Background: The hepatic venous pressure gradient (HVPG) is used to evaluate portal hypertension.
Methods: We measured HVPG in two separate liver veins in 169 liver vein catheterizations in 102 cirrhosis patients and in 27 patients with no liver disease (controls).
Results: In the controls, the two measurements differed by 0.0 +/- 1.8 mmHg (mean +/- s, n = 27), upper 95% confidence limit 3.6 mmHg (mean + 2 s). HVPG ranged from -0.1 to 8.3 mmHg, upper 95% confidence limit 6.7 mmHg. In cirrhosis, the two measurements agreed within +/- 3.6 mmHg in 39%. In 61%, the measurements differed by 4-34 mmHg. In 35%, fluoroscopy demonstrated hepatic vein-to-hepatic-vein shunting in veins with low HVPG values. In some patients with HVPG measurements above 30 mmHg, Doppler ultrasound examination showed arterialization of the hepatic vasculature.
Discussion: Our results demonstrate a hitherto unrecognized notable heterogeneity of the intrahepatic vasculature and HVPG measurements in cirrhosis. The presumption of interposition of non-flowing blood between the catheter tip and the portal system for the measurement of HVPG may thus be violated in about one-third of the cirrhosis cases because of abnormal outlet into hepatic venous shunts and in a minor fraction because of abnormal arterial inlet. In 26%, one measurement was below 12 mmHg, the other measurement above. If the HVPG had been measured in only one liver vein, 13% of the cases would have been classified in a lower risk group than appropriate according to the 12 mmHg concept of risk of bleeding from oesophageal varices.