Detection of transjugular intrahepatic portosystemic shunt dysfunction: value of duplex Doppler sonography

Recent reports have shown that a high percentage of patients with transjugular intrahepatic portosystemic shunts (TIPS) have postprocedural shunt complications, including thrombosis of the stent, stenosis of the stent, or stenosis of the hepatic vein draining the stent. We did a prospective study to determine the utility of Doppler sonography as a screening technique for the detection of these complications.

From September 1991 to September 1992 we placed TIPS in 45 patients. After the procedure, patients were routinely evaluated with both Doppler sonography and angiography. The sonographic protocol consisted of insonation of the stent, portal vein, and hepatic vein to determine the presence of flow, peak velocity, and direction of flow. The angiograms were evaluated for stenoses of the stent or hepatic vein that caused an increase in the portosystemic pressure gradient greater than 15 mm Hg, increased intrahepatic portal venous filling, retrograde filling of the draining hepatic vein, or opacification of varices. The sonographic findings were statistically evaluated to determine if sonography could demonstrate the complications shown by angiography.

Adequate follow-up was obtained in 29 of the 45 patients. Sixteen of the 29 patients had shunt complications that consisted of one stent thrombosis, three stent stenoses, nine hepatic vein stenoses, and three concomitant stenoses of the stent and hepatic vein. Flow was not detected by sonography in the stent of the patient with thrombosis. There was a significant difference (p = .003) between the temporal change in peak stent velocity in patients with stenoses versus those without. Use of a change (increase or decrease) in peak stent velocity greater than 50 cm/sec from the post-TIPS baseline sonogram as the diagnostic criterion for the detection of shunt stenoses resulted in a 93% sensitivity and 77% specificity. Five patients with stenosis had reversed flow in the draining hepatic vein. Only one patient with a stenosis had a peak stent velocity less than 50 cm/sec.

Our results suggest that Doppler sonography is an excellent noninvasive screening technique for the detection of complications of TIPS. We have found a temporal change in peak stent velocity greater than 50 cm/sec to be a more sensitive sonographic sign of TIPS stenosis than the previously reported low-velocity parameters. Our experience suggests that nearly all complications of TIPS can be detected by using three criteria: (1) no flow for thrombosis, (2) a temporal change in peak stent velocity greater than 50 cm/sec for stent and/or hepatic vein stenosis, and (3) reversed flow in the hepatic vein draining the stent for hepatic vein and, rarely, stent stenosis.