To evaluate the feasibility of determining patency of the transjugular intrahepatic portosystemic shunt (TIPSS) by non-invasive CT angiography (CTA).(1) Non-enhanced scanning of the shunt. (2) Bolus tracking by injecting 20 ml of non-ionic contrast material through a cubital vein access to determine the time to maximal shunt enhancement. (3) Contrast-enhanced spiral CT study applying a delay according to the time to peak of the shunt, 3 mm collimation, 5 mm table feed and 3 mm reconstruction interval. (4) 3D and multiplanar reconstructions. (5) Evaluation of the questions: intrahepatic shunt patent or not; evidence of intimal hyperplasia; evidence of stenosis and potential location. (6) Transjugular portography via the stent. (7) Comparison of angiographic findings and CT morphology.Eight patients had inconspicuous CTA. Four of them had a normal shunt at angiography, four had slight intimal hyperplasia. No intervention was necessary in these patients. CTA of three patients showed intimal hyperplasia (lumen reduction between 10% and 50%). The diagnosis was angiographically confirmed in all cases. Due to a high portosystemic gradient intervention was required in all. In five patients CTA and angiography showed a stenosis (reduction of shunt lumen > 50%). All required a revision including stent placement or PTA of the shunt tract. Four shunts were occluded; all occlusions were shown both in CTA and angiography.None of the shunts with normal findings at CTA required revision. All shunts conspicuous on CTA resulted in revision. In this study, CTA turned out to be an accurate, non-invasive method to evaluate the patency of TIPSS.