BACKGROUND/AIMS: A problem in pharmacotherapy for bleeding varices in portal hypertension is non-responders. The aim of this study was to elucidate the features of hemodynamic response to vasopressin in the gastroesophageal collateral vein in patients with esophageal varices.

METHODS

Flow velocity in the portal and the collateral left gastric vein was measured with an echo-Doppler flowmeter before and during infusion of vasopressin, 0.2 U/min, in 41 patients with cirrhosis and esophageal varices.

RESULTS

The decrease in flow velocity in the left gastric vein with vasopressin (-29 +/- 25%) was significantly smaller than that in the portal vein (-56 +/- 20%). There was no or only minimal change in flow velocity in the left gastric vein in 39% of the patients, especially in those with large-size varices. In 28 patients examined by portal catheterization, changes in flow velocity in the left gastric vein were correlated with portal pressure, and portal pressure in non-responders was significantly higher than that in responders (non-responders: 363 +/- 49, responders: 312 +/- 41 mmH2O, p < 0.05).

CONCLUSIONS

It was concluded that hepatofugal blood flow in the gastroesophageal collateral is not readily reduced by vasopressin. However, as the study was performed in a stable condition without variceal bleeding, whether these hemodynamic features will apply during acute variceal bleeding in patients who are known to have a poor hemodynamic response to vasopressin remains to be elucidated.