A Prospective, Randomized Trial of Butyl Cyanoacrylate Injection Versus Band Ligation in the Management of Bleeding Gastric Varices

A Prospective, Randomized Trial of Butyl Cyanoacrylate Injection Versus Band Ligation in the Management of Bleeding Gastric Varices GIN-HO LO, KWOK-HUNG LAI, JIN-SHIUNG CHENG, MEI-HSIU CHEN, AND HUNG-TING CHIANG Gastric variceal bleeding is a catastrophic event. Both cyanoacrylate injection and banding ligation have been proven to be effective in the management of bleeding gastric varices. This study was performed to compare the efficacy and complications of both the modalities. Cirrhotic patients with a history of gastric variceal bleeding were randomized to 2 groups. The group receiving endoscopic obturation (group A) comprised 31 patients and the group receiving band ligation (group B) comprised 29 patients. Butyl cyanoacrylate and pneumatic-driven ligator were applied, respectively. Treatment was repeated regularly until obliteration of gastric varices. Active bleeding occurred in 15 patients in group A and 11 patients in group B. Initial hemostatic rate (defined as no bleeding for 72 hours after treatment) was 87% in group A and 45% in group B (P.03). The sessions required to achieve variceal obliteration and obliteration rates were similar in both the groups. However, rebleeding rates were significantly higher in group B (54%) than group A (31%) (P.0005). Treatment-induced ulcer bleeding occurred in 2 patients (7%) in group A and 8 patients (28%) in group B (P.03). The amount of blood transfusions required were also higher in group B than group A (4.2 1.3 vs. 2.6 0.9 units, respectively) (P <.01). Nine patients of group A and 14 patients of group B died (P.05). In conclusion, endoscopic obturation using cyanoacrylate proved more effective and safer than band ligation in the management of bleeding gastric varices. (HEPATOLOGY 2001;33:1060-1064.) Hemorrhage from esophagogastric varices is a serious complication of portal hypertension. 1 With the advent of endoscopic therapy, the efficacy in management of bleeding esophageal varices has been greatly enhanced. The rebleeding rate has been reduced and the survival rate improved. 2,3 In recent years, endoscopic variceal ligation (EVL) has been used to replace sclerotherapy as the treatment of choice for the prevention of rebleeding from esophageal varices. 4,5 Abbreviations: EVL, endoscopic variceal ligation; EIS, endoscopic injection sclerotherapy; GVL, gastric variceal ligation; GVO, gastric variceal obturation; GOV1, type 1 gastric varices; GOV2, type 2 gastric varices; IGV, isolated gastric varices. From the Division of Gastroenterology, Department of Medicine, Kaohsiung Veterans General Hospital, National Yang-Ming University, Taipei, Taiwan, Republic of China. Received September 29, 2000; accepted February 22, 2001. Supported by a grant from Kaohsiung Veterans General Hospital (No. VGHKS 87-07). Part of this work was presented in 2000, at the 8th United European Gastroenterology Week in Brussels, Belgium. Address reprint requests to: Gin-Ho Lo, M.D., Division of Gastroenterology, Department of Medicine, Veterans General Hospital, 386 Ta-Chung 1st. Road, Kaohsiung 813, Taiwan, Republic of China. E-mail: ghlo@isca.vghks.gov.tw; fax: (886) 7-3468237. Copyright 2001 by the American Association for the Study of Liver Diseases. 0270-9139/01/3305-0008$35.00/0 doi:10.1053/jhep.2001.24116 1060 On the other hand, the management of gastric varices is still a great challenge for endoscopists. Traditional methods including vasoconstrictors and balloon tamponades do not effectively reduce the rebleeding rate. 6 Endoscopic injection sclerotherapy (EIS) using ordinary sclerosants such as ethanolamine oleate or sodium tetradecyl sulfate also has a high risk of rebleeding. 6,7 Nonetheless, injection with cyanoacrylate glue has been reported to have a higher success rate. 7-11 Although the efficacy and safety of band ligation in the management of esophageal varices has been well documented, 4,5,12-14 the application of ligation in the treatment of gastric varices has rarely been reported. 15 This study was performed to compare the efficacy and safety of endoscopic obturation using butyl cyanoacrylate and band ligation in the management of gastric variceal bleeding. PATIENTS AND METHODS From July 1996 to December 1999, patients with the history of gastric variceal bleeding admitted to Kaohsiung Veterans General Hospital were considered for enrollment. All the patients received emergency endoscopy within 12 hours of admission. The inclusion criteria of this study were that patients were (1) cirrhotic, (2) had the history of gastric variceal bleeding, and (3) between 20 and 70 years of age. The diagnosis of cirrhosis was based on liver biopsy or results of clinical, laboratory, and imaging studies. The definition of gastric variceal bleeding included (1) active spurting or oozing of blood from gastric varices during endoscopy and (2) stigmata of recent hemorrhage such as blood clots coating on the gastric varices or the presence of hematocystic spots or white nipples on gastric varices. The exclusion criteria were (1) undetermined origin of bleeding from esophageal varices or gastric varices; (2) presence of deep jaundice (serum bilirubin 10 mg/dl), hepatic encephalopathy, or hepatorenal syndrome; (3) association with advanced hepatocellular carcinoma, Okuda-staging III, 16 uremia, cerebral vascular accident, or other debilitating disease; (4) prior history of sclerotherapy or shunt operation; and (5) life expectancy less than 24 hours. Betablockers were not administered during the study. Informed consent was obtained from all the patients. Our study was approved by the Ethics Committee of our hospital. Eligible patients were randomized into 2 groups, using opaquesealed envelopes numbered according to a table of random numbers. The group that underwent gastric variceal ligation was the GVL group, and the group that received obturation of the gastric varices was the GVO group. After the establishment of the diagnosis of gastric variceal bleeding, randomization was performed by an assistant, and endoscopic treatment was administered at once. Standard therapy, including blood and frozen plasma transfusion, fluid and electrolytes replacement, systemic antibiotics, and lactulose, were administered in patients of both the groups as clinically indicated. Our technique for GVL was as follows: premedication of 20 mg of buscopan was given intramuscularly. The pneumatic-active ligating device (Sumitomo, Tokyo, Japan) was attached to the endoscope (Olympus XQ 230, Tokyo, Japan) and an overtube was used. Ligation was performed initially on the active bleeding sites or the hematocystic spots on the gastric varices whenever possible. Subse-

HEPATOLOGY Vol. 33, No. 5, 2001 LO ET AL. 1061 quently, ligation was performed on prominent varices. One to 4 rubber bands were applied during 1 session. The interval between 2 sessions of GVL was 3 to 4 weeks. GVL was performed regularly until varices were obliterated or reduced to residual small varices, which could not be ligated. The method for GVO was as follows: premedication was similar to that given to the GVL group. Commercial needles (Olympus NM-1k) were rinsed with distilled water and lipiodol before and after injections. The obturation agent was n-butyl-2-cyanoacrylate (Histoacryl; B. Braun, Melsungen AG, Germany) 0.5 ml mixed with 1.5 ml Lipiodol ultra-fluide (Guerbet, Bois Cedex, France). If bleeding points were identified, the injection sites were focused at bleeding point. If no bleeding point could be identified, the injection site was focused at the hematocystic spots or the most prominent varices. During each session, 2 ml (rarely 4 ml) of obturation agent was used. GVO was regularly performed at 3- to 4-week intervals until obliteration of gastric varices. For patients with concomitant esophageal varices, EVL was performed immediately after the treatment of gastric varices for both groups during the same session. Our technique for EVL of esophageal varices has been described elsewhere. 11 Ligation was initiated at the gastroesophageal junction and advanced proximally. The treatment interval was similar to the management of gastric varices until obliteration of esophageal varices. To hasten the healing of ulcers induced by endoscopic therapy, mucosal protector sucralfate granules were routinely given to patients in both groups. After each session of endoscopic treatment, 1 g sucralfate granules were given 4 times per day and administered 1 hour before meals and at bedtime for 1 month. At the time of enrollment, the severity of liver disease was classified according to Child- Pugh classification. 17 The size of esophageal varices was classified according to criteria of Beppu et al. 18 The classification of gastric varices was based on criteria by Sarin et al. 19 : type 1 (GOV1), continuation of esophageal varices that extend for 2 to 5 cm below the gastroesophageal junction along the lesser curve of the stomach; type 2 (GOV2), varices extend beyond the gastroesophageal junction into the fundus of stomach; and IGV, isolated gastric varices. Clinical evaluations were performed monthly. Liver function tests and serum -fetoprotein measurements were performed every 3 months. After obliteration of gastroesophageal varices, endoscopy was performed every 6 months to detect recurrence of varices. Primary end point was initial hemostasis, which was defined as cessation of bleeding for more than 72 hours. A secondary end point was rebleeding, which occurred after 72 hours after endoscopic treatment among patients presenting with active bleeding or occurred after endoscopic treatment in patients presenting with stigmata of recent hemorrhage. The definition of rebleeding was recurrence of hematemesis, melena, and bleeding source proven to originate from gastric varices by emergency endoscopy. Only those who had a drop in hemoglobin and required blood transfusion of 2 or more units were considered to have rebleeding. Patients who presented with insignificant bleeding such as scanty tarry stool and clear aspirates from a nasogastric tube but who did not need blood transfusions were not considered to have rebleeding. Endoscopy was performed whenever possible to identify the cause of bleeding. Among patients receiving treatments for both esophageal varices and gastric varices, determination of origin from gastric varices was based on whether the ulcers on gastric varices are coated with blood clots or not. When recurrent varices or rebleeding from varices was encountered, repeated sessions of treatment, which were the same as the original treatments were performed. Blood was transfused to keep hemodynamics stable when a patient bled actively or when hemoglobin concentration was less than 9 g/dl. The requirements of blood transfusion were recorded during hospitalization for patients with acute bleeding and rebleeding during the study period. Complications were defined as any untoward events (e.g., ulcer bleeding from gastric varices, bacteremia, or bacterial peritonitis) that required active treatment or prolonged hospitalization. An ulcer of greater than 2 cm in diameter on gastric varices was defined as a huge ulcer. Treatment failure was defined as 2 or more rebleeding episodes from gastric varices or death from gastric variceal bleeding. The choice of 2 or more rebleeding episodes was based on that most favored in a consensus meeting. 20 The cross-over was generally not admissible except for emergency situations. Further management was at physician s discretion in patients rebled twice or more. The Student s t test, 2 test, and Fisher s exact test were used to assess the differences between both the groups with respect to clinical data and variceal recurrence when appropriate. Statistical analyses of both the groups were based on the intention-to-treat principle. Kaplan-Meier estimation was applied to examine the time to a first occurrence of recurrent bleeding from gastroesophageal varices and the time to death. The Log rank test was used to examine the variation of rebleeding episodes and survival. All P values were 2-tailed. A P value.05 was considered significant. Based on previous trials, 8,9,15 the expected acute hemostatic rate of GVO was around 95% and of band ligation was 88%, with a 2-tailed test to achieve a statistical power of 80% and an error of 5%; a sample size of 242 patients in each group was required. It is unlikely to enroll so many cases of gastric variceal bleeding by a single center. After an enrollment period of more than 3 years, an interim analysis showed that a significant difference was reached between the study groups, and we decided to terminate the trial. RESULTS During the study period, a total of 293 patients with esophagogastric variceal bleeding were encountered in our unit. Among them, 74 patients (25%) were noted to bleed from gastric varices. The time intervals between the onset of bleeding and hospital admission was about 3 hours (range, 1-8 hours). Fourteen patients were excluded: 1 patient was older than 72 years, 2 patients had undetermined esophageal or gastric variceal bleeding, 1 patient had deep jaundice, 1 patient had severe encephalopathy, 7 patients were associated with advanced hepatocellular carcinoma, 1 patient had uremia, and 1 patient had received a shunt operation. Thus, a total of 60 patients were enrolled in our trial. The GVO group comprised 31 patients and the GVL group comprised 29 patients. Twenty-seven patients (87%) in the EIS group and 28 patients (96%) in the GVL group had concomitant esophageal varices or had prior esophageal varices. Six patients in the GVO group and 4 patients in the GVL group had secondary gastric varices (i.e. developed after endoscopic treatment of esophageal varices). None of them had ever received treatment with -blockers. Among the GVO group, 21 patients were GOV1, 6 patients were GOV2, and 4 patients had isolated gastric varices. Among the GVL group, 20 patients were GOV1, 8 patients were GOV2, and 1 patient had isolated gastric varices. The characteristics of both groups are shown in Table 1. Both groups were comparable regarding the age, gender, etiologies of portal hypertension, severity of liver disease, and sizes of gastric varices. The prevalence of patients associated with hepatocellular carcinoma was higher in the GVL group than in the GVO group; however, the difference was not significant. The size of esophageal varices and gastric varices before performing endoscopic therapy were also similar between both groups. The mean follow-up period was 14 months in the GVO group and 9 months in the GVL group. One patient in each group was lost to follow-up. The results are shown in Table 2. Initial hemostasis was achieved in 13 of 15 patients (87%) in the GVO group and in 5 of the 11 patients in the GVL group (45%) (P.03). The higher success rate in the EIS group applied to both GOV1 and GOV2 subgroups. The treatment of 1 patient in the GVO group and 2 patients in the GVL group failed because of the

1062 LO ET AL. HEPATOLOGY May 2001 TABLE 1. Characteristics GVO GVL P No. of patients 31 29 Age 58 17 55 13 NS Men/Women 24/7 22/7 NS Etiology of portal hypertension Alcohol 10 6 NS HBV 11 11 NS HCV 7 10 NS HBV HCV 3 2 NS Child-Pugh class A/B/C 8/16/7 5/17/7 NS Ascites present 15 (48%) 17 (59%) NS Albumin (g/dl) 3.0 1.1 3.2 1.2 NS Bilirubin (mg/dl) 2.3 0.9 2.0 1.0 NS Association with EV 27 (87%) 28 (96%) NS Size of gastric varices F1/F2/F3 1/21/9 2/19/8 NS NS Association with HCC 7 10 NS Abbreviations: NS, not significant; HBV, hepatitis B virus; HCV, hepatitis C virus; EV, esophageal varice; HCC, hepatocellular carcinoma. difficulty in approaching during active bleeding. One patient in the GVL group was rescued by histoacryl injection under an emergency situation to control active bleeding. Two patients in the GVO group and 3 in the GVL group died of exsanguination despite the use of vasoconstrictor or balloon tamponades. Clinical course is shown in the flow chart in Fig. 1. Obliteration of gastric varices was achieved in 51% of the GVO group and 45% of the GVL group (P.78). The number of sessions required to obliterate gastric varices were 2.7 0.8 in the GVO group and 2.3 1.0 in the GVL group (P.60). Obliteration of esophageal varices was achieved in 20 of 26 patients (77%) in the GVO group and 15 of 24 patients (62%) in the GVL group (P.42). The number of sessions required to achieve esophageal variceal obliteration was 3.3 1.2 and 3.5 1.3 for the GVO and GVL group, respectively. Rebleeding from gastric varices occurred in 9 patients (31%) in the GVO group and 14 patients (54%) in the GVL group. Minor rebleeds occurred in 2 patients in both groups. The actuarial probability of rebleeding from gastric varices is shown in Fig. 2 (P.0005). The rebleeding rate was higher in the GVL group regardless of GOV1, GOV2, or IGV1 subgroup. Control of rebleeding was successful in 7 of the 8 patients (87%) in the GVO group and 6 of the 12 patients (50%) in the GVL group (P.10). Treatment failure was encountered in 4 patients (13%) in the GVO group and 11 TABLE 2. Results of Treatment GVO (n 31) GVL (n 29) P Control of active bleeding 13/15 (87%) 5/11 (45%).03 GOV1 11/13 (85%) 3/7 (43%).08 GOV2 2/2 (100%) 2/4 (50%).4 GV obliteration 16/31 (51%) 13/29 (45%).78 Sessions 2.7 0.8 (1-4) 2.3 0.7 (1-3).75 Rebleeding 9/29 (31%) 14/26 (54%).0005 GOV1 5/19 (26%) 10/18 (55%).14 GOV2 2/6 (33%) 3/7 (43%).58 IGV1 2/4 (50%) 1/1 (100%).60 Control of bleeding 7/8 (87%) 6/12 (50%).10 Blood transfusion (units) 2.6 0.9 4.2 1.3.01 FIG. 1. Clinical course of patients in both treatment groups. SRH, signs of recent hemorrhage. patients in the EIS group (38%)(P.05). Two patients in the GVO group and 3 in the GVL group were rescued using transjugular intrahepatic portosystemic stent shunts. Blood transfusion requirements were significantly higher in the GVL group than in the GVO group (4.2 1.3 vs. 2.6 0.9 units, respectively) (P.01). Table 3 shows complications and mortality rates of both groups. Six (19%) patients in the GVO group compared with 11 (38%) patients in the GVL group experienced significant complications (P.05). Although 2 patients bled from huge ulcers on gastric varices in the GVO group, a significantly higher prevalence of bleeding from the ulcers on gastric varices was noted in the GVL group than in the GVO group (P.03). Both patients with huge ulcers on gastric varices in the GVO group developed after 3 sessions of cyanoacrylate injection. In contrast, bleeding from the ulcers on gastric varices in the GVL group usually occurred in the first few days after GVL. One patient in the GVO group developed cerebral vascular accident 12 hours after injection. No complications could be attributed to EVL in both groups. A cumulative survival curve showed a significantly higher survival rate in the GVO group than in the GVL group (Fig. 3) (P.05). The majority of patients died of hepatic failure and variceal bleeding. Two (6%) patients in the GVO group compared with 7 patients (24%) in the GVL group died of gastric variceal bleeding (P.07). DISCUSSION It is generally accepted that bleeding from gastric varices is less frequently encountered than from esophageal varices. Approximately 14% to 36% of acute variceal bleeding arises from gastric varices. 21-23 However, bleeding from gastric varices is usually more severe than from esophageal varices. 6,19 It is noted that the bleeding from gastric varices requires greater amounts of blood transfusion and is more fatal than from esophageal variceal bleeds. 19,24 According to our experience, the number of patients presenting with gastric variceal bleeding have been increasing in recent years. This could be ascribed to an increasing awareness or a late complication of successful management of esophageal varices. 5,19,25 The search for a more effective method to reduce rebleeding and hence to enhance survival rate is urgently needed. Unfortunately, the optimal treatment for bleeding gastric varices is still unclear. Several modalities, including balloon tamponades, EIS, surgery, 26 and TIPS, 27 have been tried but none is well documented to be satisfactory.

HEPATOLOGY Vol. 33, No. 5, 2001 LO ET AL. 1063 FIG. 2. The actuarial probability of free from rebleeding from gastric varices was significantly higher in the GVO group than in the GVL group. FIG. 3. The actuarial probability of survival in the 2 treatment groups. In recent years, EVL has proven to be a viable substitute for EIS in the management of esophageal variceal bleeding. 4,5,12-14 EVL has been associated with fewer complications compared with EIS. Additionally, EVL can achieve variceal obliteration more rapidly, thereby resulting in lower rebleeding rates. It is still unknown whether ligation of gastric varices (GVL) is of similar effectiveness. Preliminary trials have shown that both endoscopic obturation using cyanoacrylate (Histoacryl) 7-11,28 and GVL 15 are promising tools in the management of gastric variceal bleeding. The mechanism of arresting variceal bleeding is different in both the modalities. Cyanoacrylate glue polymerizes rapidly and plugs the lumen if injected into a varix. 8 In contrast, ligation of varices results in strangulation and necrosis of varices. 29 To our knowledge, no controlled studies comparing ligation versus cyanoacrylate injection have ever been undertaken on the issue of gastric variceal bleeding. Our controlled trial showed that cyanoacrylate injection was superior to GVL in terms of arresting acute bleeding and prevention of rebleeding. Our control rate of acute gastric variceal bleeding was 87% and rebleeding rate was 31% in the GVO group. This is comparable with most trials, which showed hemostatic rates around 95% and rebleeding rates TABLE 3. Complications and Mortality GVO Group (n 31) GVL Group (n 29) P Complications 6 (19%) 11 (38%).05 Huge ulcer on GV 2* (6%) 0 (0%).70 Bleeding from ulcers on GV 2* (6%) 8 (28%).03 SBP 1 1 1 CVA 1 0 1 Bacteremia 2 2 1 Mortality 9 (29%) 14 (48%).05 Hepatic failure 3 3 1 Esophageal variceal bleeding 1 0 1 Gastric variceal bleeding 2 7.07 CVA 1 0 1 Ulcer bleeding 1 2.60 Sepsis 1 2.60 Abbreviations: GV, gastric varices; SBP, spontaneous bacterial peritonitis; CVA, cerebral vascular accident. * The 2 patients with huge ulcers were the same patients with bleeding ulcers. ranging between 10% to 50%. 7-11,28 However, the success rate of controlling active gastric variceal bleeding was 45% and the rebleeding rate was 54% in our GVL group. The higher rebleeding rate in the GVO group applied to GOV1, GOV2, and IGV1 subgroups. This was apparently inferior to a previous trial that showed that the hemostatic rate was 88% and the rebleeding rate was only 18%. 15 It cannot be simply ascribed to different techniques of expertise or different ligation devices. As shown in our previous trial of emergency EVL, we achieved a 97% success rate in the control of active esophageal variceal bleeding, and the rebleeding rate was only 17%. 30 There were slightly fewer patients belonging to Child-Pugh class A and more patients with hepatocellular carcinoma in the GVL group than the GVO group. However, this does not appear to explain the differences in effectiveness of both modalities. One possible reason to explain the discrepancy was the presence of acid milieu in the stomach. In our trial, we used mucosal protector sucralfate 31,32 to promote the healing of anticipated ulcers induced by endoscopic therapy, whereas Shiha et al. used proton pump inhibitor omeprazole to prevent ulcers induced by GVL. 15 As shown in our trial, a number of patients bled on the second or third day after GVL, which was rarely encountered in our patients receiving EVL. 5 Indeed, Lee et al. performed a preliminary trial of GVL and found that 4 of the 6 patients (67%) who received sucralfate had rebleeding whereas only 3 of 23 (13%) who received omeprazole experienced rebleeding. 33 The reason we did not use proton pump inhibitor was because of lack of sufficient data during the trial. Another factor that may influence the result of GVL was the presence of food materials in the stomach. It is possible that peristalsis of stomach with food materials may enhance sloughing of ligated tissue whereas it had little impact on the tissue injected with cyanoacrylate. On the other hand, one third of the patients in the trial by Shiha et al. had noncirrhotic portal hypertension. 15 The prognosis in such patients appeared much better than in those with other etiologies of portal hypertension, which thereby may have contributed to the lower rebleeding rate. 34 The interval between our GVL sessions was 3 to 4 weeks. It is possible that lower rebleeding can be achieved by shortening this interval. Nowadays, many centers use multiple ligators and no longer need an overtube for EVL. In this trial, a single ligator with an overtube was applied during EVL and GVL. In our experience, it is easier to perform ligation of gastric

1064 LO ET AL. HEPATOLOGY May 2001 varices using a single ligator rather than a multiple ligator. The overtube was used to facilitate endoscopic procedures. Like our previous trial, 30 the modified overtube did not result in any significant complication. On the other hand, it has been claimed that the use of overtube may protect the airway from aspiration and possibly reduce the chance of bacteremia during active bleeding. 4,35 The complication rates were higher in the GVL group than in the GVO group. This is because of the higher frequency of ulcers induced by ligation. Although GVO using cyanoacrylate may induce deep ulcers, the incidence was rather uncommon. Bleeding from gastric ulcers on the gastric varices was more frequently noted in the GVL group. This result appears to be contradictory to the results obtained from those carried out in patients with esophageal varices. 4,5,12-14 This may also be related to the hindrance of gastric acid on the healing of ulcers induced by ligation. Although cyanoacrylate injection has been found to be remarkably safe, severe complications related to embolization including cerebral stroke, pulmonary embolism, portal vein embolism, splenic infarction and retro-gastric abscess have been reported. 36 In our trial, a case of cerebral vascular accident occurred after receiving GVO and unfortunately the patient died. The follow-up period was shorter in our GVL group. This is because patients in the GVL group had a higher mortality rate due to the higher incidence of uncontrollable variceal bleeding. In conclusion, our trial showed that endoscopic obturation using cyanoacrylate was superior to GVL in the management of bleeding gastric varices in terms of controlling active bleeding, prevention of rebleeding, and incidence of complications, as well as enhancing survival. Further studies are justified to prove the efficacy and safety of GVL. REFERENCES 1. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-809. 2. Westaby D, Macdougall BR, Williams R. Improved survival following injection sclerotherapy for esophageal varices: final analysis of a controlled trial. HEPATOLOGY 1985;5:827-830. 3. Infante-Rivard C, Esnaola S, Villeneuve JP. Role of endoscopic variceal sclerotherapy in the long-term management of variceal bleeding: a metaanalysis. Gastroenterology 1989;96:1087-1092. 4. Stiegmann GV, Goff JS, Michaletz-Onody PA, Korula J, Lieberman D, Saeed ZA, Reville RM, et al. Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992; 326:1527-1532. 5. Lo GH, Lai KH, Cheng JS, Hwu JH, Chang CF, Chen SM, Chiang HT. A prospective, randomized trial of sclerotherapy versus ligation in the management of bleeding esophageal varices. HEPATOLOGY 1995;22:466-471. 6. Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices. Gastrointest Endosc 1986;32:264-268. 7. Oho K, Iwao T, Toyonaga A, Tanikawa K. Ethanolamine oleate versus butyl cyanoacrylate for bleeding gastric varices: a nonrandomized study. Endoscopy 1995;27:349-354. 8. Soehendra N, Nam VC, Grimm H, Kempeneers I. Endoscopic obliteration of large esophagogastric varices with bucrylate. Endoscopy1986;18: 25-26. 9. Ramond MJ, Valla D, Mosnier JF, Degott C, Bernuan J, Rueff B, Benhamou JP. Successful endoscopic obliteration of gastric varices with butyl cyanoacrylate. HEPATOLOGY 1989;10:488-493. 10. Binmoeller KF, Soehendra N. Superglue the answer to variceal bleeding and fundal varices. Endoscopy 1995;27:392-396. 11. Feretis C, Dimopoulos C, Benakis P, Kalliakmanis B, Apostolidis N. N-butyl-2-cyanoacrylate (Histoacryl) plus sclerotherapy versus sclerotherapy alone in the treatment of bleeding esophageal varices: a randomized prospective study. Endoscopy 1995;27:355-357. 12. Gimson AES, Ramage JK, Panos MZ, Hayllar K, Harrison PM, Williams R, Westaby D. Randomised trial of variceal banding ligation versus injection sclerotherapy for bleeding esophageal varices. Lancet 1993;342:391-394. 13. Laine L, El-Newihi HM, Migikovsky B, Slloane R, Garcia. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices. Ann Intern Med 1993;119:1-7. 14. Hou MC, Lin HC, Kuo BIT, Chen CH, Lee FY, Lee SD. Comparison of endoscopic variceal injection sclerotherapy and ligation for the treatment of esophageal variceal hemorrhage: a prospective randomized trial. HEPA- TOLOGY 1995;21:1517-1522. 15. Shiha G, El-Sayed SS. Gastric variceal ligation: a new technique. Gastrointest Endosc 1999;49:437-441. 16. Okuda K, Ohtsuki T, Obata H. Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Cancer 1985;56:918-928. 17. Pugh RNN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of esophagus for bleeding esophageal varices. Br J Surg 1973; 60:646-649. 18. Beppu K, Inokuchi K, Koyanagi N, Nakayama S, Sakata H, Kitano B, Kobayashi M, et al. Prediction of variceal hemorrhage by esophageal endoscopy. Gastrointest Endosc 1981;27:213-218. 19. Sarin SK, Lahoti D, Saxena SP, Murthy NS, Makwana UK. Prevalence, classification and natural history of gastric varices: a long-term follow-up study in 568 portal hypertension patients. HEPATOLOGY 1992;16:1343-1349. 20. Burroughs AK, Alexandrino P, Cales P, Fleig W, Grace N, Minoli G, Siringo S. Sore points A review of the points where there was disagreement at Baveno I, and an attempt to reach consensus. In: De Franchis R, ed. Portal Hypertension II. Proceedings of the Second Baveno International Consensus Workshop on Definitions, Methodology and Therapeutic Strategies on Portal Hypertension. Oxford: Blackwell Science, 1996:10-17. 21. Teres J, Cecilia A, Bordas JM, Rimola A, Bru C, Rodes J. Esophageal tamponade for bleeding varices. Controlled trial between the Sengstaken Blakemore tube and the Linton Nachlas tube. Gastroenterology 1978;75: 566-569. 22. Mitchell KJ, Macdougall BRD, Silk DBA, Williams R. A prospective reappraisal of emergency endoscopy in patients with portal hypertension. Scand J Gastroenterol 1982;17:965-968. 23. Korula J, Chin K, Young KO, Yamada S. The demonstration of two distinct subsets of gastric varices: observations during a seven-year study of endoscopic therapy. Dig Dis Sci 1991;36:303-307. 24. Bretagne JF, Dudicourt JC, Morisot D. Is endoscopic variceal sclerotherapy effective for the treatment of gastric varices? Dig Dis Sci 1986;31: A505S. 25. Lo GH, Lai KH, Lee SD, Tsai YT, Lo KJ. Does propranolol maintain post-sclerotherapy variceal obliteration? J Gastroenterol Hepatol 1993; 8:358-362. 26. Thomas PG, D cruz AJ. Distal splenorenal shunting for bleeding gastric varices. Br J Surg 1994;81:241-244. 27. Chau TN, Patch D, Chan YW, Nagral A, Dick R, Burroughs AK. Salvage transjugular intrahepatic portosystemic shunts: gastric fundal compared with esophageal variceal bleeding. Gastroenterology 1998;114:981-987. 28. Huang YH, Yeh HZ, Chen GH, Chang CS, Wu CY, Poon SK, Lien HC, et al. Endoscopic treatment of bleeding gastric varices by N-butyl-2-cyanoacrylate (Histoacryl) injection: long-term efficacy and safety. Gastrointest Endosc 2000;52:160-167. 29. Stiegmann GV, Cambre T, Sun JH. A new endoscopic elastic band ligating device. Gastrointest Endosc 1986;32:230-233. 30. Lo GH, Lai KH, Cheng JS, Lin CK, Huang JS, Hsu PI, Chiang HT. Emergency banding ligation versus sclerotherapy for the control of active bleeding from esophageal varices. HEPATOLOGY 1997;25:1101-1104. 31. Polson RJ, Westaby D, Gimson AES, Hayes PC, Stellon AJ, Hayllar K, Williams R. Sucralfate for the prevention of early rebleeding following injection sclerotherapy for esophageal varices. HEPATOLOGY 1989;10:279-282. 32. Lo GH, Lai KH, Cheng JS, Chen MH, Huang HC, Hsu PI, Lin CK. Endoscopic variceal ligation plus nadolol and sucralfate compared with ligation alone for the prevention of variceal rebleeding: a prospective, randomized trial. HEPATOLOGY 2000;32:461-465. 33. Lee TH, Lin JT. Application of variceal ligation to treat gastric varices: emphasis on use of acid suppression [Abstract]. Chinese Taiwan Gastroenterol Soc 1999 Autumn Convention & The International Symposium. FP 24. 34. Binmoeller KF, Borsatto R. Variceal bleeding and portal hypertension. Endoscopy 2000;32:189-199. 35. Lo GH, Lai KH, Shen MT, Chang CF. A comparison of the incidence of transient bacteremia and infectious sequelae after sclerotherapy and rubber band ligation of bleeding esophageal varices. Gastrointest Endosc 1994;40:675-679. 36. Binmoeller KF. Glue for gastric varices: some sticky issues. Gastrointest Endosc 2000;52:298-301.