Primary Prevention of Bleeding from Esophageal Varices
The New England Journal of Medicine — April 1, 1999 — Vol. 340, No. 13
Most patients with portal hypertension in North America and Europe have cirrhosis. Of those with compensated cirrhosis (i.e., cirrhosis with no ascites, encephalopathy, or severe jaundice), 30 percent have esophageal varices, as compared with 60 percent of those with decompensated cirrhosis. (1) The risk of a first episode of bleeding from esophageal varices is higher among patients with cirrhosis who have severe liver dysfunction and those whose varices are large or have red signs — so-called varices on varices — identified endoscopically. (2) Patients with cirrhosis may also have bleeding due to portal hypertensive gastropathy, nonvariceal mucosal abnormalities seen on endoscopy. These abnormalities account for at least 10 percent of first bleeding episodes. (2) The mortality due to bleeding among these patients is high: between 30 percent and 50 percent die within six weeks of the first bleeding episode. (1)
This dismal outcome has led to attempts both to identify those at high risk for bleeding and to prevent bleeding. Initial studies of the endoscopic characteristics originally thought to predict variceal bleeding — such as the presence of large varices or the presence of red signs (3) — proved to overestimate the risk of bleeding. In one subsequent study in which the severity of liver disease was assessed by modified Child’s criteria (with the classification based on ascites, encephalopathy, serum albumin and bilirubin concentrations, and prothrombin time), only 19 percent of 321 patients in the two categories at highest risk had an incidence of bleeding of 40 percent or more one year after initial endoscopy. (2) In another study, which included patients with cirrhosis who had varices at high risk as determined by endoscopy, only 18 percent of these patients had varices that bled within one year. (4)
Further data related to the risk of bleeding have come from studies of untreated groups in randomized trials designed to assess methods for the prevention of the first bleeding episode. Two therapies have been used for the prevention of first bleeding from varices: beta-adrenergic-antagonist drugs and variceal obliteration. Propranolol or nadolol has been compared with no active therapy or placebo in nine studies. A meta-analysis (5) found that the drugs reduced the incidence of bleeding, with a pooled odds ratio of 0.5 (95 percent confidence interval, 0.4 to 0.7); an average of 11 patients required treatment in order to prevent one episode of bleeding. The mortality rate was lower in the treatment group in all but one study. (5) The results of a second meta-analysis, (6) of data on 589 patients with at least medium-sized but mostly large varices (diameter, greater than or equal to 5 mm), all followed for two years, revealed an actuarial rate of bleeding of 22 percent among the patients treated with a beta-adrenergic-antagonist drug, as compared with 35 percent among the untreated patients (P=0.002). The rate of fatal bleeding was also lower among the treated patients (P=0.01), but overall survival was not.(6) The addition of isosorbide mononitrate may improve the effectiveness of medical therapy: in patients with red signs on their varices, irrespective of the size of the varices, or with varices 5 mm or more in diameter, bleeding occurred after an average follow-up of 30 months in 6 percent of patients treated with nadolol and isosorbide mononitrate and in 14 percent of those treated with nadolol alone. (7)
The other treatment option is endoscopic sclerotherapy, which has been compared with no treatment in 20 randomized trials that included 1756 patients, most of whom had either medium-sized or large varices. (5) In the sclerotherapy groups there was a significant reduction in bleeding in 5 trials, an increase in bleeding in 2 trials, and no difference in 13 trials. Although the pooled odds ratio for bleeding was 0.6 (95 percent confidence interval, 0.5 to 0.7), there was significant heterogeneity in the results with regard to bleeding and mortality among the different trials (P<0.001). (5) For example, the Veterans Affairs Cooperative Variceal Sclerotherapy Group study of 281 patients was stopped prematurely because mortality was higher in the sclerotherapy group than in the sham-treatment group (32 percent vs. 17 percent); the rates of bleeding were 22 percent and 17 percent, respectively. (8) The complications of sclerotherapy included bleeding from ulcers caused by the sclerosing procedure, dysphagia, esophageal strictures, and (rarely) perforation. (8) Given these results, nonselective beta-adrenergic-antagonist drugs are currently the treatment of choice for the primary prevention of bleeding from esophageal varices in patients with cirrhosis. (5)
The above-mentioned studies were performed before the advent of endoscopic ligation, which has replaced injection sclerotherapy as the endoscopic method of choice for the prevention of rebleeding (secondary prevention). In comparative trials and in a meta-analysis of studies including 547 patients, (9) ligation was more effective than sclerotherapy in preventing rebleeding, in part because it resulted in faster eradication of varices and had fewer complications.
The logical next step was to compare ligation of esophageal varices with drug therapy for the primary prevention of variceal bleeding. In a trial reported in this issue of the Journal, Sarin et al. (10) studied patients who had varices larger than 5 mm in diameter with one or more red signs, corresponding to the highest-risk categories. (2,3) After 18 months, the actuarial rate of bleeding in the ligation group was 15 percent. In contrast, the bleeding rate in the propranolol group was 43 percent, an unusually high rate. (10) The latter rate is similar to the 46 percent rate found by Sarin et al. at 18 months among untreated patients in a trial of esophageal ligation for primary prevention in patients selected according to the same criteria and followed for a similar period of time. (11)
The higher-than-expected rate of bleeding in the propranolol group in the current study by Sarin et al. is unexplained, particularly because the authors do not include information about bleeding due to portal hypertensive gastropathy. In addition, two other difficulties exist in interpreting the results. First, the mean dose of propranolol, 70 mg per day, was lower than that in previous trials, in which the mean dose was 123 mg per day (6); this low dose could in part explain the high rate of bleeding among patients in the propranolol group. Second, despite a significant reduction in bleeding in the ligation group as compared with the propranolol group, there was no significant difference between the two treatment groups either in overall mortality or in mortality due to bleeding, in contrast to the results of previous primary-prevention studies in which either drugs or sclerotherapy was used. (5,6)
Should the study by Sarin et al. change the clinical practice of giving propranolol for the primary prevention of variceal bleeding? Though the question is arguable, we think not, because of the difficulties in interpreting the results, described above, and because of the similarity of the results to those of some trials of sclerotherapy. (5) Moreover, the preliminary results of another recent trial suggest that ligation has no advantage over propranolol.(12)Finally, propranolol is very cheap and safe. We think that patients with cirrhosis should continue to be examined for varices by endoscopy and that those with varices should be considered for primary-prevention therapy. Patients deemed to be at high risk for bleeding should be given a nonselective beta-adrenergic drug, and for now, ligation should be reserved for patients who have contraindications to or intolerance of these drugs.
Andrew K. Burroughs, F.R.C.P.
David Patch, M.R.C.P.
Royal Free Hospital
London NW3 2QG, United Kingdom
1. D’Amico G, Luca A. Natural history: clinical-haemodynamic correlations: prediction of the risk of bleeding. Baillieres Clin Gastroenterol 1997;11:243-56.
2. The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices: a prospective multicenter study. N Engl J Med 1988;319:983-9.
3. Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal hemorrhage by oesophageal endoscopy. Gastrointest Endosc 1981;27:213-8.
4. Rigo GP, Merighi A, Chahin NJ, et al. A prospective study of the ability of three endoscopic classifications to predict hemorrhage from esophageal varices. Gastrointest Endosc 1992;38:425-9.
5. Pagliaro L, D’Amico G, Sorensen TI, et al. Prevention of first bleeding in cirrhosis: a meta-analysis of randomized trials of nonsurgical treatment. Ann Intern Med 1992;117:59-70.
6. Poynard T, Cales P, Pasta L, et al. Beta-adrenergic-antagonist drugs in the prevention of gastrointestinal bleeding in patients with cirrhosis and esophageal varices — an analysis of data and prognostic factors in 589 patients from four randomized clinical trials. N Engl J Med 1991;324:1532-8.
7. Merkel C, Marin R, Enzo E, et al. Randomised trial of nadolol alone or with isosorbide mononitrate for primary prophylaxis of variceal bleeding in cirrhosis. Lancet 1996;348:1677-81.
8. The Veteran Affairs Cooperative Variceal Sclerotherapy Group. Prophylactic sclerotherapy for esophageal varices in men with alcoholic liver disease — a randomized, single-blind, multicenter clinical trial. N Engl J Med 1991;324:1779-84.
9. Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of esophageal variceal bleeding: a meta-analysis. Ann Intern Med 1995;123:280-7.
10. Sarin SK, Lamba GS, Kumar M, Misra A, Murthy NS. Comparison of endoscopic ligation and propranolol for the primary prevention of variceal bleeding. N Engl J Med 1999;340:988-93.
11. Sarin SK, Guptan RK, Jain AK, Sundaram KR. A randomized controlled trial of endoscopic variceal band ligation for primary prophylaxis of variceal bleeding. Eur J Gastroenterol Hepatol 1996;8:337-42.
12. Stanley AJ, Forrest EH, Lui HF, et al. Band ligation versus propranolol or isosorbide mononitrate in the primary prophylaxis of variceal haemorrhage: preliminary results of a randomised controlled trial. Gut 1998;42:Suppl 1:A19. abstract.
Copyright © 1999 by the Massachusetts Medical Society. All rights reserved.