Portal flow and arterioportal shunting after transjugular intrahepatic portosystemic shunt creation

PURPOSE: It was postulated that a transjugular intrahepatic portosystemic shunt (TIPS) produces arterioportal shunting and accounts for reversed flow in the intrahepatic portal veins (PVs) after creation of the TIPS. This study sought to quantify this shunting in patients undergoing TIPS creation and/or revision with use of a direct catheter-based technique and by measuring changes in blood oxygenation within the TIPS and the PV. MATERIALS AND METHODS: This prospective study consisted of 26 patients. Median Model for End-stage Liver Disease and Child-Pugh scores were 13 and 9, respectively. Primary TIPS creation was attempted in 21 patients and revision of failing TIPS was undertaken in five. In two patients, TIPS creation was unsuccessful. All TIPS creation procedures but one were performed with use of polytetrafluoroethylene-covered stent-grafts. Flow within the main PV (Q(portal)) was measured with use of a retrograde thermodilutional catheter before and after TIPS creation/revision, and TIPS flow (Q(TIPS)) was measured at procedure completion. The amount of arterioportal shunting was assumed to be the increase between final Q(portal) and Q(TIPS), assuming Q(TIPS) was equivalent to the final Q(portal) plus the reversed flow in the right and left PVs. Oxygen saturation within the TIPS and the PV was determined from samples obtained during TIPS creation and revision. RESULTS: Mean Q(portal) before TIPS creation was 691 mL/min; mean Q(portal) after TIPS creation was 1,136 mL/min, representing a 64% increase (P = .049). Mean Q(TIPS) was 1,631 mL/min, a 44% increase from final Q(portal) (P = .0009). Among cases of revision, baseline Q(portal) was 1,010 mL/min and mean Q(portal) after TIPS revision was 1,415 mL/min, a 40% increase. Mean Q(TIPS) was 1,693 mL/min, a 20% increase from final Q(portal) (P = .42). Arterioportal shunting rates were 494 mL/min after TIPS creation and 277 mL/min after TIPS revision, representing 30% of total Q(TIPS) after TIPS creation and 16% of Q(TIPS) after TIPS revision. No increase in oxygen tension or saturation was seen in the PV or TIPS compared with initial PV levels. Q(TIPS) did not correlate with the portosystemic gradient. CONCLUSION: TIPS creation results in significant arterioportal shunting, with less arterioportal shunting seen among patients who undergo TIPS revision. Further work is necessary to correlate Q(TIPS) with the risk of hepatic encephalopathy and liver failure.