Starting point David Patch and colleagues (Gastroenterology 2002; 123: 1013-19) randomised 102 patients surviving a variceal bleeding to EBL or drug therapy with propranolol with the addition of ISMN if target reductions in portal pressure (evaluated by the hepatic venous pressure-gradient [HVPG]) were not achieved at 3 months. Overall, results of drug therapy were similar to those of EBL (44% vs 54% rebleeding at 1 year). There were no differences in survival or nonbleeding complications. Christophe Bureau and colleagues (Hepatoloy 2002; 36: 1361-66) treated 34 patients with cirrhosis and portal hypertension with propranolol and measured HVPG after a median of 4 days. Target HVPG reductions were achieved in 13 "responders". ISMN was added in the 21 "non-responders" and HVPG, measured again: seven more patients achieved target HPVG reduction. Re-bleeding rates were lower in responders than in nonresponders (10% vs 64%). Both studies suggest that drug therapy-can be improved by adding ISMN to b blockers in those patients with an insufficient decrease in HVPG.

Where next? Long-term drug therapy is emerging as effective treatment for the prevention of variceal rebleeding. The role of HVPG monitoring as a guide to identifying patients requiring further treatment needs to be further evaluated. Trials to determine the best treatment for, patients who do not respond to drug therapies are also required.