Portal Hypertension Related Complications After Acute Portal Vein Thrombosis: Impact of Early Anticoagulation

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2008;6: Portal Hypertension Related Complications After Acute Portal Vein Thrombosis: Impact of Early Anticoagulation JUAN TURNES,*, JUAN CARLOS GARCÍA PAGÁN,*, MONICA GONZÁLEZ,, CARLES ARACIL,, JOSÉ LUÍS CALLEJA, CRISTINA RIPOLL,,# JUAN G. ABRALDES,, RAFAEL BAÑARES,,# CÁNDIDO VILLANUEVA,, AGUSTÍN ALBILLOS,, JUAN RAMÓN AYUSO,** ROSA GILABERT,, ** and JAIME BOSCH*, *Hepatic Hemodynamic Laboratory, Liver Unit, Institut Malalties Digestives i Metabòliques, Hospital Clínic, Institut d Investigacions Biomèdiques August Pi i Sunyer, Barcelona; Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (Ciberehd), Barcelona; Department of Gastroenterology, Hospital Universitario Ramón y Cajal, University of Alcalá, Madrid; Department of Gastroenterology, Hospital Sant Pau, Barcelona; Clínica Puerta de Hierro, Madrid; # Hospital Gregorio Marañón, Madrid; and **Centre de Diagnostic per l Imatge, Hospital Clinic, Barcelona, Spain Background & Aims: Acute portal vein thrombosis (APVT) is a rare disorder that causes chronic portal hypertension if recanalization is not obtained. However, response to anticoagulation and long-term prognosis of APVT are not well-defined. Methods: Thirty-eight patients diagnosed with APVT between 1995 and 2003 from 5 Spanish referral hospitals, in whom cirrhosis and malignancy were specifically excluded, were included in this retrospective study. The response to anticoagulation therapy and development of portal hypertension related complications during follow-up were evaluated. Results: Mean follow-up was 43 months (range, months). Recanalization occurred in 12 of 27 patients receiving anticoagulation versus 0 of 11 patients who did not receive anticoagulation (P.008). Rates of recanalization were influenced by the precocity of heparin administration and the number of underlying prothrombotic conditions. Follow-up upper endoscopy performed in 29 patients disclosed gastroesophageal varices in 16 (55%). Varices appeared as early as 1 month after APVT. However, in most patients varices were detected in successive endoscopies, mainly during the first year. Two-year actuarial probability of variceal bleeding was 12% and for ascites 16%. Five-year survival was 87%. Mortality was related to the APVT episode in 2 cases and to an underlying hematologic disorder in one. Conclusions: Anticoagulation achieved recanalization in about 40% of patients. Most patients not achieving recanalization will develop gastroesophageal varices during follow-up. However, development of variceal bleeding and ascites is infrequent, and survival is satisfactory. Acute thrombosis of the portal venous system (APVT) is a rare disorder in adults. In the acute phase, APVT can lead to mesenteric ischemia and intestinal necrosis. If recanalization of the portal venous system is not achieved in those surviving patients, then the development of APVT is followed by the formation of a portal cavernoma and chronic portal hypertension along with its related complications. 1 5 Spontaneous recanalization after APVT is quite uncommon, 6,7 and recanalization after early anticoagulation has only been reported in case reports and in a small series of patients. 1,8,9 In the largest series, reported by Condat et al, 1 recanalization was observed in 90% of patients receiving early anticoagulation therapy. However, up until now these results have not been confirmed. In addition, the incidence and the rate of development of esophageal varices, portal hypertension related bleeding, and ascites after APVT are not well-known. The aims of the present study were to establish the response to anticoagulation therapy, the development of portal hypertension related complications, and survival during long-term follow-up in a cohort of patients with acute noncirrhotic nontumoral portal vein thrombosis. Methods Study Cohort This retrospective study included all patients who had been diagnosed with AVPT in the gastroenterology or surgical units of 5 reference hospitals in Spain between January 1995 and September The diagnosis of APVT was made on the basis of recent abdominal pain and without signs of chronic portal hypertension. The exclusion critieria included those patients with portal cavernoma or the presence of collateral portosystemic circulation by using Doppler ultrasound, computed tomography (CT) scan, magnetic resonance imaging, and/or angiography. Those with liver cirrhosis, Budd-Chiari syndrome, or malignancy were excluded by means of clinical history and follow-up, biochemical data (including markers of viral infection), liver imaging, or liver biopsy where necessary. All patients were studied for known prothrombotic and myeloproliferative disorders including blood marrow biopsy. 10 JAK2/V617F mutation was assessed in half of the patients. All subjects underwent evaluation to determine the initial extent of thrombosis, and the response to anticoagulant therapy on follow-up was performed by Doppler ultrasound, contrast-enhanced CT scan, magnetic resonance angiography, and/or splanchnic angiography. The cause of thrombosis, whether in the portal, mesenteric, or splenic vein, was considered to be a local condition when one of the following was present: (1) septic pylephlebitis from intra-abdominal sources, (2) acute or chronic pancreatitis, or (3) recent abdominal surgery. 10 Despite the presence of one of these, the presence of a prothrombotic or myeloproliferative disorder was evaluated. Although a definitive protocol for performing follow-up endoscopies was not fol- Abbreviations used in this paper: APVT, acute portal vein thrombosis; CT, computed tomography by the AGA Institute /08/$34.00 doi: /j.cgh

2 December 2008 ACUTE PORTAL THROMBOSIS OUTCOME 1413 lowed, at least one upper gastrointestinal endoscopy was performed in 29 of 38 patients at some point during follow-up to assess the presence of gastroesophageal varices. Follow-up data were collected up to July 2004 or death. The end points evaluated during follow-up included the response to anticoagulation therapy, development of esophageal or gastric varices, incidence of variceal bleeding, ascites, new venous thrombosis episodes, and death. According to Condat et al, 1 recanalization was considered complete when the portal vein trunk, one or both of the 2 main branches, and the splenic vein were patent when reevaluated after at least 6 months of anticoagulant therapy. Recanalization was considered partial if it was achieved in the portal vein trunk and in at least one of its main branches, in the splenic vein (if previously thrombosed), or in the mesenteric vein (if previously thrombosed). Esophageal varices were classified as large ( 5 mm) or small, 11 whereas gastric varices were classified according to classification by Sarin et al. 12 Variceal bleeding was defined according to the Baveno criteria. 11 Ascites was considered a chronic complication of APVT if it developed after 1 month of the thrombosis episode. Statistical Analysis The SPSS 11.0 package (SPSS Inc, Chicago, IL) was used to perform the analysis. Data are reported as frequencies, means with standard error of the mean, or range. Categorical variables were assessed by means of the Fisher exact test. Cumulative risk of developing variceal bleeding, ascites, or survival end points was evaluated by Kaplan-Meier curves. Significance was established at P.05. Results Thirty-eight patients were included in the study (Table 1). Mean age at diagnosis was 49 years (range, years). Mean follow-up in 36 patients surviving the acute episode was 43 months (range, months). Diagnosis of thrombosis of portal, mesenteric, or splenic veins was based on data obtained from Doppler ultrasound in 31 subjects (81.6%), CT scan in 32 subjects (84.2%), magnetic resonance angiography in 9 subjects (23.7%), and angiography in 7 subjects (18.4%). Two or more imaging techniques were used in 31 (81.6%) subjects. The portal vein was the only thrombosed vessel in 10 patients (26%); 16 Table 1. Main Characteristics of the 38 Patients Included in the Study Age a (y) 49 (17 74) Gender (female) b 20 (53) Thrombosis extension: Portal vein 10 (26) Portal vein splenic or mesenteric vein 16 (43) All 3 veins 10 (26) Extension undefined 2 (5) Complete etiologic study b 34 (89) Anticoagulation therapy b 27 (71) Upper gastrointestinal endoscopy b 29 (76) Follow-up c (mo) 43.6 (6 112) a Mean (range). b n (%). c Follow-up of those patients surviving the acute episode. Table 2. Etiology of PVT in the 34 Patients With a Complete Etiologic Study Chronic myeloproliferative disorders 12 a Polycythemia vera 5 b Essential thrombocythemia 4 b Idiopathic myelofibrosis 1 b Undefined myeloproliferative disorders 2 b Paroxysmal nocturnal hemoglobinuria 1 Primary antiphospholipid syndrome 1 b Prothrombin gene G20210A mutation 3 b Protein C deficiency 3 Protein S deficiency 1 Factor V Leiden mutation 1 Local factor alone 7 Unknown (idiopathic) 8 a In 3 patients 2 procoagulant conditions were found, polycythemia vera protein S deficiency, undefined myeloproliferative disorder protein S deficiency and undefined myeloproliferative disorder prothrombin gene G20210A mutation. b In these 6 patients a predisposing local factor was also found. patients had portal vein thrombosis plus thrombosis of either the mesenteric or the splenic vein; 10 additional patients had involvement of the 3 vessels; and the extension of the thrombus was not accurately assessed in 2 patients. Etiology Despite the retrospective nature of the study, a complete etiologic work-up was available in almost all patients (34 of 38 patients [89.5%]). A causal factor was identified in 26 subjects (Table 2). In 4 patients in whom the etiologic work-up was not complete, at least one potential cause of thrombosis was identified; 2 patients had local factors (pylephlebitis caused by acute cholecystitis and severe acute pancreatitis, respectively), and 1 had an essential thrombocythemia. Anticoagulation Treatment and Recanalization Twenty-seven patients (anticoagulation group) were placed on anticoagulation with intravenous or low molecular weight heparin within 30 days of the onset of symptoms (mean, 6 days; range, 0 30 days). Parenteral heparin was maintained during a median time of 15 days (range, 3 91 days) and was then shifted to acenocoumarol. In 4 of the remaining 11 patients (no anticoagulation group), anticoagulation was delayed until the etiologic diagnosis was completed, and a prothrombotic disorder was identified (median time from onset of symptoms, 68 days; range, days). For reasons unknown, anticoagulation was never administered in the other 7 patients, even though no formal contraindication was present. Anticoagulation was stopped after 6 months (median, 14 months; range, months) in those patients in whom no prothrombotic disorder or only a local factor was found. However, in those patients in whom a prothrombotic disorder was detected, oral anticoagulation was maintained for life. International normalized ratio was targeted to 2 and 3, except in the patient with primary antiphospholipid syndrome, when it was targeted between 2.5 and 3.5. No patient received thrombolytic therapy. Recanalization was assessed by various means: Doppler ultrasound in 33 patients (86.8%), CT scan in 9 (23.7%), and magnetic resonance angiography in 13 (34.2%). Recanalization

3 1414 TURNES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 12 Figure 1. (A) Comparison of the effects of anticoagulation therapy (grey bar) on recanalization rate (complete or partial). As shown, half of the patients achieved recanalization with anticoagulation, but spontaneous recanalization was not observed. (B) Comparison of the effects of anticoagulation (grey bar) on recanalization rate (complete or partial) according to the number of etiologic factors identified. As shown, recanalization was observed in 59% of patients with a single factor, but in none with more than one. was demonstrated in 12 of the 27 patients receiving anticoagulation therapy but in none of the patients who did not receive anticoagulation (P.008, Figure 1A). Precocity of heparin administration influenced the rate of recanalization. Venous thrombosis recanalization happened in 10 of the 16 patients in whom heparin was initiated within the first week and in only 2 of the 11 subjects in whom anticoagulation was started later (P.05). Of the 12 patients in whom recanalization was demonstrated, this was complete in 6 patients and partial in the remaining 6. Of the 6 patients achieving partial recanalization, 4 recanalized the portal vein trunk, maintaining the splenic vein thrombosed, and 1 recanalized the splenic vein and 1 the mesenteric vein but not the portal vein trunk. Consequently, 24 of the 36 patients surviving more than 6 months developed chronic portal vein thrombosis. The number of prothrombotic conditions present also influenced the efficacy of anticoagulation. Of those receiving anticoagulation, 24 patients had a complete etiologic work-up. Recanalization occurred in 10 of the 17 with 1 prothrombotic condition identified but in none of those patients who had 2 or more prothrombotic conditions (P.019, Figure 1B). The extent of the thrombosis did not influence the recanalization rate (2 of 6 patients with only the portal vein involved vs 10 of 21 with more extended thrombosis, P.6). One patient experienced a spontaneous retroperitoneal hematoma related to anticoagulation that spontaneously resolved when anticoagulation was stopped. Portal Hypertension Related Complications, Rethrombosis, and Survival At least 1 follow-up upper gastrointestinal endoscopy was performed in 29 of the 38 patients. In 16 patients (55%) varices were diagnosed after a median follow-up of 7 months (range, 1 40 months). Esophageal varices were large in 11 patients and small in 4; one patient had isolated gastric varices. Endoscopy was not performed in 9 patients (2 because of early death and 7 because of refusal). The initial endoscopy was performed at a median of 4 months (range, 1 46 months) after APVT. At the initial endoscopy, 12 patients had already developed esophagogastric varices (2 of them as early as 1 month after APVT). In the remaining 17 patients who had not developed varices, a second endoscopy was performed in 13, the results of which showed 4 patients who had developed varices at 5, 9, 9, and 17 months after APVT. Of these 4 patients the endoscopy was performed in 2 because of a variceal bleeding episode (at 9 and 17 months, respectively), and in the remaining 2 it was scheduled. No varices were detected in the remaining 9 patients after a median follow-up of 8 months (range, months). Four patients experienced a variceal bleeding episode. One of the 27 patients of the anticoagulation group bled at 2 years after anticoagulation withdrawal, compared with 3 of the 11 patients in the noncoagulation group (P.06). The 2-year probability of being free of a first variceal bleeding episode was 88% (Figure 2). None of the 12 patients with complete or partial recanalization had variceal bleeding during follow-up. Variceal bleeding was controlled with vasoactive drugs plus endoscopic therapy in all patients. Therapy for the prevention of rebleeding consisted of nonselective beta-blockers plus endoscopy. There were no bleeding-related deaths or rebleeding during a mean follow-up of 24 months (range, 6 56 months). Acceptance to perform an upper endoscopy was obtained from 3 of 6 patients with complete recanalization and 3 of 6 with partial recanalization. Esophageal varices were present in only 1 patient with partial recanalization and in none of those Figure 2. Actuarial probability of being free of first variceal bleeding.

4 December 2008 ACUTE PORTAL THROMBOSIS OUTCOME 1415 with complete recanalization. Therefore, among those survivors without complete recanalization in whom an upper endoscopy was performed, 16 of 25 (64%) developed esophageal varices during follow-up. Development of ascites was observed in only 5 patients, 3 with complete thrombosis and 2 with partial recanalization, at 6, 9, 13, 25, and 102 months of follow-up, and in all cases it was easily controlled with low sodium diet (n 4) and diuretics (n 1). No patient with complete recanalization developed ascites. Two-year probability of being free of ascites was 84% (Figure 3). Of the 36 patients surviving the APVT episode (19.4%), seven experienced new thrombotic events during follow-up (4 in extrasplanchnic vessels and 3 in a vessel of the portal venous axis not initially involved). Six of 20 patients with an inherited or acquired prothrombotic disorder (30%) had a rethrombotic event, whereas this only happened in 1 of the 16 patients with only local factors or idiopathic etiology (6%). Fourteen of the 20 patients with prothrombotic disorders were on long-term anticoagulation; 3 patients had a rethrombotic episode. Of the remaining 6 who were not receiving anticoagulation, despite the presence of a prothrombotic condition, 3 experienced a rethrombotic episode. Three patients died during follow-up: 1 within the first month as a result of complications of acute thrombosis, 1 at 1.3 months as a result of infectious complications related to acute pancreatitis, and 1 patient at 49 months of follow-up in the setting of terminal idiopathic myelofibrosis. Five-year actuarial survival was 87% (Figure 4). Discussion During the past 2 decades, several studies have focused on the etiology, clinical manifestations, and management of complications related to portal hypertension related complications in patients with portal vein cavernoma However, the natural history of patients with symptomatic APVT and the impact of early anticoagulation on outcome have hardly been Figure 3. Actuarial probability of being free of ascites. Figure 4. Actuarial survival probability. studied. Up until now there has only been one small study assessing the etiology and response to anticoagulant therapy in APVT. 1 The present study performed in a large cohort of 38 patients with APVT provides additional information about the natural history of this disease. Despite the heterogeneous nature of patients included in the study and the lack of a standardized monitoring, the main drawbacks of the study, there are several important findings that should be emphasized. As in other previously published series of chronic PVT, prothrombotic conditions are the main cause of thrombosis. 1,16,20 22 In the present study more than 55% of the patients had an underlying prothrombotic condition. In 35% of subjects, the existence of myeloproliferative disorders was the leading cause of thrombosis. Despite the fact that most patients underwent a thorough screening for procoagulant disorders and myeloproliferative diseases, in 23.5% of subjects we did not find any predisposing factor, a figure similar to that previously reported. 1,13 It is important to note that in 6 of the 13 patients in whom a local factor was found, an additional procoagulant disorder was present. This is in keeping with the current opinion that venous thrombosis is a multifactorial disease in which multiple risk factors are a prerequisite for thrombosis to develop. 23 In our series, no recanalization was observed in patients with 2 or more etiologic factors predisposing thrombosis. If this finding is confirmed by other groups, an exhaustive etiologic study would help to have prognostic information for the recanalization event. The cumulative results of these findings strongly support the careful screening for prothrombotic disorders, even if a local factor is found as a possible cause of thrombosis. The rate of spontaneous recanalization after recent thrombosis, although not well-established, seems to be nil or extremely low. 6,7 However, the results of the present study confirmed those of Condat et al, 1 clearly showing that anticoagulation facilitates recanalization in a significant proportion of patients. In the present cohort of patients with APVT, recanalization (complete or partial) was observed in more than 60% of patients in whom anticoagulation was initiated within the

5 1416 TURNES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 12 first week but in less than 20% of those patients who started thereafter. In addition, it is important to mention that similar to a previous study, 1 no major complications associated with anticoagulation were observed. Our study was unable to confirm the previous observed relationship between extension of thrombosis and recanalization rate, 1 although in more than 80% of patients, more than 1 imaging technique was used to evaluate the extent of thrombosis. It is clear that the retrospective nature of our and the previous study 1 might have influenced the results. Further studies in this area are clearly needed, but until then, early anticoagulation should be recommended independently of thrombosis extension. In our experience, with this approach a complete recanalization might be expected in up to 25% of patients and partial recanalization in an additional 25% of patients. The fact that no patient with complete recanalization developed esophageal varices or ascites during a mean follow-up of 33 months (range, months) clearly shows that achieving recanalization has a major impact on the outcome of these patients. Also achieving partial recanalization seems not to be irrelevant. Indeed, although some patients with partial recanalization developed esophageal varices, no patient bled, and in the 2 patients who developed ascites, this was easily controlled with mild sodium restriction. Our study showed that esophageal varices might develop as soon as 1 month after the initial symptoms of thrombosis. The lack of gastroesophageal varices at an early endoscopy does not exclude that these might develop during the following months. Indeed, in 1 patient without varices at initial endoscopy, varices were subsequently diagnosed at month 9 during a bleeding episode. Although we cannot completely rule out the presence of varices before the diagnosis of acute PVT, this is highly unlikely because the presence of liver cirrhosis or clinical or imaging signs of portal hypertension at diagnosis precluded the inclusion of the patient in the study. Mild ascites is a common transient finding at the acute episode as a result of the inflammation process. However, the probability of developing ascites as a consequence of the development of the portal cavernoma is low and usually easily controlled. The precise mechanism for the development of ascites in the setting of chronic PVT is not clear. Other relevant data of the study are the impact of the presence of inherited or acquired prothrombotic disorders and anticoagulation in rethrombotic events during follow-up. Indeed, our data showed that only 6% of those patients without a recognized prothrombotic disorder had a rethrombotic event during follow-up, thus supporting the recommendation of not using long-term anticoagulation in these patients but only during the initial 6 months after the APVT. In contrast, rethrombosis occurred in 30% of patients with prothrombotic disorders. However, this happened in 50% of those not receiving anticoagulation and in only 21% of those under long-term anticoagulation. Again this finding supports the recommendation of using long-term anticoagulation in patients with a recognized prothrombotic disorder, although it only reduces, not abolishes, the risk of rethrombosis. The long-term survival of this cohort of patients with APVT was extremely good and was not related to the development of portal hypertension but to the underlying disease or to complications related to the acute thrombotic episode. However, we can rule out that survival of patients not achieving recanalization might be impaired because of the development of portal hypertension related complications after more prolonged follow-up. However, studies performed in cohorts of patients diagnosed at the phase of portal cavernoma have also shown that mortality is low (5% 10% 5-year mortality) and is mainly related to associated diseases rather than to complications of portal hypertension. 1,13,16,20 22 References 1. Condat B, Pessione F, Helene Denninger M, et al. Recent portal or mesenteric venous thrombosis: increased recognition and frequent recanalization on anticoagulant therapy. Hepatology 2000;32: Valla DC, Condat B. Portal vein thrombosis in adults: pathophysiology, pathogenesis and management. J Hepatol 2000;32: Sarin SK, Agarwal SR. Extrahepatic portal vein obstruction. Semin Liver Dis 2002;22: de Franchis R. Evolving consensus in portal hypertension: report of the Baveno IV consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol 2005;43: Webster GJ, Burroughs AK, Riordan SM. Review article: portal vein thrombosis new insights into aetiology and management. Aliment Pharmacol Ther 2005;21: Spahr L, Willems B, Gianfelice D, et al. Spontaneously reversible extensive portal vein thrombosis after gallbladder puncture during transjugular liver biopsy. J Hepatol 1996;24: Davies M, Satyadas T, Akle CA. Spontaneous resolution of a superior mesenteric vein thrombosis after laparoscopic nissen fundoplication. Ann R Coll Surg Engl 2002;84: Baril N, Wren S, Radin R, et al. The role of anticoagulation in pylephlebitis. Am J Surg 1996;172: Sheen CL, Lamparelli H, Milne A, et al. Clinical features, diagnosis and outcome of acute portal vein thrombosis. Q J M 2000; 93: Winslow ER, Brunt LM, Drebin JA, et al. Portal vein thrombosis after splenectomy. Am J Surg 2002;184: de Franchis R. Updating consensus in portal hypertension: report of the Baveno III Consensus Workshop on definitions, methodology and therapeutic strategies in portal hypertension. J Hepatol 2000;33: Sarin SK, Lahoti D, Saxena SP, et al. Prevalence, classification and natural history of gastric varices: a long-term follow-up study in 568 portal hypertension patients. Hepatology 1992;16: Denninger MH, Chait Y, Casadevall N, et al. Cause of portal or hepatic venous thrombosis in adults: the role of multiple concurrent factors. Hepatology 2000;31: Janssen HL, Meinardi JR, Vleggaar FP, et al. Factor V Leiden mutation, prothrombin gene mutation, and deficiencies in coagulation inhibitors associated with Budd-Chiari syndrome and portal vein thrombosis: results of a case-control study. Blood 2000; 96: Condat B, Pessione F, Hillaire S, et al. Current outcome of portal vein thrombosis in adults: risk and benefit of anticoagulant therapy. Gastroenterology 2001;120: Janssen HL, Wijnhoud A, Haagsma EB, et al. Extrahepatic portal vein thrombosis: aetiology and determinants of survival. Gut 2001;49: Condat B, Vilgrain V, Asselah T, et al. Portal cavernoma-associated cholangiopathy: a clinical and MR cholangiography coupled with MR portography imaging study. Hepatology 2003;37:

6 December 2008 ACUTE PORTAL THROMBOSIS OUTCOME Vleggaar FP, van Buuren HR, Schalm SW. Endoscopic sclerotherapy for bleeding oesophagogastric varices secondary to extrahepatic portal vein obstruction in an adult Caucasian population. Eur J Gastroenterol Hepatol 1998;10: Zargar SA, Javid G, Khan BA, et al. Endoscopic ligation compared with sclerotherapy for bleeding esophageal varices in children with extrahepatic portal venous obstruction. Hepatology 2002; 36: Bombeli T, Basic A, Fehr J. Prevalence of hereditary thrombophilia in patients with thrombosis in different venous systems. Am J Hematol 2002;70: Primignani M, Martinelli I, Bucciarelli P, et al. Risk factors for thrombophilia in extrahepatic portal vein obstruction. Hepatology 2005;41: Amitrano L, Guardascione MA, Scaglione M, et al. Prognostic factors in noncirrhotic patients with splanchnic vein thromboses. Am J Gastroenterol 2007;102: Rosendaal FR. Venous thrombosis: a multicausal disease. Lancet 1999;353: Address requests for reprints to: Juan Carlos García-Pagán, MD, Hepatic Hemodynamic Laboratory, Liver Unit, Hospital Clínic, Villarroel 170, Barcelona 08036, Spain. jcgarcia@clinic.ub.es; fax: The authors disclose the following: The authors are indebted to Ms M. A. Baringo, L. Rocabert, and R. Saez for their expert technical assistance. The authors also express their gratitude to Ms Clara Esteva for her secretarial support. Supported in part by grants from Ministerio de Educación y Ciencia (SAF-07/61298, BFU /BFI), from Instituto de Salud Carlos III (FIS 06/0623), and from Fundación Española para el Estudio de las Enfermedades Hepáticas. Ciberehd is funded by Instituto de Salud Carlos III.