Caval backflow: a potential problem during blood sampling from the hepatic vein.

A proper measurement of splanchnic metabolism involves sampling blood from the hepatic vein without backflow contamination of blood from the caval vein. We have investigated the potential problem of caval backflow in human volunteers with an indwelling hepatic vein catheter by sampling blood with different amounts of suction on the syringe (ie, sampling speeds). We also investigated the potential problem in pigs in which a balloon catheter was inserted in the hepatic vein. Pure hepatic vein samples were obtained with the balloon inflated and compared with samples obtained from the same catheter in the conventional manner. In overnight fasted humans, drawing blood samples from the hepatic vein with minimal suction ("slow" drawing) resulted in glucose values 9.6% higher than drawing the samples with greater suction ("fast" drawing). The calculated arterial-venous balance across the splanchnic bed was 4.8 times greater with "slow" blood drawing as compared with "fast" drawing. Values obtained from the pigs showed no concentration differences between pure hepatic vein samples and "slow" drawing from the hepatic vein. The current study indicates that it is possible to obtain a "true" hepatic vein sample, but backflow from the caval vein is a potential pitfall that can have a physiologically significant impact on calculated balance data.