Clinical vs haemodynamic response to drugs in portal hypertension

Background/Aims: The combination of non-selective beta-blockers and nitrates is an effective therapy for the prevention of rebleeding from oesophageal varices. However, a significant number of patients fail to respond and have further haemorrhage. It has been suggested that measurement of the hepatic venous pressure gradient response to long-term drug therapy may allow early selection of non-responders. We aimed to test this hypothesis in 63 patients with cirrhosis and variceal bleeding treated with propranolol±isosorbide mononitrate. Methods: Hepatic venous pressure gradient was measured before and during treatment. Response was defined as a reduction of 20% or more in hepatic venous pressure gradient, or a fall in hepatic venous pressure gradient to 12 mmHg or less. Results: Forty-four patients were evaluable: 28 responders and 16 non-responders. Hepatic venous pressure gradient fell significantly in the responder group (17.5±0.5 mmHg vs 12.2±0.5 mmHg; p<0.01) but not in the non-responders (18.0±1.0 vs 17.9 ±1.2 mmHg; p=n.s.). Overall, there was no difference in rebleeding rates between the two groups: responders 43%, non-responders 25%. However, rebleeding was uncommon in compliant patients with alcoholic cirrhosis, in whom the hepatic venous pressure gradient fell to less than 12 mmHg (9%). Conclusions: In this study a fall in hepatic venous pressure gradient at 20% was not a reliable predictor of clinical response. A threshold value of 12 mmHg was useful, but applied to relatively few patients.