Randomized Comparison of Ligation Versus Ligation Plus Sclerotherapy in Patients With Bleeding Esophageal Varices

GASTROENTEROLOGY 1996;110:529 533 Randomized Comparison of Ligation Versus Ligation Plus Sclerotherapy in Patients With Bleeding Esophageal Varices LOREN LAINE, CRAIG STEIN, and VISHVINDER SHARMA Division of Gastrointestinal and Liver Diseases, University of Southern California School of Medicine, Los Angeles, California dium tetradecyl sulfate proximal to the ligature) could See editorial on page 635. lead to more rapid eradication of varices than the use of ligation alone. They hypothesized that injection of a Background & Aims: Combining endoscopic sclerother- small amount of sclerosant into the stagnant variceal apy with ligation has been proposed to hasten variceal channel above the band would hasten obliteration witheradication. A randomized trial was performed comout increasing side effects. Subsequently, Koutsomanis, 7 paring combination ligation plus sclerotherapy with liin a small controlled trial presented only in abstract gation alone in patients with major bleeding from form, reported that combined therapy in 9 patients led esophageal varices. Methods: Forty-one patients were randomly assigned to receive ligation or ligation plus to variceal eradication in an average of only 1.2 treatment 1 ml 1.5% tetradecyl injected just above each band. sessions, as compared with a mean of 4.2 sessions in 7 Treatment was repeated weekly until varices were erad- patients in whom sclerotherapy alone was used. icated. Repeat endoscopy was performed for rebleedferred form of endoscopic treatment for patients with Because ligation has replaced sclerotherapy as our pre- ing and every 3 months after eradication. Results: No significant differences were found between combined bleeding esophageal varices, we sought to determine if therapy and ligation in rebleeding (29% vs. 30%), blood treatment with ligation plus sclerotherapy held any adtransfused (3.1 { 1.1 vs. 2.0 { 0.8 U), hospital days vantages over the use of ligation alone. We therefore (9.3 { 2.1 vs. 7.5 { 1.2), complications (29% vs. performed a prospective, randomized comparison of liga- 10%), or deaths (14% vs. 15%) during a mean followtion plus sclerotherapy vs. ligation. Our primary aim up period of 28 weeks. Combined therapy required sigwas to assess whether combined therapy led to variceal nificantly more sessions to achieve eradication (4.9 { eradication more quickly than ligation alone. Secondary 0.6 vs. 2.7 { 0.4) and greater time per treatment sesaims included comparisons of efficacy (e.g., rebleeding sion (18.3 { 1.7 vs. 13.3 { 0.5 minutes). Conclusions: Combined ligation plus sclerotherapy does not reduce mortality), safety (complications), and the time required the number of treatment sessions required for variceal for the treatment sessions. eradication as compared with ligation alone. Combined Materials and Methods therapy lengthens the time required for treatment without improving efficacy or decreasing complications. Patients Thus, combined ligation and sclerotherapy should not Patients with chronic liver disease were eligible if they be used to treat patients with bleeding esophageal had (1) witnessed hematemesis, bloody nasogastric aspirate, varices. melena, or hematochezia; (2) systolic blood pressure 90 mm Hg, heart rate 110/min, or orthostatic changes in blood ndoscopic variceal ligation is now considered by pressure of 20 mm Hg or in heart rate of 20/min; transfu- Emany physicians to be the treatment of choice for sion of 2 U of blood; or a decrease in hematocrit of 6% within 12 hours; and (3) endoscopy within 24 hours of admission patients with bleeding esophageal varices. Individual showing active variceal bleeding or grade 2 4 esophageal varicontrolled trials indicate that ligation is at least as safe ces (grading scale of Korula et al. 8 ) without other potential and effective as endoscopic sclerotherapy, 1 4 and a recent bleeding lesions in the upper gastrointestinal tract (patients meta-analysis 5 found that endoscopic ligation had sig- with gastric varices were therefore not eligible). nificant advantages over sclerotherapy in terms of re- Patients were excluded if they had received endoscopic therbleeding, mortality, esophageal stricture formation, and apy for varices in the past 6 months, if they had hepatocellular number of sessions required for variceal obliteration. carcinoma or other malignancy, if they were unable or unwill- In an attempt to further improve the results achieved with ligation, Reveille et al. 6 suggested that ligation 1996 by the American Gastroenterological Association combined with low-volume sclerotherapy (1 ml of so- 0016-5085/96/$3.00

530 LAINE ET AL. GASTROENTEROLOGY Vol. 110, No. 2 ing to sign the informed consent, or if they were homeless. using a t test. Proportional data were compared with the Fisher The study was performed between April 1993 and January Exact Test. A t-tailed P value of õ0.05 was considered signifi- 1995 at the Los Angeles County / University of Southern cant. Before beginning the study, we determined that we California Medical Center. The protocol was approved by the would need a sample size of 80 patients to show a difference Research Committee of the Los Angeles County / University of 1.5 treatment sessions between the study groups (with an of Southern California Medical Center, and all participants a value of 0.05 and a power of 0.80). However, an interim signed a written consent form before endoscopic examination analysis after half of the projected patients were enrolled and treatment. showed that we had a significant difference between the study groups in the direction opposite that predicted by our primary Treatment hypothesis. This effectively ruled out the possibility that our Ligation was performed using the endoscopic ligating results would support our hypothesis, even with the full comdevice and plastic esophageal overtube (Bard Interventional plement of patients, and led us to terminate the study at this Products, Tewksbury, MA). Ligation was begun in the region point. of the gastroesophageal junction with subsequent ligatures applied more proximally. Up to 8 bands were placed per session. Results In patients receiving combined therapy, after each liga- Seventy-six patients meeting the inclusion criteria ture was placed, 1 ml of 1.5% sodium tetradecyl sulfate was were seen during the study period. Forty-one patients injected into the varix immediately above the ligature. We were enrolled in the study, and 35 were excluded for the chose 1 ml of sclerosant per injection site to reproduce the following reasons: 14 had sclerotherapy in the preceding technique of Reveille et al. 6 Three percent sodium tetradecyl 6 months, 10 were unable to give consent, 6 were homesulfate was diluted in a 1:1 ratio with sterile normal saline to less, 3 refused to participate, 1 had hepatocellular carciachieve the 1.5% dilution. The time required for the endonoma, and the investigators were not called for 1 eligible scopic treatment of the varices was recorded at each procedure. Patients did not receive adjunctive therapy, such as octreotide patient. Six patients had a history of prior episodes of or vasopressin, in concert with endoscopic treatment in this upper gastrointestinal bleeding: (1) variceal bleeding 3 study. weeks before entry with no endoscopic therapy, (2) Mal- Eligible patients were randomly assigned to receive ligation lory Weiss tear 1 year before entry, (3) ulcer bleeding or ligation plus sclerotherapy at the time of the initial endoscopy 5 years before entry, (4) variceal bleeding 5 years before according to a computer-generated randomization se- entry with no follow-up endoscopy or treatment, (5) ulcer quence. Treatment was repeated weekly until variceal eradica- bleeding 7 years before entry, and (6) unknown source tion was achieved. If circumferential ulceration or stricture of bleeding 10 years before entry with no follow-up enwas found, the treatment was withheld. In patients with less doscopy or treatment. extensive ulceration, treatment was delivered to areas not in- Selected characteristics at entry are shown in Table 1. volved by the ulcers. After obliteration, patients had repeat The mean age of the participants was 47 years, and alcoendoscopies performed every 3 months or for any episode of rebleeding. Recurrent varices were treated with the originally assigned form of therapy. Table 1. Selected Characteristics of the Study Groups at Analysis Randomization End points of patient follow-up were death, treatment Ligation plus failure, definitive therapy for portal hypertension (e.g., porta- sclerotherapy (n Å 21) caval shunt, transjugular intrahepatic portosystemic shunt), or Ligation (n Å 20) patient lost to follow-up. Rebleeding was ascribed to a treat- Age (yr) 46 { 2 48 { 3 ment-induced ulcer when active bleeding was found or stig- Men/women 15/6 15/5 Alcoholic cirrhosis (%) 15 (71) 16 (80) mata of recent hemorrhage were present in the ulcer base. Child Pugh score 9.3 { 0.5 9.2 { 0.5 Treatment failure was defined as (1) death related to esophageal Child Pugh class A (%) 1 (5) 3 (15) variceal bleeding or a complication of bleeding or endoscopic Child Pugh class B (%) 11 (52) 8 (40) treatment; (2) three rebleeding episodes due to esophageal Child Pugh class C (%) 9 (43) 9 (45) varices or treatment-induced ulcers that required hospitaliza- Hematocrit (%) 21.9 { 1.3 24.7 { 1.5 Blood transfusions (U) 2.5 { 0.3 a 1.4 { 0.3 tion and transfusion; or (3) further bleeding in a single hospi- Active bleeding (%) 4 (19) 4 (20) talization that failed to respond to endoscopic therapy and Grade of variceal size (%) required transfusion of at least 6Uofblood after endoscopic 2 4 (19) 4 (20) therapy. An esophageal stricture was diagnosed if a patient 3 11 (52) 13 (65) reported dysphagia and had endoscopic evidence of narrowing; 4 6 (29) 3 (15) dilation was performed for all esophageal strictures. Red color signs (%) 18 (86) 16 (80) Quantitative data (expressed as mean { SE) were compared a P õ 0.05.

February 1996 LIGATION VERSUS LIGATION PLUS SCLEROTHERAPY 531 Table 2. Results of Treatment in the Study Groups despite repeated endoscopic treatments, 1 patient was considered a treatment failure due to multiple rebleeding Ligation/sclerotherapy Ligation (n Å 21) (n Å 20) episodes in the hospital requiring more than 6 U of blood transfusion, and 1 patient died from hepatic failure Index hospitalization Recurrent bleeding (%) 2 (10) 3 (15) without recurrent bleeding. Blood transfusions (U) 2.8 { 1.2 1.9 { 0.8 Six patients receiving combined therapy developed Hospital days 8.0 { 2.1 6.6 { 1.2 complications: 2 had esophageal strictures (requiring 1 Treatment failure (%) 2 (10) 3 (15) Complications (%) 3 (14) 1 (5) and 2 dilations each), 2 had bleeding esophageal ulcers, Deaths (%) 3 (14) 3 (15) 1 had bacterial peritonitis, and 1 had a cardiorespiratory Entire study arrest at the initial endoscopy. Two patients in the liga- Recurrent bleeding (%) 6 (29) 6 (30) Due to esophageal tion group had complications: 1 developed bacterial perivarices 4 (19) 5 (25) tonitis and 1 had a bleeding esophageal ulcer. Complica- Due to esophageal tions contributed to death in 2 of the patients in the ulcers 2 (10) 1 (5) combined therapy group (cardiorespiratory arrest at en- Blood transfusions (U) 3.1 { 1.1 2.0 { 0.8 Hospital days 9.3 { 2.1 7.5 { 1.2 doscopy and persistent bleeding from esophageal ulcers) Treatment failure (%) 2 (10) 3 (15) and none of the patients in the ligation group. Deaths (%) 3 (14) 3 (15) Variceal eradication was achieved in 12 (60%) of Complications (%) 6 (29) 2 (10) Follow-up (wk) 28 { 6 27 { 6 the 20 patients in the ligation group and 15 (71%) of 21 in the combined therapy group (Table 3). However, variceal eradication occurred in all patients who received hol was the cause of the liver disease in approximately more than three treatment sessions (12 in the three quarters of the patients. The mean Child Pugh ligation and 15 in the combined therapy group). The score 9 and the proportion of patients in Child Pugh mean number of treatment sessions required for eradi- classes A, B, and C were similar in the two groups. At the cation was 2.7 { 0.4 in the ligation group and 4.9 { time of entry, the combined therapy group had received a 0.6 in the combination therapy group (P Å 0.005). In significantly greater mean number of units of blood. Four addition, a treatment session was prevented due to exten- patients in each group had active bleeding at initial enligation sive ulcerations 5 times in a total of 5 patients in the doscopy. Hemostasis was achieved in all 4 patients in group and 11 times in 8 patients in the com- each group with endoscopic therapy. bined therapy group (these endoscopies were not in- Results in the two study groups during the initial cluded in the treatment session calculations). The mean hospitalization are shown in Table 2. No significant difsignificantly number of bands used during the first treatment was not ferences were identified between the two groups in outalone, different in the 2 treatment groups (ligation come parameters including rebleeding, transfusions, 6.9 { 0.3; combined therapy, 6.1 { 0.4; P Å length of hospital stay, or mortality. Similarly, no sigment 0.14), whereas the mean number of bands used per treatnificant differences were found between the study groups session up to the time of eradication was signifi- in these parameters during the entire course of the study, a mean follow-up period of 28 weeks (Table 2). Table 3. Characteristics of Endoscopic Treatment in the Four patients in the combination therapy group rebled Study Groups after their initial hospitalization; all were hospitalized Ligation plus again, but only 2 received blood transfusions (2 and 4 sclerotherapy Ligation U). Three patients in the ligation group had recurrent (n Å 21) (n Å 20) bleeding; they were rehospitalized, but only 1 received Variceal eradication (%) 15 (71) 12 (60) a blood transfusion (3 U). Treatment sessions to eradication 4.9 { 0.6 a 2.7 { 0.4 All treatment failures and deaths occurred during the Bands per treatment (to eradication) 3.8 { 0.3 a 5.3 { 0.3 initial hospitalization. In the combined therapy group, 1 Total treatment time to patient had persistent bleeding from treatment-induced eradication (min) 85.8 { 12.8 a 37.1 { 6.3 esophageal ulcers, underwent a transjugular intrahepatic Time per treatment (min) (all sessions) 18.3 { 1.7 a 13.3 { 0.5 portosystemic shunt, and eventually died with multior- Time per band placed (min) (all gan system failure and bleeding; 1 patient died after a sessions) b 4.6 { 0.3 a 3.0 { 0.2 cardiorespiratory arrest that occurred during the initial a P õ 0.01. endoscopy; and 1 patient died from hepatic failure. In the b Time per band placed includes placement of band and injection of ligation group, 2 patients died with persistent bleeding sclerosant in the ligation plus sclerotherapy group.

532 LAINE ET AL. GASTROENTEROLOGY Vol. 110, No. 2 cantly higher in the ligation group than the combined quired significantly more time for performance. Therefore, therapy group (5.3 { 0.3 vs. 3.8 { 0.3; P Å 0.002). combined therapy with ligation and sclerotherapy The time required to perform combined ligation and does not seem to have any advantages over the use of sclerotherapy was significantly greater than the time to ligation alone, and its use cannot be recommended. perform ligation alone (Table 3). The mean value for We attempted to reproduce the technique proposed total minutes of treatment to eradication was 37.1 { by Reveille et al. 6 by injecting 1 ml of sclerosant above 6.3 for ligation and 85.8 { 12.8 for combined therapy each band placed. Despite our use of sodium tetradecyl (P Å 0.003). When all treatment sessions were assessed, sulfate in a concentration of 1.5% (higher than the 1% the mean time per session was 13.3 { 0.5 minutes for concentration used by Reveille et al. 6 ), we were unable ligation and 18.3 { 1.7 minutes for combined therapy to confirm their hypothesis of more rapid obliteration of (P Å 0.01), whereas the mean time per ligature placed varices with combined therapy. We cannot rule out the was 3.0 { 0.2 minutes for ligation and 4.6 { 0.3 mi- possibility that injection of greater volumes of sclerosant nutes for combined ligation and sclerotherapy (P õ may have hastened variceal obliteration, but it seems 0.001) (this time, per ligature included the time for likely that increased volume might also increase the rate placement of the band and injection of sclerosant in the of local complications. combined therapy group). The explanation for the increased number of sessions to Discussion achieve eradication with combined therapy is uncertain. Fewer ligatures were used per treatment session in the Bleeding esophageal varices continue to be a diftensive combined therapy group than in the ligation group. Exficult management problem. Patients bleeding from ulceration was more common in the group receiv- esophageal varices have higher rebleeding rates, transfunumber ing sclerotherapy with ligation, and this did limit the sion requirements, lengths of hospitalization, and a of bands placed in some cases. In addition, it is greater risk of death than do patients bleeding from possible that the significantly longer period of time re- nonvariceal sources. 10 Endoscopic therapy decreases but quired to perform combined therapy as compared with by no means eliminates the morbidity and mortality ligation alone may have led the endoscopists to terminate associated with variceal bleeding. For example, approxifewer some combination treatment sessions after applying mately 50% of patients treated with sclerotherapy will bands than they applied in the ligation group. have recurrent bleeding during a follow-up period of 6 Preliminary results from another recent trial compar- months or more, and approximately one third may die. 5 ing combination ligation plus sclerotherapy with ligation Although, compared with sclerotherapy, ligation seems alone also failed to show any benefit of combined therto lessen rebleeding, mortality, local complications, and apy 11 ; complications were significantly more common in the number of treatment sessions required to achieve patients treated with combined therapy. On the other variceal eradication, 5 further improvements in therapy hand, a recent abstract comparing ligation plus sclero- are still desirable. therapy with sclerotherapy alone indicated that combined We studied the combination of ligation and sclerotherapy therapy required fewer treatment sessions than sclerotherapy with the hypothesis that this therapy would alone to achieve variceal eradication, 12 support- achieve more rapid eradication of varices than the use of ing the results of Koutsomanis. 7 Because ligation is ligation alone. A more rapid obliteration of varices would known to require fewer treatment sessions than sclero- be attractive for several reasons. Fewer treatment sessions therapy, 5 it is not surprising that combined ligation plus should translate into lower costs as well as less discomfort sclerotherapy would result in more rapid eradication than and risk for the patient. In addition, more rapid eradicacontradictory sclerotherapy alone. Thus, these latter studies are not tion of varices would be likely to reduce variceal rebleeding, to the results of our trial. which occurs most commonly before varices disap- The failure to improve the treatment of variceal bleeding pear. by adding sclerotherapy to ligation requires us to However, the results of our study showed that com- turn our attention to other strategies that have the potential bined ligation and sclerotherapy did not reduce the number to improve outcome in patients with bleeding varibined of treatment sessions required for variceal eradication. ces. The use of intravenous octreotide during the first Rather, the number of treatments was increased with days of hospitalization combined with endoscopic treatcombined therapy. No benefit was found with combined ment seems to hold promise. 13,14 Endoscopic treatment therapy in any of the outcome parameters, such as rebleeding, with tissue adhesives (e.g., cyanoacrylate), not available transfusions, length of hospital stay, treatment in the United States, may also be useful, especially in the failure, or death. Furthermore, combined therapy re- difficult-to-treat patient with bleeding gastric varices. 15

February 1996 LIGATION VERSUS LIGATION PLUS SCLEROTHERAPY 533 In summary, combined ligation plus sclerotherapy does not reduce the number of treatment sessions revariceal sclerotherapy. Hepatology 1985;5:584 589. quired for variceal eradication as compared with ligation 9. Pugh RNH, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. alone. Combined ligation plus sclerotherapy significantly Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646 649. lengthened the time required for treatment sessions com- 10. Silverstein FE, Gilbert DA, Tedesco FJ, Buenger NK, Persing H. pared with ligation alone without improving efficacy or The national ASGE survey on upper gastrointestinal bleeding. II. decreasing complications. Thus, combined endoscopic Clinical prognostic factors. Gastrointest Endosc 1981;27:80 therapy with ligation and sclerotherapy does not have 93. 11. Saeed ZA, Stiegmann GV, Ramirez RC, Reville M, Goff JS, Hepps advantages over ligation alone and should not be used KS, Cole RA. Endoscopic variceal ligation versus combined ligation and low dose sclerotherapy: a multi-center randomized to treat patients with esophageal variceal hemorrhage. trial 8. Korula J, Balart LA, Radvan G, Zweiban BE, Larson AW, Kao HE, Yamada S. A prospective randomized trail of chronic esophageal (abstr). Am J Gastroenterol 1994;89:1626. References 12. Jensen DM, Kovacs TOG, Jutabha R, Randall G, Cheng S, Jensen ME, Gornbein J. Initial results of a randomized, prospective study 1. Stiegmann GV, Goff JS, Michaletz-Onody PA, Korula J, Lieberman of combination banding & sclerotherapy vs. sclerotherapy alone D, Saeed ZA, Reveille M, Sun JH, Lowenstein SR. Endoscopic for hemostasis of bleeding esophagogastric varices (abstr). Gassclerotherapy as compared with endoscopic ligation for bleeding trointest Endosc 1995;41:351. esophageal varices. N Engl J Med 1992;326:1527 1532. 13. Besson I, Ingrand P, Person B, Boutroux D, Heresbach D, Bernard 2. Laine L, El-Newihi HM, Migikovsky B, Sloane R, Garcia F. Endo- P, Hochain P, Larricq J, Nousbaum JB, Ribard D, Mostefa Kara scopic ligation compared with sclerotherapy for the treatment of N, Legoux JL, Pillegand B, Becker MC, Di Costanzo J, Métreau bleeding esophageal varices. Ann Intern Med 1993;119:1 7. JM, Silvain C, Beauchant M, Multicentric Group, France. A com- 3. Gimson AES, Ramage JK, Panos MZ, Hayllar K, Harrison PM, parison of octreotide combined with emergency sclerotherapy to Williams R, Westaby D. Randomised trial of variceal banding sclerotherapy alone in the treatment of acute variceal bleeding: ligation versus injection sclerotherapy for bleeding esophageal a randomized double-blind placebo-controlled trial (abstr). Gas- varices. Lancet 1993;342:391 394. troenterology 1995;108:A1034. 4. Lo GH, Lai KH, Cheng JS, Hwu CH, Chang CF, Chiang HT. A 14. Sung JY, Chung SCS, Leung VKS, Li MKK, Lau JYW, Li AKC. prospective, randomized trial of sclerotherapy versus ligation in Octreotide as an adjuvant therapy to endoscopic variceal ligation the management of bleeding esophageal varices. Hepatology for acute variceal hemorrhage (abstr). Gastrointest Endosc 1995;22:466 471. 1995;41:358. 5. Laine L, Cook D. Endoscopic ligation compared with sclerother- 15. Khandelwal M, Jones WF, Akerman P, Raijman I, Siemens M, apy for treatment of esophageal variceal bleeding: a meta-analyof Kandel G, Kortan P, Marcon NE, Haber GB. Preliminary results sis. Ann Intern Med 1995;123:280 287. a prospective randomized trial of histoacryl vs endoscopic 6. Reveille RM, Goff JS, Stiegmann GV, Stauffer JT. Combination band ligation (EBL) for acute esophago-gastric variceal hemorendoscopic variceal ligation (EVL) and low-volume endoscopic rhage (abstr). Gastrointest Endosc 1994;40:P76. sclerotherapy (ES) for bleeding esophageal varices: a faster route to variceal eradication (abstr)? Gastroinest Endosc 1991;37: Received July 3, 1995. Accepted September 18, 1995. 243. Address requests for reprints to: Loren Laine, M.D., Gastrointesti- 7. Koutsomanis D. Endoscopic variceal ligation combined with lowvolume sclerotherapy: a controlled study (abstr). Gastroenterology 1992;102:A835. nal Division (LAC 12-137), Department of Medicine, University of Southern California School of Medicine, 2025 Zonal Avenue, Los Angeles, California 90033. Fax: (213) 226-7573.