Bleeding from esophageal or gastric varices is the major complication of portal hypertension in cirrhotic patients and carries a high mortality. If it could be prevented, one might expect survival to improve. Portacaval shunts were the first type of prophylactic therapy for variceal bleeding to be evaluated in four controlled randomized clinical trials. Unfortunately, survival was not improved even though bleeding was prevented (1-3). This failure was attributed to a combination of immediate operative mortality and later deaths due to hepatic failure, the shunt contributing to the hepatic failure in these patients. In recent years, sclerotherapy of esophageal varices and pharmacological treatment of portal hypertension (P-blockade) have renewed the possibility of treating patients' varices prophylactically. In preliminary studies, both appear to be safer than shunt surgery in the short-term, and they appear to have fewer longterm complications, such as hepatic encephalopathy. However, several factors make it difficult to interpret their value in the prophylaxis of variceal haemorrhage and create problems in the design of new studies.
AIMS OF PROPHYLACTIC THERAPY
The primary aim of prophylactic therapy for esophageal varices is to improve survival by preventing the first bleeding episode. A secondary aim might be to reduce bleeding even if mortality was not reduced, as was the case with prophylactic shunt surgery. It might be considered a legitimate aim in view of the distressing nature of variceal hemorrhage. However, evaluation would require measurement of the quality of life in cirrhotic patients, and the effect on this of both the disease and the prophylactic treatment might prove difficult to assess in practice. We believe, therefore, that trials of prophylaxis for variceal bleeding should use survival as the major criterion for therapy.
NATURAL HISTORY
When considering prophylactic treatment of varices in cirrhotic patients, it is important to know the natural