Acute Hemodynamic Response to β-Blockers and Prediction of Long-term Outcome in Primary Prophylaxis of Variceal Bleeding

Background & Aims

Studies of variceal bleeding have shown that a hemodynamic response to treatment of portal hypertension is appropriate when the hepatic venous pressure gradient (HVPG) decreases below 12 mmHg or by >20% from baseline. However, in primary prophylaxis, many nonresponders do not bleed and 2 invasive procedures are needed to assess response. We investigated the long-term prognostic value of an acute response to β-blockers and whether the target reduction in HVPG can be improved in primary prophylaxis.

Methods

An initial hemodynamic study was performed in patients with large varices and without previous bleeding. After baseline measurements were made, propranolol was administered intravenously and measurements were repeated 20 minutes later. Patients were given nadolol daily and a second hemodynamic study was performed.

Results

Of 105 patients, 15% had variceal bleeding. Using receiver operating characteristic curve analysis, a decrease of HVPG ≥10% was the best value to predict bleeding. In the initial study, 75 patients (71%) were responders (HVPG decreased to ≤12 mmHg or by ≥10%) and had a lower probability of first bleeding than nonresponders (4% vs 46% at 24 months; P < .001). Acute responders also had a lower risk of developing ascites (P = .001). Chronic responders had a lower probability of bleeding than nonresponders (P < .001). There was a correlation between acute and chronic changes in HVPG (r = 0.62; P = .01).

Conclusion

The acute hemodynamic response to β-blockers can be used to predict the long-term risk of first bleeding. An HVPG reduction >10% from baseline is the best target to define response in primary prophylaxis.

Abbreviations used in this paper: HR, hazard ratio, HVPG, hepatic venous pressure gradient, NIEC, North Italian Endoscopic Club, PPV, positive predictive value