Transjugular Intrahepatic Portosystemic Shunt in the Treatment of Refractory Bleeding From Ruptured Gastric Varices

Transcription

1 Transjugular Intrahepatic Portosystemic Shunt in the Treatment of Refractory Bleeding From Ruptured Gastric Varices KARL BARANGE, 1 JEAN-MARIE PÉRON, 1 KAMRAN IMANI, 1 PHILIPPE OTAL, 2 JEAN-LOUIS PAYEN, 1 HERVE ROUSSEAU, 2 JEAN-PIERRE PASCAL, 1 FRANCIS JOFFRE, 2 AND JEAN-PIERRE VINEL 2 The optimal management of ruptured gastric varices in patients with cirrhosis has not been codified yet. The present study reports the use of transjugular intrahepatic portosystemic shunt (TIPS) in patients with refractory gastric variceal bleeding. Thirty-two consecutive patients were included. All had been unresponsive to vasoactive agents infusion, sclerotherapy, and/or tamponade and were considered poor surgical candidates. They were followed-up until death, transplantation, or at least 1 year (median: 509 days; range 4 to 2,230). Hemostasis was achieved in 18 out of 20 patients actively bleeding at the time of the procedure. In the whole sample of 32 patients, rebleeding rates were 14%, 26%, and 31%, respectively at 1 month, 6 months, and 1 year. De novo encephalopathy was observed in 5 (16%) patients. Seven patients experienced complications and consequently 4 of these patients died. TIPS primary patency rates were 84%, 74%, and 51%, respectively, at 1 month, 6 months, and 1 year. For the same periods of time, survival rates were 75%, 62%, and 59%. These results suggest that TIPS can be used in cirrhotic patients with refractory gastric variceal bleeding and are effective in achieving hemostasis as well as in preventing rebleeding. (HEPATOLOGY 1999;30: ) In the absence of any specific controlled trial the management of bleeding gastric varices in patients with cirrhosis is still a difficult problem. These bleedings are generally considered more severe than those originating from ruptured esophageal varices and more prone to relapse. 1 Transjugular intrahepatic portosystemic shunt (TIPS) has been extensively used within the last 10 years. It allows decompression of the portal system through a percutaneous approach and has consequently been used to treat both ascites 2 and variceal bleeding Its use as salvage therapy after failure of other usual therapies was reported in patients bleeding from ruptured esophageal varices Its use in patients bleeding from gastric varices has seldom been addressed. 11,13,14 Abbreviation: TIPS, transjugular intrahepatic portosystemic shunt. From 1 Service d Hépato-Gastro-Entérologie, Fédération Digestive, CHU Purpan, and 2 Service de Radiologie, CHU Rangueil, Toulouse, France. Received October 8, 1998; accepted August 10, Address reprint requests to: J. M. Péron, M.D., Service d Hépato-Gastro-Entérologie, Fédération Digestive, CHU Purpan, 31059, Toulouse, France. peron.jm@chutoulouse.fr; fax: Copyright 1999 by the American Association for the Study of Liver Diseases /99/ $3.00/ In the present study, we report the use of TIPS in 32-consecutive patients bleeding from gastric varices after failure of usual treatments including sclerotherapy and vasoactive drugs. PATIENTS AND METHODS Patient Selection. From January 1992 to July 1997, 48 patients were admitted to the intensive care unit of our department for bleeding from ruptured gastric varices. Four patients died within a few hours after admission from massive uncontrolled hemorrhage, 10 were successfully treated with vasoactive drugs and intravariceal injection of 1% polidocanol or cyanoacrylate, 2 were treated with surgery (portocaval shunting in 1, transsection of the esophagus in 1). In the remaining 32 patients, hemostasis could not be obtained and they were treated by TIPS. The main clinical and biochemical characteristics of these 32 patients are presented in Table 1. All patients had experienced at least 1 previous bleeding episode. The site of this episode was known in 20 patients. According to a classification by Sarin and Kumar 15 it was GOV-1 in 9 patients (28%), GOV-2 in 3 patients (9%), and esophageal varices in 8 patients (25%). The median interval between the prior bleeding episodes and the acute episode evaluated in the study was 13 days (1 to 1,096 days). Six prior bleeding episodes were not treated, 1 was treated with glypressine, 9 by sclerotherapy, 3 by sclerotherapy and octreotide or glypressine, 1 by surgery and octreotide, 1 by surgery, sclerotherapy, and a Linton tube. The treatment of the other prior bleeding episodes was unknown. The bleeding episode was ascribed to ruptured gastric varices according to 1 of the following criteria: actively bleeding varix at endoscopy, adherent clot on a varix, or presence of blood in the stomach with gastric varices as the only possible cause of the bleeding. The bleeding varix was found along the lesser curvature of the stomach in 22 patients and in the fundus in 10 patients. They corresponded, respectively, to GOV-1 and GOV-2 of the classification by Sarin and Kumar. 15 In all 32 patients, TIPS was indicated because of failure of vasoactive agents (glypressine in 10, octreotide in 9, somatostatine in 4, and 3 unknown) and/or sclerotherapy. In 11 patients, a Linton tube had to be used because of persistent active bleeding, while awaiting for TIPS insertion (Table 2). In 20 instances, TIPS was inserted in emergency conditions in patients whose bleeding was active as assessed by the need to tranfuse 2 U ofblood or more to increase and maintain hematocrit at 25% within 24 hours. Two of these patients had previously been treated by surgical mesentericoportal shunt, the obstruction of which was responsible for the hemorrhage. In the remaining 12 patients, TIPS had been indicated because of early rebleeding after hemostasis had been achieved. These patients were considered poor surgical candidates because of the severity of the underlying liver disease (6 Child s C patients), general contraindication (age older than 65, severe obesity, respiratory failure) Whitaker syndrome in 1 case, and human immuondeficiency virus infection in another case.

2 1140 BARANGE ET AL. HEPATOLOGY November 1999 TABLE 1. Main Clinical and Biochemical Characteristics of the 32 Cirrhotic Patients Treated With TIPS for Bleeding Gastric Varices Sex (M/F) 23/9 Age (mean SD) Cause of cirrhosis, no. (%) Alcoholic 19 (59) Postviral 4 (12) Alcohol virus 6 (19) Hemochromatosis 1 Primary biliary cirrhosis 1 Whitaker syndrome 1 Hepatocellular carcinoma 2 Child-Pugh class, no. (%) A 3 (9) B 14 (44) C 15 (47) Previous bleeding episodes Mean 2.3 Range 1-5 Interval between previous bleeding episodes and episode evaluated in the study (d) Median 13 Range (1-1,096) Transfusion needs (U of blood) Mean 8 Range 1-26 Follow-up (d) Median 509 Range 4-2,230 Procedure. The TIPS technique has been previously described. 16,17 Briefly, the procedure was performed under general anesthesia with tracheal intubation. The main right hepatic vein was catheterized by the transjugular route. A needle pushed through the liver parenchyma allowed insertion of a wire guide into a branch of the portal vein. The intraparenchymal tract was thereafter dilated to 8 mm using an angioplasty catheter. Wallstents (Pfeizer, Bülach, Switzerland), Memotherm (Angiomed, Karlsruhe, Germany), or Cragg (Mintech, La Ciotat, France) stents were used, respectively, in 26 patients, 5 patients, and 1 patient. Inferior vena cava and portal vein pressures were measured before and after stenting. Whenever the pressure gradient decrease was less than 50%, the shunt was dilated to 10 mm. Patients were not given anticoagulants. All were treated with antibiotics, which are routinely given in our unit to patients with cirrhosis admitted for variceal bleeding. 18 Shunt patency was assessed using Doppler ultrasonography 1 day, 7 days, 1 month, and every 3 months thereafter TIPS, or when shunt dysfunction was suspected. Whenever blood velocity was less than 50 cm/sec in the shunt, dysfunction was suspected and angiography performed. Definitions. The following definitions were used: (1) bleeding: hematemesis and/or need to transfuse at least 2 U of blood within 24-consecutive hours to maintain and/or increase hematocrit to 25%; (2) refractory bleeding: failure of vasoactive agents (glypressine, octreotide, or somatostatine) and/or sclerotherapy in achieving hemostasis; (3) hemostasis: absence of any symptoms of active bleeding for at least 24-consecutive hours; (4) rebleeding: reoccurence of a bleeding episode after hemostasis had been achieved. Statistical Analysis. Results were expressed as mean SD unless otherwise stated. Comparisons were performed using Student s t test for paired series and 2 test. Survival, rebleeding, and shunt patency probability rates were calculated using the Kaplan-Meier method (1 transplant recipient was censored at the date of transplantation). P.05 was considered the level of statistical significance. RESULTS TIPS insertion was successful in all 32 patients. In 4 patients, 2 coaxial prostheses were necessary to entirely cover the intraparenchymal tract. In 1 patient, complete deployment of the prosthesis within the liver could not be obtained and a second parallel shunt had to be performed. Pressure gradient between the inferior vena cava and the portal vein dropped from mm Hg to mm Hg, respectively, before and after TIPS (P.001). Rebleeding. Hemostasis was obtained in 18 of the 20 patients (90%) who were actively bleeding at the time of the procedure. In 1 patient, 6Uofblood were necessary within 24 hours after TIPS insertion to maintain hematocrit at 25%. Injection of 1% polidocanol in a gastric varix presenting an adherent clot, finally obtained hemostasis. The other patient was transfused 3Uofblood during the 48 hours after TIPS and thereafter no symptom of rebleeding was observed and hematocrit remained stable. Actuarial probability of remaining free of rebleeding is presented in Fig. 1. Within a 2-year period, 9 patients rebled; 8 of whom rebled within the first year after TIPS. Four patients died from bleeding before endoscopy and Doppler ultrasonography could be performed, so that the source of rebleeding and TIPS status remained unkown. The other 5 patients were all found to have shunt stenosis or obstruction. One patient had total shunt and portal vein thrombosis and died from bleeding 5 days later. One patient refused a new procedure and was treated with sclerotherapy; he rebled 4 months later and died. Two patients were treated with TABLE 2. Treatments Used in 32 Cirrhotic Patients With Bleeding Gastric Varices Before TIPS Insertion Vasoactive Drug Linton Tube Endoscopic Sclerotherapy No. of Patients No. of patients FIG. 1. Actuarial probability of remaining free of rebleeding in 32 cirrhotic patients treated with TIPS for bleeding gastric varices. (Figures in parentheses on the x-axis represent number of patients at risk.)

3 HEPATOLOGY Vol. 30, No. 5, 1999 BARANGE ET AL angioplasty, and a second TIPS was performed in 1 patient because of complete occlusion of the first shunt. One of these patients rebled 4 months later and azygoportal deconnection with gastric transsection was performed; he died 3 weeks after surgery from sepsis and liver failure. There was no difference in the rebleeding rate between GOV-1 and GOV-2 bleeding sites (26.1% vs. 37.5%; P.66). Clinical Complications. Comparison of the main clinical and biochemical parameters, before and 1 month after TIPS, showed no significant difference except for the prothrombin index that decreased (Table 3). One patient died from acute hepatic failure, which appeared 48 hours after treatment and was ascribed to TIPS. The ischemic origin of this complication was confirmed by transjugular liver biopsy. One patient had peritoneal bleeding and was treated with surgery, which found a tear in the extrahepatic portion of the right hepatic vein. The patient died from multiorgan failure 13 days later. TIPS was considered the cause of systemic infections in 2 patients. One patient had a Staphylococcus aureus infection, which lead to his death from respiratory failure. Three portal vein thromboses were diagnosed. One was limited to the left branch of the portal vein and remained asymptomatic. In the second, Doppler ultrasonography on day 7 found thrombosis of both branches of the portal vein and stenosis of the shunt; after angioplasty and insertion of a second coaxial prosthesis, the thrombus completely resolved. The third patient had complete obstruction of both the TIPS and the portal vein and was previously mentioned. Five patients (15.6%) had de novo hepatic encephalopathy after TIPS. One patient, previously mentioned, had acute hepatic failure; 2 patients had acute pulmonary infection and died 4 days after TIPS; 1 patient was efficiently treated with lactitol; in the fifth patient symptoms persisted until death, which was 1 month after TIPS. On the other hand, in 4 patients, encephalopathy which was present at admission, disappeared after TIPS. Accordingly, as a whole, TIPS did not significantly change the rate of encephalopathy. Shunt Obstruction or Stenosis. As a whole, 18 episodes of stenosis or obstruction were diagnosed in 10 patients. One patient died from bleeding before any treatment could be performed; 1 patient refused treatment and is, to date, alive and well; 4 patients were treated by angioplasty with insertion of a second coaxial prosthesis, and 3 patients had a second TIPS. The last patient, a Child s A woman, had a Warren shunt that was found occluded 3 months later. A second TIPS was then performed which, to date, has stayed patent. Survival. Figure 2 shows the actuarial probability of survival. A total of 19 patients died within a mean of 290 days (range 4 to 1,428 days). Causes of death are listed in Table 4. TIPS was held responsible for 4 of these deaths: 1 from acute liver FIG. 2. Actuarial probability of survival in 32 cirrhotic patients treated with TIPS for bleeding gastric varices. (Figures in parentheses on the x-axis represent number of patients at risk.) failure, 1 from peritoneal bleeding, 1 from sepsis, and 1 from massive bleeding secondary to total portal vein and shunt thrombosis. DISCUSSION The present study reports the results of TIPS in 32 patients with cirrhosis treated in emergency conditions for gastric variceal bleeding refractory to usual treatments. To our knowledge only 1 similar study, which included 28 patients, has been published so far. 14 Only bleeding episodes, the origin of which had been clearly shown to be ruptured gastric varices were considered. All of our patients had previously presented at least 1 episode of bleeding. The index bleed was particularly severe if one considers it required a mean of 8 U of blood. By comparison, transfusion needs were 4.8 in the series by Sarin et al. 19 Moreover, the bleeding episode could not be controlled by usual treatments, including sclerotherapy and vasoactive agents in any of our patients. The effectiveness of sclerotherapy in this setting varies tremendously in the literature, ranging from 26% to 80%. These discrepancies could be accounted for by the location of the varices. In GOV-1 varices, hemostasis is obtained in more than 80% of the cases, whereas in GOV-2 varices, the rate of hemostasis ranges from 26% 22 to 70%. 15 However, all authors agree on the very high rebleeding rate that can be as high as 89%. 23 Surgery was not used in emergency conditions in our patients because they were considered poor candidates for such therapy according to the severity of liver disease, age above 65 years, or associated diseases. Out of our 3 Child s A patients, one was 72-years-old at the time of the index bleed and thus TIPS was preferred over surgery, another had a TABLE 3. Comparison of Main Biochemical Data 1 Month Before and 1 Month After TIPS in 32 Cirrhotic Patients Treated for Bleeding Gastric Varices Before TIPS After TIPS P Albumin (g/l) NS Bilirubin (µmol/l) NS Prothrombin index (%) Child-Pugh score NS TABLE 4. Causes of Death in 32 Cirrhotic Patients Treated With TIPS for Bleeding Gastric Varices Cause of Death No. of Patients Peritoneal bleeding 1 Variceal bleeding 6 Sepsis 5 Liver failure 6 Unknown 1

4 1142 BARANGE ET AL. HEPATOLOGY November 1999 severe hemorrhage requiring a Linton tube and in this emergency situation was referred to TIPS. The third patient could have been referred to surgery, but TIPS was decided by the practitioner at the time. Child s A patients may do just as well with a surgical option, but a controlled study would be needed to observe the results. Moreover, emergency surgery, whether derivation or devascularization procedures, carries such high morbidity and mortality rates that it has been abandoned in most centers Bleeding esophageal or gastric varices can be injected with the tissue-adhesive cyanoacrylate, as described in a few uncontrolled studies. 29,30 This technique has not been popularized mainly because of the expertise it requires, particularly in avoiding damage to the endoscope and because of potentially severe complications. 31 More recently, the injection of thrombin was reported to achieve hemostasis in 100% of 11 patients. 32 However, these preliminary results should be confirmed and the risk of prion transmission precludes the use of this technique. Finally, embolization, whether through the percutaneous transhepatic 33,34 or the transjugular route, 35 can be performed. However, because portal hypertension is not relieved, rebleeding rates after embolization are reported to be greater than 60% Hemostasis was obtained by TIPS insertion in 18 out of 20 (90%) actively bleeding patients. In the 2 patients who continued to bleed after TIPS, hemostasis was finally obtained within 48 hours using endoscopic sclerotherapy in 1 case, and occurred spontaneously in the other case. These results are similar to the 96%-success rate reported by Chau et al. 14 Taking into account the critical status of our patients, despite the absence of any control group, this strongly suggests TIPS is an effective treatment of refractory gastric variceal bleeding in patients with cirrhosis. The rebleeding rate was 31% at 1 year. Thereafter, only 1 more patient bled during follow-up. This figure is very close to the 29% reported by Chau et al. 14 It is somewhat higher than the usual percentage reported after surgical shunting. 36 Rebleeding episodes were severe since 4 patients died before any treatment could be performed. Among the patients whose TIPS patency could be assessed by Doppler ultrasonography, all who were found to have stenosis or occlusion rebled. On the other hand, none of those found to have a patent shunt rebled. Therefore, one can conclude that TIPS is a very effective treatment to prevent rebleeding, at least as long as it remains patent. Were it not for strict follow-up, using serial sonography, helping diagnose early asymptomatic TIPS dysfunction, the rebleeding rates would have been higher. Indeed, as it has been well demonstrated, stenosis or occlusion is the main drawback of TIPS. In the present series, rates of TIPS dysfunction were 49% and 61%, respectively, 1 year and 2 years after insertion. These figures are similar to those reported in other series. 37 Our first month dysfunction rate of 14% could be ascribed to thrombosis. Some investigators advocate the use of anticoagulants to prevent early thrombosis of TIPS. However, in their experience obstruction rate was the same as our, 38 and the efficacy of such a preventive treatment has not been established. However, in 1 of our patients both the TIPS and the portal vein were found thrombosed, which led to severe rebleeding and death. Therefore, it is possible that a subgroup of patients could benefit from anticoagulants. How to select these patients is unknown and one would be quite reluctant to give anticoagulants to actively bleeding Child s C patients. Sixteen percent of our patients had de novo encephalopathy. The 95%-confidence interval of this percentage ranges from 5% to 33%, which is very similar to rates usually reported. 37 However, most Child s C cirrhotic patients have encephalopathy during or shortly after active bleeding. Therefore, a treatment that achieves hemostasis is very likely to improve this complication. This was confirmed in 4 patients in whom encephalopathy, which was present at admission, disappeared after TIPS. This is probably why no significant difference could be observed in our patients when comparing pre-tips and post-tips encephalopathy rates. Moreover, except in 1 patient whose encephalopathy symptoms persisted until death 1 month after TIPS, those episodes were controlled by usual medical therapy. No late encephalopathy was observed in this sample of patients. Although the Child-Pugh score did not change, prothrombin index, a well-known index of liver function, significantly decreased after TIPS. This could indicate that our patients experienced some impairment of their liver function. This is in keeping with a case of acute liver failure we observed. Such a complication of TIPS has already been reported 39 and ascribed to an arterial blood flow towards the liver too low to compensate for the drop in portal vein blood flow. Other complications ascribed to TIPS were 1 case of peritoneal bleeding, 2 systemic infections, and 3 thromboses of the portal vein or its branches. Three of these complications were considered lethal. Such a rate is similar to those usually reported. 37 Accordingly, neither the fact that the procedure was performed in emergency conditions, nor the poor general conditions of most of our patients should be considered contraindications for TIPS. The one-year survival rate was 36% for Child s C patients and 51% for the entire group of 32 patients. These figures compare favorably to those usually reported in similar conditions. 1 In conclusion, TIPS seems to be an effective treatment for cirrhotic patients with gastric variceal bleeding refractory to usual treatments, including sclerotherapy. REFERENCES 1. D amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. HEPATOLOGY 1995;22: Ochs A, Rossle M, Haag K, Hauenstein KH, Deibert P, Siegerstetter V, Huonker M, et al. The transjugular intrahepatic portosystemic stentshunt procedure for refractory ascites. N Engl J Med 1995;332: Sanyal AJ, Freedman AM, Purdum PP, Luketic VA, Shiffman ML, Tisnado J, Cole PE. Transjugular intrahepatic portosystemic shunt (TIPS) versus sclerotherapy for prevention of recurrent variceal hemorrhage: a randomized prospective trial [Abstract]. Gastroenterology 1994;106:A Groupe d étude des anastomoses intra-hépatiques. TIPS versus sclerotherapy plus propranolol in the prevention of variceal rebleeding: preliminary results of a multicenter randomized trial [Abstract]. HEPATOL- OGY 1995;22:297A. 5. Cabrera J, Maynar M, Granados R, Gorriz E, Reyes R, Pulido Duque JM, Rodriguez SanRoman JL, et al. Transjugular intrahepatic portosystemic shunt versus sclerotherapy in the elective treatment of variceal hemorrhage. Gastroenterology 1996;110: Sauer P, Theilmann L, Stremmel W, Benz C, Richter G, Stiehl A. Transjugular intrahepatic portosystemic stent shunt versus sclerotherapy plus propranolol for variceal bleeding. Gastroenterology 1997; 113: Rössle M, Deibert P, Haag K, Ochs A, Olschewski M, Siegerstetter V, Hauenstein K, et al. Randomised trial of transjugular intrahepatic

5 HEPATOLOGY Vol. 30, No. 5, 1999 BARANGE ET AL portosystemic shunt versus endoscopy plus propranolol for prevention of variceal rebleeding. Lancet 1997;349: Jalan R, Forrest EH, Stanley AJ, Redhead DN, Forbes J, Dillon JF, MacGilchrist AJ, et al. A randomized trial comparing transjugular intrahepatic portosystemic stent-shunt with variceal band ligation in the prevention of rebleeding from esophageal varices. HEPATOLOGY 1997;26: Cello JP, Ring EJ, Olcott EW, Koch J, Gordon R, Sandhu J, Morgan DR, et al. Endoscopic sclerotherapy compared with percutaneous transjugular intrahepatic portosystemic shunt after initial sclerotherapy in patients with acute variceal hemorrhage. A randomized, controlled trial. Ann Intern Med 1997;126: Merli M, Salerno F, Riggio O, De Franchis R, Fiaccadori F, Meddi P, Primignani M, et al. Transjugular intrahepatic portosystemic shunt versus endoscopic sclerotherapy for the prevention of variceal bleeding in cirrhosis: a randomized multicenter trial. HEPATOLOGY 1998;27: Bilodeau M, Rioux L, Willems B, Pomier-Layrargues G. Transjugular intrahepatic portacaval stent shunt as a rescue treatment for lifethreatening variceal bleeding in a cirrhotic patient with severe liver failure. Am J Gastroenterol 1992;87: McCormick P, Dick R, Panagou E, Chin J, Greenslade L, McIntyre N, Burroughs A. Emergency transjugular intrahepatic portasystemic stent shunting as salvage treatment for uncontrolled variceal bleeding. Br J Surg 1994;81: Sanyal AJ, Freedman AM, Luketic VA, Purdum PP, Shiffman ML, Tisnado J, Cole PE. Transjugular intrahepatic portosystemic shunts for patients with active variceal hemorrhage unresponsive to sclerotherapy. Gastroenterology 1996;111: Chau T, Patch D, Chan Y, Nagral A, Dick R, Burroughs A. Salvage transjugular intrahepatic portosystemic shunts: gastric fundal compared with oesophageal variceal bleeding. Gastroenterology 1998;114: Sarin SK, Kumar A. Gastric varices: profile, classification, and management. Am J Gastroenterol 1989;84: Richter G, Noeldge G, Palmaz J, Rössle M. The transjugular intrahepatic portosystemic stent-shunt (TIPSS): experience results of a pilot study. Cardiovasc Intervent Radiol 1990;13: Rössle M, Haag K, Ochs A, Sellinger M, Nöldge G, Perarnau J, Berger E, et al. Transjugular intrahepatic portosystemic stent-shunt procedure for variceal bleeding. N Engl J Med 1994;330: Soriano G, Guaner C, Thomas A, Villanueva C. Norfloxacin prevents bacterial infection in cirrhotics with gastrointestinal hemorrhage. Gastroenterology 1992;103: Sarin SK, Lahoti D, Saxena S, Marthy N, Makwana U. Prevalence, classification and natural history of gastric varices: a long term follow-up study in 568 portal hypertension patients. HEPATOLOGY 1992;16: Korula J, Chin K, Ko Y, Yamada S. Demonstration of two distinct subsets of gastric varices. Observation during a seven-year study of endoscopic sclerotherapy. Dig Dis Sci 1991;36: Hosking S, Johnson A. Gastric varices: a proposed classification leading to management. Br J Surg 1988;75: Gimson A, Westaby D, Williams R. Endoscopic sclerotherapy in the management of gastric variceal hemorrhage. J Hepatol 1991;13: Trudeau W, Pindiville T. Endoscopic injection sclerosis in bleeding gastric varices. Gastrointest Endosc 1986;32: Cello J, Grendell J, Crass R, Weber T, Trunkey D. Endoscopic sclerotherapy versus portacaval shunt in patients with severe cirrhosis and acute variceal hemorrhage. Long term follow-up. N Engl J Med 1987;316: Houssin D, Brunet A, Brochard L, D Hubert E, Benmansour A, Roche A, Bismuth H. Traitement des hémorragies incontrôlables dues aux varices oesophagiennes chez le cirrhotique. Embolisation trans-hépatique ou dérivation portale en urgence. Press Med 1986;15: Levine B, Gaskill H, Sirinek K. Portasystemic shunting remains the procedure of choice for control of variceal hemorrhage. Arch Surg 1985;120: Orloff MJ, Bell RH, Orloff MS, Hardison WGM, Greenburg AG. Prospective randomized trial of emergency portocaval shunt and emergency medical therapy in unselected cirrhotic patients with bleeding varices. HEPATOLOGY 1994;20: Villeneuve J, Pomier-Layrargues G, Duguay L, Lapointe R, Tanguay S, Marleau D, Willems B, et al. Emergency portacaval shunt for variceal hemorrhage. Ann Surg 1987;206: Soehendra N, Nam V, Grimm H, Kempeneers I. Endoscopic obturation of large esophagogastric varices wih bucrylate. Endoscopy 1986;18: Ramon M, Valla D, Gottlib J, Rueff B, Benhamou J. Obturation endoscopique des varices eso-gastriques par le bucrylate. Etude clinique de 49 malades. Gastroenterol Clin Biol 1986;10: Sée A, Florent C, Lamy P, Levy V, Bouvry M. Accidents vasculaires cérébraux après obturation endoscopique des varices oesophagiennes par l isobutyl-2-cyanoacrylate chez 2 malades. Gastroenterol Clin Biol 1986;10: Williams S, Peters R, Westaby D. Thrombin: an effective treatment for gastric variceal hemorrhage. Gut 1994;35: Lunderquist A, Borjesson B, Owman T, Bengmark S. Isobutyl 2-cyanoacrylate (bucrylate) in obliteration of gastric coronary vein and esophageal varices. AJR Am J Roentgenol 1978;130: L Herminé C, Chastanet P, Delemazure O, Bonnière P, Durieu J, Paris J. Percutaneous transhepatic embolization of gastroesophageal varices: result in 400 patients. AJR Am J Roentgenol 1988;152: Vinel JP, Scotto J, Levade M, Teisseire M, Cassigneul J, Calès P, Putois J, et al. Embolisation des varices oesophagiennes par voie transjugulaire en urgence dans le traitement des hémorragies digestives grâves du cirrhotique. Etude prospective de 84 patients. Gastroenterol Clin Biol 1985;9: Grace N, Conn H, Resnick R, Groszmann R, Atterbury C, Wright S, Gusberg R, et al. Distal splenorenal versus portal-systemic shunts after hemorrhage from varices: a randomized controlled trial. HEPATOLOGY 1988;8: Barange K, Rousseau H, Vinel JP. Anastomoses porto-systémiques intra-hépatiques et traitement de l hypertension portale. Gastroenterol Clin Biol 1996;20: Perarnau J, Raabe J, Schwing D, Rucin B, Monchouet S, Rössle M, Arbogast J. Anastomose porto-cave intra-hépatique par voie transjugulaire: résultats préliminaires. Gastroenterol Clin Biol 1993;17: Freedman A, Sanyal A, Tisnado J, Cole P, Shiffman M, Luketic V, Purdum P, et al. Complications of transjugular intrahepatic portosystemic shunt: a comprehensive review. Radiographics 1993;13: