Following the demonstration that somatostatin lowered portal pressure in cirrhotic patients with portal hypertension, 2 uncontrolled reports suggested that the hormone might be useful in the control of acute variceal haemorrhage. Subsequently, a number of randomised controlled trials have indicated that somatostatin may have an efficacy as good as or better than either vasopressin or combined vasopressin and nitroglycerin therapy and is associated with fewer side effects. Somatostatin has an efficacy comparable to balloon tamponade, histamine-2-receptor antagonists and injection sclerotherapy. One double-blind randomised controlled trial demonstrated a significant benefit of somatostatin over placebo in the control of variceal bleeding whereas a second did not show any significant difference between treatments. In all the controlled trials, the average control rate achieved with somatostatin administration was 69% and it was not associated with any major side effects. Somatostatin administration has also been shown in uncontrolled series to be very effective in controlling postinjection sclerotherapy bleeding from the varices per se, and from oesophageal ulcers and oesophagitis. Few data are available on the long acting analogue of somatostatin, octreotide, but preliminary data suggest that it may be as effective and safe as the native hormone in controlling the acute variceal bleeding and postinjection sclerotherapy haemorrhage. It is concluded that there may be a case for instituting somatostatin therapy as soon as the patient enters hospital to facilitate sclerotherapy, and for continuing treatment for 5 days after sclerotherapy when the risk of recurrent bleeding is highest.