequate blood and bone marrow samples were available. In 13 of the 36 patients, decreased levels of several coagulation inhibitors were associated with decreased levels of several clotting factors, suggesting multiple acquired deficiencies owing to impaired liver synthesis or activation of coagulation. In these 13 patients, and for conservative purposes, the decreased levels of coagulation inhibitors were not considered as prothrombotic coagulation disorders. The prevalences of causal factors are presented in Tables 1 and 2. Paroxysmal nocturnal hemoglobinuria was not found in any patients. Myeloproliferative disorders were found in 11 patients, 5 of whom had an overt form and 6 a forme fruste. In 3 of these 11 patients, the myeloproliferative disorder was associated with 1 or several prothrombotic coagulation disorders. These combinations are set out in Table 3. In 15 other patients, 1 or several prothrombotic coagulation disorders were evidenced in the absence of myeloproliferative disorder. Of these 15 patients, a single prothrombotic coagulation TABLE 1. Prevalence of Factors That Were Simultaneously Investigated in 36 Patients With Portal Vein Thrombosis and 32 Patients With Hepatic Vein Thrombosis Portal Vein Thrombosis (36 Patients) Hepatic Vein Thrombosis (32 Patients) n % 95% CI n % 95% CI Total number of patients with prothrombotic disorder Primary myeloproliferative disorder Prothrombotic coagulation disorder (1 or more) Primary myeloproliferative disorder plus prothrombotic coagulation disorder Total number of patients without prothrombotic disorder 10 4 Local factor alone 1 0 Oral contraceptive use alone 0 2 Local factor oral contraceptive use 1 0 No predisposing factors 8 2 TABLE 2. Specific Prothrombotic Coagulation Disorders Portal Vein Thrombosis (36 Patients) Hepatic Vein Thrombosis (32 Patients) n % 95% CI n % 95% CI Antiphospholipid antibodies G1691A factor V gene mutation G20210A factor II gene mutation T677T MTHFR gene mutation Protein S deficiency * * 0 Protein C deficiency 0 0* * 0-14 Antithrombin deficiency 1 4.5* * 0 *Percent of patients without decreased level of coagulation factors. disorder could be evidenced in 13, whereas several prothrombotic coagulation disorders were associated in 2 (Table 3). Only 1 incidence of coagulation inhibitor deficiency (antithrombin) could be shown as inherited by a familial study. Thus, 1 or several prothrombotic disorders were evidenced in 26 of the 36 patients with portal vein thrombosis. Among the 10 patients in which no prothrombotic disorder could be identified, a local precipitating factor was found in 2. A local precipitating factor was also identified in 8 of the 26 patients with a prothrombotic disorder. Thus, a local precipitating factor was identified in 10 of the 36 patients with portal vein thrombosis (splenectomy, Bacteroides pylephlebitis, and cholecystectomy in 2 patients each; umbilical cannulation, abdominal trauma, colon adenocarcinoma, and intestinal obstruction in 1 patient each). Among the 14 child-bearing aged women, 3 used oral contraceptives, 2 of whom had a prothrombotic disorder and 1 had a local precipitating factor. There was no significant difference in the respective prevalences (number positive/number investigated) of individual prothrombotic disorders, local precipitating factors, and other risk factors for thrombosis in the group of 36 patients who could be extensively investigated and in the whole group of 58 patients (data not shown). Hepatic Vein Thrombosis. There were 44 patients with hepatic vein thrombosis, 32 of whom could be completely investigated. In 23 of the 32 patients, decreased levels of several coagulation inhibitors were associated with decreased levels of several clotting factors, suggesting multiple acquired deficiencies owing to impaired liver synthesis or activation of TABLE 3. Combination of Prothrombotic Disorders, Risk Factors for Thrombosis, and Local Factors in Patients With Portal Vein Thrombosis Patient PVT 1 PVT 2 PVT 5 PVT 12 PVT 17 PVT 19 PVT 25 Primary myeloproliferative disorder, G1691A factor V gene mutation, G20210A factor II gene mutation, cholecystectomy protein S deficiency, schistosomiasis, splenectomy Primary myeloproliferative disorder, protein S deficiency, splenectomy Antiphospholipid antibodies, oral contraceptive use, Bacteroides pylephlebitis G20210A factor II gene mutation, protein S deficiency, resection of colon adenocarcinoma C677T MTHFR gene mutation, protein S deficiency G20210A factor II gene mutation, recent CMV infection, oral contraceptive use

3 HEPATOLOGY Vol. 31, No. 3, 2000 DENNINGER ET AL. 589 coagulation. In these 23 patients, and for conservative purposes, the decreased levels of coagulation inhibitors were not considered as prothrombotic coagulation disorders. The prevalence of causal factors is presented in Table 1. Paroxysmal nocturnal hemoglobinuria was not found in any patients. Myeloproliferative disorders were found in 17 patients, 12 with overt forms and 5 with a forme fruste. In 7 of these 17 patients, the myeloproliferative disorder was associated with 1 or several prothrombotic coagulation disorders (Table 4). In 11 other patients, 1 or several prothrombotic coagulation disorders were evidenced in the absence of myeloproliferative disorder. Of these 11 patients, a single prothrombotic coagulation disorder could be evidenced in 9, whereas associated prothrombotic coagulation disorders were evidenced in 2. These combinations are set out in Table 4. None of the 2 patients with protein C deficiency could be shown as inherited. Overall, 1 or several prothrombotic disorders were evidenced in 28 of the 32 patients with hepatic vein thrombosis. As shown in Table 4, there were 2 associated prothrombotic disorders in 8 patients and 3 associated disorders in 1 patient. Local precipitating factors were not identified in any of the 32 patients. Among the 20 child-bearing aged women, 10 used oral contraceptives and 2 were pregnant. Eight of these 12 women had a prothrombotic disorder. Two of the 4 patients without prothrombotic precipitating disorder used oral contraceptives. There was no significant difference in the respective prevalences (number positive versus number investigated) of individual prothrombotic disorders, local precipitating factors, and other risk factors for thrombosis in the group of 32 patients who could be extensively investigated and in the whole group of 44 patients (data not shown). DISCUSSION The first aim of our study was to ascertain the role of individual prothrombotic disorders in the cause of portal and hepatic venous thromboses. One or several prothrombotic disorders were found in 26 of 36 patients (72%) with portal vein thrombosis and in 28 of 32 patients (87%) with hepatic vein thrombosis. Primary myeloproliferative disorders were TABLE 4. Combination of Prothrombotic Disorders, Risk Factors for Thrombosis, and Local Factors in Patients With Hepatic Vein Thrombosis Patient HVT 1 HVT 2 HVT 3 HVT 4 HVT 7 HVT 11 HVT 16 HVT 26 HVT 28 HVT 18 HVT 20 Primary myeloproliferative disorder, G1691A factor V gene mutation Primary myeloproliferative disorder, protein C deficiency, oral contraceptive use protein C deficiency, celiac disease, oral contraceptive use Primary myeloproliferative disorder, G1691A factor V gene mutation oral contraceptive use Primary myeloproliferative disorder, G20210A factor II gene mutation Primary myeloproliferative disorder, antiphospholipid antibodies C677T MTHFR gene mutation, oral contraceptive use C677T MTHFR gene mutation, oral contraceptive use G1691A factor V gene mutation, antiphospholipid antibodies G1691A factor V gene mutation, G20210 factor II gene mutation found to be the leading causal factor in portal and hepatic vein thromboses with a prevalence of 30% and 50%, respectively, in close agreement with the results of previous studies reported by us 2,3 and confirmed by other investigators. 16 Our findings confirm that in a relatively large series of patients, factor V Leiden is found in patients with hepatic vein thrombosis as frequently as in patients with deep vein thrombosis (about 25%), 17 but it is not more frequent in patients with portal vein thrombosis than in healthy white European patients (about 5%). 1,18 G20210A factor II gene mutation was found in 14% of patients (95% confidence interval [CI]; 3% to 25%) with portal venous thrombosis, a figure not significantly different from that found in hepatic venous thrombosis (6%, 95% CI; 0% to 14%) and similar to that found in a recently reported series of 20 patients with portal vein thrombosis (20%, 95% CI; 3% to 43%). 20 These prevalences are close to those found in patients with venous thrombosis (8% to 18%) and higher than those encountered in healthy control patients (1% to 4.8%). 1,9,19,20 In the present study, the prevalence of homozygous C677T MTHFR gene mutation was similar to that found in a healthy European population (about 10%), 10 which suggests that, alone, this mutation is not a strong prothrombotic factor. This suggestion is in accordance with results of recent studies indicating that whereas hyperhomocysteinemia is a risk factor for venous thromboembolism and vascular disease, C677T MTHFR mutation, per se, is only 1 of the multiple determinants of hyperhomocysteinemia and is not a risk factor for vascular disease. 21,22 Isolated protein S deficiency was most common in patients with portal vein thrombosis and isolated protein C deficiency was most common in patients with hepatic vein thrombosis. However, family studies were not able to precisely identify the inherited character of the defect, usually because all informative relatives could not be tested. However, it should be noted that our estimation of the prevalence of primary deficiencies is conservative. Indeed we have not considered the combined deficiencies in natural coagulation inhibitors and clotting factors as primary deficiencies because impaired liver function or coagulation activation could explain them. 23 Therefore, there is a possibility that some patients with primary coagulation inhibitor deficiency were missed because of superimposed deficiency secondary to liver function impairment or coagulation activation. Molecular studies will be necessary to address this issue. There have been several anecdotal case reports of portal or hepatic venous thromboses associated with antiphospholipid syndrome, but little systematic study of this potential cause. 5,19 In the present study, based on the association of antiphospholipid antibodies and deep vein thrombosis, the antiphospholipid syndrome could be considered in 19% of patients with hepatic vein thrombosis, but also in 11% of patients with portal vein thrombosis. In summary, although none of the different Western studies of hepatic and portal vein venous thromboses included large numbers of patients, there is much agreement in their results about the following points: frequent association with primary myeloproliferative disorders, frequent occurrence of the primary myeloproliferative disorders in an occult form, a prevalence of factor V Leiden that is higher than expected in the general population in hepatic vein thrombosis but not in portal vein thrombosis, and a high prevalence of protein C deficiency in patients with hepatic vein thrombosis and of

4 590 DENNINGER ET AL. HEPATOLOGY March 2000 protein S deficiency in patients with portal vein thrombosis. The small amount of patients studied do not allow yet, to be certain that portal vein and hepatic veins really differ with regard to the causes of thrombosis. Adding the results of the investigations for prothrombotic disorders to those of investigations for other factors, we succeeded in identifying a possible causal factor in a high proportion of patients. One or several causal factors could be identified in 28 of 36 patients (78%) with portal vein thrombosis, 26 of whom had a prothrombotic disorder. Likewise, 1 or several causal factors could be identified in 30 of 32 patients (94%) with hepatic vein thrombosis, 28 of whom had a prothrombotic disorder. The second aim of this study was to test the hypothesis that several causal factors need to be associated for portal or hepatic venous thrombosis to develop. Several prothrombotic disorders were combined in 5 patients (14%) with portal vein thrombosis and in 9 patients (28%) with hepatic vein thrombosis. Primary myeloproliferative disorders were associated with coagulation disorders in 3 patients (8%) with portal vein thrombosis and 7 patients (22%) with hepatic vein thrombosis. Combination with oral contraceptive use was further documented in 25% of the patients. The prevalence of individual prothrombotic disorders in Europe ranges from 5% to 8% (factor II or factor V gene mutation) to much less than that. 1 Therefore, the probability of combined disorders in an unselected population is far less than 0.7%, a value 30- to 40-fold lower than that observed in patients with portal or hepatic venous thrombosis. Thus, our findings support the hypothesis that a combination of several prothrombotic factors favors the occurrence of hepatic or portal venous thrombosis. Such a combination would nicely explain why only some patients with primary myeloproliferative disorders 24 or antiphospholipid antibodies suffer from portal or hepatic venous thrombosis. 19 Indeed, in the present study, 50% of patients with primary myeloproliferative diseases and about 30% of patients with antiphospholipid antibodies had another prothrombotic disorder or used oral contraceptives. However, in many patients, a single prothrombotic disorder could be evidenced; this was particularly well documented in the group of 23 patients with portal vein thrombosis and no evidence of impaired liver synthesis or coagulation activation, in whom natural coagulation inhibitors could be reliably assessed. Fifteen of these 23 patients were found to have only 1 prothrombotic disorder. In these patients with a single prothrombotic disorder, thrombosis might still be explained by the association to 1 or several yet unknown prothrombotic disorders. Alternatively, failure to identify additional prothrombotic disorders could mean that hepatic or portal venous thrombosis can occur in a patient with a single prothrombotic state, or even no prothrombotic state, for example, because local factors play an important role. Indeed, the latter appear to play an ancillary role in the development of portal vein thrombosis, being found in about 25% of patients, usually in combination with a prothrombotic disorder. By contrast, local precipitating factors were not found in most patients with hepatic vein thrombosis. The latter finding implies either that no local factor is necessary for triggering hepatic venous thrombosis or that the local factors are difficult to identify. The high prevalence of localized hepatic vein stenoses in patients with prothrombotic disorders favors the second interpretation. 25 Compression by the diaphragm during deep breathing might be the mechanical factor responsible for thrombosis of the hepatic vein or suprahepatic inferior vena cava. 25,26 Taken together, these findings invite one to regard splanchnic vein thrombosis as an index of 1 or several prothrombotic disorders, whether or not oral contraceptives are used, thrombosis occurs in pregnancy or postpartum, and a local precipitating factor can be found. This conclusion raises 2 issues. First, what is the benefit-risk ratio of anticoagulation in patients prone both to thrombosis and to bleeding from portal hypertension? Controlled studies, although difficult to perform in such uncommon conditions as hepatic or portal venous thromboses, are badly needed. Second, what is the benefit of extensive screening for underlying prothrombotic disorders if anticoagulation is to be administered to every patient? We believe that the potential for specific treatment in primary myeloproliferative diseases and antiphospholipid syndromes, as well as the potential for decreasing the risk of thrombosis in relatives with inherited coagulation disorders, argues for extensive screening. REFERENCES 1. Shafer AI. Hypercoagulable states: molecular genetics to clinical practice. Lancet 1994;344: Valla D, Casadevall N, Huisse MG, Tulliez M, Grange JD, Muller O, Binda B, et al. Etiology of portal vein thrombosis in adults. A prospective evaluation of primary myeloproliferative disorders. Gastroenterology 1988;94: Valla D, Casadevall N, Lacombe C, Varet B, Goldwasser E, Franco D, Malliard J-N, et al. Primary myeloproliferative disorder and hepatic vein thrombosis. A prospective study of erythroid colony formation in vitro in patients with Budd-Chiari syndrome. Ann Intern Med 1985;103: Pelletier S, Landi B, Piette JC, Ekert P, Coutelier A, Fadlallah JP, Herson S, et al. Antiphospholipid syndrome as the second cause of nontumorous Budd-Chiari syndrome. J Hepatol 1994;21: Denninger MH, Helley D, Valla D, Guillin MC. Prospective evaluation of the prevalence of factor V Leiden mutation in portal or hepatic vein thrombosis. Thromb Haemost 1997;78: Van Gansbeke D, Avni EF, Delcour C, Engelholm L, Struyven J. Sonographic features of portal vein thrombosis. AJR Am J Roentgenol 1985;144: Valla D, Benhamou JP. Obstruction of the hepatic veins or suprahepatic inferior vena cava. Dig Dis 1996;14: Bertina RM, Koeleman RPC, Koster T, Rosendaal FR, Dirven RJ, de Ronde H, Van der Velpen PA, et al. Mutation in blood coagulation factor V associated with resistance to activated protein C. Nature 1994;369: Poort SR, Rosendaal FR, Reitsma PH, Bertina RM. A common genetic variation in the 3 untranslated region of the prothrombin gene is associated with elevated plasma prothrombin levels and an increase in venous thrombosis. Blood 1996;88: Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, Boers GJH, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nature Genet 1995;10: Berlin NI: Classification of the polycythemias and initial clinical features in polycythemia vera. In: Wasserman LR, Berk PD, Berlin NI, eds. Polycythemia Vera and the Myeloproliferative Disorders. Philadelphia: WB Saunders, 1995; Iland H, Laszlo J, Murphy S: Essential thrombocythemia. In: Wasserman LR, Berk PD, Berlin NI, eds. Polycythemia Vera and the Myeloproliferative Disorders. Philadelphia: WB Saunders, 1995; Reid CD, Chanarin I, Lewis J. Formes frustes in myeloproliferative disorders. Identification by the growth of an endogenous erythroid clone in vitro in patients with arterial vascular disease. Lancet 1982;1: Schubert J, Alvarado M, Uciechowski P, Zielinska-Skowronek M, Freud M, Vogt H, Schmidt RE. Diagnosis of paroxysmal nocturnal haemoglobinuria using immunophenotyping of peripheral blood cells. Br J Haematol 1991;79: Rosner E, Pauzner R, Lusky A, Modan M, Many A. Detection and quantitative evaluation of lupus circulating anticoagulant activity. Thromb Haemos 1987;57:

5 HEPATOLOGY Vol. 31, No. 3, 2000 DENNINGER ET AL De Stefano V, Teofili L, Leone G, Michiels JJ. Spontaneous erythroid colony formation as the clue to an underlying myeloproliferative disorder in patients with Budd-Chiari syndrome or portal vein thrombosis. Semin Thromb Hemost 1997;23: Zöller B, Hilm J, Dahlback B. Resistance to activated protein C due to a factor V gene mutation. The most common inherited risk factor of thrombosis. Trends Cardiovasc Med 1996;6: Mahmoud AEA, Elias E, Beauchamp N, Wilde JT. Prevalence of the factor V Leiden mutation in hepatic and portal vein thrombosis. Gut 1997;40: Chamouard P, Pencrach E, Maloisel F, Grunebaum L, Ardizzone JF, Meyer A, Gaub M-P, et al. Frequent factor II G20210 mutation in idiopathic portal vein thrombosis. Gastroenterology 1999;116: Maraglione M, Brancaccio V, Giuliani N, D Andrea G, Cappucci G, Iannaconne L, Vecchione G, et al. Increased risk for venous thrombosis in carriers of the prothombin G-A gene variant. Ann Intern Med 1998;129: Den Heijer M, Rosendaal, Blom HJ, Gerrits WBJ, Bos GMJ. Hyperhomocysteinemia and venous thrombosis: a meta-analysis. Thromb Haemost 1998;80: Wicken DEL. MTHFR 677C-T mutation, folate intake, neural tube defect, and risk of cardiovascular disease. Lancet 1997;350: Robson SC, Kahn D, Kruskal J, Bird AR, Kirsch RE. Disordered hemostasis in extrahepatic portal hypertension. HEPATOLOGY 1993;18: Wanless IR, Peterson P, Das A, Boitnott JK, Moore GM, Bernier V. Hepatic vascular disease and portal hypertension in polycythemia vera and agnogenic myeloid metaplasia: a clinicopathological study of 145 patients examined at autopsy. HEPATOLOGY 1990;12: Valla D, Hadengue A, El Younsi M, Azar N, Zeitoun G, Boudet MJ, Molas G, et al. Hepatic venous outflow block ( Budd Chiari syndrome ) due to short-length hepatic vein stenoses ( hepatic vein webs ). HEPATOLOGY 1997;25: Shrestha SM, Okuda K, Uchida T, Maharjan KG, Shresta S, Joshi B, Larson S, et al. Endemicity and clinical picture of liver disease due to obstruction of the hepatic portion of the inferior vena cava in Nepal. J Gastroenterol Hepatol 1996;11: