Current issues of ACP Journal Club are published in Annals of Internal Medicine


Propranolol reduced bleeding during endoscopic sclerotherapy

ACP J Club. 1992 Sept-Oct;117:37. doi:10.7326/ACPJC-1992-117-2-037

Related Content in the Archives
Propranolol to prevent recurrent bleeding from portal hypertension
Sclerotherapy for prophylaxis of hemorrhage from esophageal varices increased mortality

Source Citation

Vinel JP, Lamouliatte H, Cales P, et al. Propranolol reduces the rebleeding rate during endoscopic sclerotherapy before variceal obliteration. Gastroenterology. 1992 May;102:1760-3.



To compare the effectiveness of endoscopic sclerotherapy alone with endoscopic sclerotherapy plus propranolol in reducing the rate of recurrence of bleeding before variceal obliteration in patients with cirrhosis and variceal bleeding.


Randomized controlled trial.


Hospital in France.


74 patients (mean age, 54 y; 59 men) hospitalized with cirrhosis and upper gastrointestinal hemorrhage endoscopically proven to originate from ruptured esophageal varices. Exclusion criteria were previous treatment with propranolol, endoscopic sclerotherapy, or shunt or deconnection surgery; hepatocellular carcinoma; and contraindication to β-blockers. No patients were lost to follow-up.


The initial bleeding episode was treated with balloon tamponade, vasopressin infusion, and blood transfusions as required. As soon as patients were considered to be hemodynamically stable, they were randomized to endoscopic sclerotherapy alone (n = 35) or endoscopic sclerotherapy plus propranolol (n = 39). The propranolol dose was adjusted to decrease resting heart rate by 25% or to a maximum dose of 320 mg/d (mean dose, 105 mg/d).

Main outcome measure

Recurrence of bleeding, defined as any gastrointestinal hemorrhage necessitating a transfusion of ≥ 2 units of blood.

Main results

7 of 39 patients (18%) in the endoscopic sclerotherapy plus propranolol group had recurrence of bleeding compared with 14 of 35 patients (40%) in the endoscopic sclerotherapy alone group ( P < 0.05). {This absolute risk reduction of 22% means that 5 patients would need to be treated with sclerotherapy plus propranolol (rather than sclerotherpy alone) to prevent 1 additional case of rebleeding 95% CI 2 to 67; the relative risk reduction was 55%, CI 5% to 80%.}* The difference remained significant after adjusting for the initial size of esophageal varices. A total of 25 rebleeding episodes occurred, 17 of which originated from ruptured esophageal varices (13 in the endoscopic sclerotherapy alone group vs 4 in the combined therapy group, P < 0.01). The mean time from enrollment in the study to recurrence of bleeding was longer in the combined therapy group compared with the endoscopic sclerotherapy alone group (40.9 vs 17.6 d, respectively, P < 0.05). Mean total blood requirement per patient was lower in the combination therapy group compared with the endoscopic sclerotherapy alone group (1.4 vs 2.8 units of blood, respectively, P < 0.01).


Propranolol reduced the rate of recurrence of bleeding during endoscopic sclerotherapy in patients with cirrhosis and variceal hemorrhage.

Source of funding: Not stated.

Address for article reprint: Dr. J.P. Vinel, Service d'hépato-gastro-entérologie, Centre Hospitalier Universitaire Purpan, 31059 Toulouse Cedex, France.

*Numbers calculated from data in article.


In this study, Vinel and colleagues showed that sclerotherapy plus propranolol is more effective than sclerotherapy alone in reducing the rate of recurrence of bleeding from esophageal varices in patients with cirrhosis. Theoretically, combined therapy should be more effective than single therapy. Propranolol lowers portal pressure pharmacologically, whereas sclerotherapy obliterates varicose veins in the esophagus, the usual site of bleeding in these patients with cirrhosis.

Two earlier trials that compared sclerotherapy plus propranolol to sclerotherapy alone gave conflicting results. Jensen and Krarup (1) noted a 27% decrease in bleeding with combined therapy compared with sclerotherapy alone whereas Westaby and colleagues (2) noted no difference. The discordant findings in the study of Westaby and colleagues is probably a result of differences in the patient populations; only 36% of patients had alcoholic cirrhosis and 17% did not have cirrhosis. This finding contrasts with 89% of the patients in Vinel's study having confirmed alcoholic cirrhosis. A second difference in these trials is that most patients in Westaby's study bled from mucosal ulcerations, whereas most patients in Vinel's study bled from ruptured varices. Finally, patients in Westaby's study were more seriously ill—30% died during the study. In Vinel's study, 16 of 91 potential study patients died before randomization and were not included in the trial; of the 74 patients who were randomized and followed, 10 died (5 in each group). Exclusion of the former group of patients, however, makes comparison between the trials difficult.

The data presented by Vinel and colleagues are convincing enough to recommend the addition of propranolol to sclerotherapy in hemodynamically stable patients with gastrointestinal hemorrhage from esophageal varices. Careful consideration of the contraindications to β-blocker therapy is warranted for patients at risk for hypovolemic shock.

Marshall M. Kaplan, MD
Tufts UniversityBoston, Massachusetts, USA


1. Jensen LS, Krarup N. Scand J Gastroenterol. 1989;24:339-45.

2. Westaby D, Melia W, Hegarty J, et al. Hepatology. 1986;6:673-5.