In unselected cirrhotic patients with bleeding gastric varices to compare the influence on mortality rate, duration of life, control of bleeding, quality of life, and economic costs of treatment of: portacaval shunt, endoscopic variceal sclerotherapy and/or variceal ligation.

BACKGROUND: In patients with cirrhosis, bleeding gastric varices (BGV) are associated with a high mortality rate, but have received much less investigation than bleeding esophageal varices. Various therapeutic measures have been used to treat BGV, including endoscopic, radiographic, and surgical procedures, but there have been few prospective evaluations of therapy involving sizable groups of patients that have received acceptable follow-up. Management of this serious disorder has been uncertain and often unsuccessful. Herein is a prospective randomized controlled trial in unselected cirrhotic patients with BGV that compared the effectiveness of endoscopic therapy (ET) and portacaval shunt (PCS) during follow-up for more than 5 years or until death.

STUDY DESIGN: 518 unselected patients with cirrhosis and BGV were randomized to ET or PCS performed as an emergency in 220 and electively in 298. All patients received the same diagnostic workup, initial therapy, post-treatment therapy, and rigorous follow-up. One-, 5-, 10-, and 15-year follow-up rates were 100%, 97%, 97%, and 92%, respectively. ET consisted of repetitive sessions of intravariceal injection sclerotherapy and/or variceal band ligation aimed at variceal obliteration. PCS consisted of a direct anastomosis, side to side in 95%. ET and PCS were compared specifically with regard to control of bleeding, survival rate, and quality of life.

Inclusion Criteria:

• All patients with upper gastrointestinal bleeding (blood in the upper gastrointestinal tract) of a magnitude that required 2 or more units of blood transfusion and entered the emergency room directly, or were referred from an area hospital, or developed bleeding while in the hospital, and were shown to have cirrhosis of the liver, and were shown by endoscopy to have bleeding gastric varices, absence of bleeding from esophageal varices, and absence of any other lesion that could reasonably account for the bleeding were included ("all comers").

Exclusion Criteria:

• None