Doppler sonography of the portal vein and hepatic artery: measurement of a prandial effect in healthy subjects.

Abstract

To determine if the prandial effect outweighs variability in the measurement of portal venous blood flow and hepatic arterial resistive index. Fourteen healthy adult volunteers fasted overnight. For eight of the subjects, two sonographers performed three determinations of portal venous blood flow and hepatic arterial resistive index before and 30 minutes after a liquid meal (5 mg per kilogram of body weight). In the remaining six volunteers, determinations were made at the same time intervals but without a meal. The sonographers were blinded to on-screen measurements and the subject's prandial status. For subjects examined after the meal by sonographers A and B, respectively, portal venous blood flow increased from 144.2 to 201.7 mL/min and from 209.2 to 331.9 mL/min and hepatic arterial resistive index increased from 0.70 to 0.77 and from 0.67 to 0.78. After repeated-measures analysis of variance, ingestion of a meal was estimated to increase portal venous blood flow by 96.3 mL/min (P < .001)--a change in sonographers could affect the measurement by 76.7 mL/min (P < .001)--and to increase hepatic arterial resistive index by 0.089 (P < .001)--a change in sonographers did not affect the measurement (P > .1). The prandial effect on portal venous blood flow is only marginally greater than the interobserver variation in the measurement. Hepatic arterial resistive index also increases after a meal, but interobserver differences between sonographers are minimal; therefore, it is a more robust measurement.