The present study investigated to what extent measurements of wedged and free hepatic venous pressures adequately reflect the effects of vasopressin at the esophageal varices in patients with cirrhosis. Eleven patients undergoing therapeutic sclerotherapy were studied by measuring wedged hepatic venous pressure, intravariceal pressure, free hepatic venous pressure, superior vena cava pressure and the intravascular pressure gradients wedged hepatic venous pressure-free hepatic venous pressure and intravariceal pressure-superior vena cava pressure, prior to and after vasopressin injection (1 IU, iv). Vasopressin caused a significant reduction in intravariceal pressure (from 22.5 +/- 9.4 to 19.2 +/- 8.4 mm Hg, p less than 0.001). Measurement of wedged hepatic venous pressure and free hepatic venous pressure closely reflected the reduction in variceal pressure. Thus, wedged hepatic venous pressure decreased by 16 +/- 11%, which is close to the 14 +/- 7% change in intravariceal pressure, and the 23 +/- 12% fall in the pressure gradient wedged hepatic venous pressure-free hepatic venous pressure was mirrored by the 26 +/- 10% change in intravariceal pressure-superior vena cava pressure. These pressure gradients decreased more than the absolute pressures (intravariceal pressure and wedged hepatic venous pressure) due to concomitant increases in superior vena cava pressure (1.9 +/- 1.9 mm Hg) and free hepatic venous pressure (0.6 +/- 1.9 mm Hg).(ABSTRACT TRUNCATED AT 250 WORDS)