A new method for the measurement of intrahepatic shunts

After transhepatic portal pressure determination, 96 patients were assessed for the presence of intrahepatic shunts by injection of microspheres (25 +/- 5 micron diameter) into the portal vein using RISA-131I as an indicator of dilution. Multiple portal vein injections in each patient allowed blood sampling from the hepatic vein (site 1) and from two inferior vena cava sampling sites (site 2, at the junction of the hepatic vein orifice with the inferior vena cava, and site 3, 2 to 3 cm closer to or within the right atrium). Intrahepatic shunting was calculated from each site: hepatic vein in 57 patients and inferior vena cava, site 2 in 43 patients and site 3 in 77 patients. At least one valid IHS calculation was available in 92 of the patients. Intrahepatic shunting calculated from sequential portal vein injections with sampling from the hepatic vein was highly correlated (r = 0.98, p less than 0.0001, slope = 1.0), with a mean difference of 1.9% +/- 1.9%. There was no significant difference by t test comparison of the mean IHS calculated from sites 1, 2, and 3. The IHShv was correlated with the IHSivc (site 2) (r = 0.79, p less than 0.0001, slope = 1.0) and IHSivc (site 3) (r = 0.82, p less than 0.0001, slope = 2.1). Occasional marked discrepancies were noted between IHS calculated from site 1 or site 2 compared with site 3, and the site 3 calculation was always greater. A shunt index in all patients included shunts calculated from the hepatic vein in 57 patients plus shunt calculation from the inferior vena cava in the remaining patients (site 2 in 26 patients and site 3 in nine). A control group with minimal chronic liver disease (10 patients) had a portal pressure (greater than IVC) of 4.1 +/- 1.4 mm Hg and shunt index of 0.5% +/- 0.6%. The 82 patients with portal hypertension or chronic liver disease had a higher portal pressure, 13.8 +/- 4.6 mm Hg, and a significantly greater shunt index, 13.7% +/- 24.5% (p less than 0.0001) compared with controls. The frequency distribution of IHS in patients with chronic liver disease demonstrated less than 2% IHS in 49% of patients and less than 5% IHS in 63%. The validity of our methods and the implications of the infrequent demonstration of a large IHS are discussed.