The shunt's patency is often compromised by dysfunction (due to occlusion or stenosis of the intrahepatic or outflow tract).^₈

Trans-shunt venography (shunt portography or portogram) continues to be the gold standard for TIPS evaluation. Nevertheless it is an expensive and invasive procedure. Doppler ultrasonography is the most common non-invasive technique used to assess TIPS patency. It is the screening tool of choice to detect early shunt dysfunction.

TIPS dysfunction criteria include:

• >50% reduction of the shunt lumen.
• Portosystemic gradient (PSG) >12-15 mmHg.
• TIPS occlusion.⁸

Reported cumulative TIPS dysfunction rates range between: 17 to 73% within the first 6 months, 23 to 87% within 12 months, and 80 to 83% within 24 months.⁸

Causes of TIPS Dysfunction

Causes of TIPS dysfunction in shunts created with bare stents include:

• Bile-related
• Non-bile-related
• Hepatic vein stenosis

Biliary duct trasection during TIPS placement has been associated with TIPS stenosis and occlusion, through the creation of biliary-TIPS fistulas. The content of bile acids, salts, cholesterol and phospholipids induces a proinflammatory and thrombogenic environment (granulomatous inflammatory response). Biliary-TIPS fistulas increase the risk of acute thrombosis and recurrent occlusions. 

Non-bile-related dysfunction arises due to the differentation of hepatic fibroblasts into myofibroblasts, which migrate into the TIPS from the hepatic parenchyma, causing tissue overgrowth (fibrotic healing response) within the shunt's lumen⁹ (see Figure 6).

Fig. 6: Causes of TIPS dysfunction

The stent must be extended to the hepatic vein - vena cava junction to avoid intimal hyperplasia and hepatic vein stenosis. Leaving the proximal segment of the hepatic vein unstented increases the risk of developing intimal hyperplasia.

Extrahepatic or hemodynamic causes of TIPS dysfunction, such as flow theft from varices or mesocaval shunts may lead to stasis and shunt thrombosis.

Hepatic vein peripheral segment approach: results in a perpendicular crossing of the hepatic vein and reduces shunt flow. Peripheral puncture of the portal vein causes intra-stent kinking, which leads to stenosis and an increased portosystemic gradient. Venogram examination plays an important role in directing the needle from an appropiate hepatic vein to a portal vein branch.⁹ The use of pre-TIPS magnetic resonance (MR) or computed tomography (CT) imaging provides important anatomic information (variants) that aids in the approach strategy (see Figure 7).

Fig. 7: Pre-TIPS evaluation. A) Patient with Budd-Chiari syndrome, Nakamura Type A portal vein (not shown). B) Patient with portal hypertension & collaterals, Nakamura Type C portal vein.

Expandable polytetrafluoroethylene (PFTE) covered stents have shown improved shunt patency. These stents don't appear to provoke an inflammatory response (preventing myofibroblasts and the extracellular collagen matrix from reaching the shunt's lumen) and are impermeable to bile.⁹

Most TIPS dysfunctions with expandable PFTE covered stents are related to technical and mechanical factors.

TIPS Follow-up

Pre-TIPS Doppler Ultrasound (within the month leading up to placement)¹ should be performed.

Baseline measurements should be obtained: 1 week after placement of wallstents, or between 1 week to 1 month after the placement of PFTE-covered stents (to avoid the ultrasonographic artifact generated by the presence of bubbles within the TIPS lumen).²

Follow-up should be performed every 3-6 months. Since PFTE-covered stents have shown improved long-term shunt patency,¹ frequent routine surveillance may not be cost-effective for this type of stents⁹ (see Figures 8, 9 and 10).

Fig. 8: A) TIPS placement in a patient with Child B cryptogenic cirrhosis. Spectral Doppler US in middle third of TIPS. B) Baseline study 15 days after placement with an intra-TIPS velocity of 121.7 cm/s & C) 1 month after with an intra-TIPS velocity of 89.8 cm/s.

Fig. 9: Same patient as in Figure 8. A) Intra-TIPS velocity reduction to 40cm/s after 9 months. B) Velocity reduction to 16cm/s after 15 months.

Fig. 10: Same patient as in Figures 8 and 9. A) Patient underwent dilation with balloon angioplasty. (B) A control study performed 3 months later showed a partial increase in velocity (35cm/s). Another control study performed 21 months after the dilation revealed a velocity of 29cm/s (not shown).

Ultrasonographic Criteria of TIPS Dysfunction

TIPS Occlusion

Diagnosis of TIPS occlusion with Doppler ultrasonography is simple and has a very high sensitivity and specificity (100%) if neither color nor duplex signal is present within the shunt's lumen (see Figure 11).

Fig. 11: Occluded TIPS with no color signal within the TIPS shunt

Color Doppler Ultrasonography and Stenosis

Color-mode intra-TIPS filling defects correspond to pseudointimal hyperplasia within the shunt's lumen. However, this sign is imprecise and has no correlation with the PSG. It has a 26% sensitivity and 50% specificity⁸ (see Figure 12).

Fig. 12: (A) B-flow technique US showing reduced flow in cephalic portion of TIPS. (B) Increase in velocity up to 247cm/s (C) Normal velocity of 95.5cm/s after TIPS dilation. A >50% stenosis was confirmed by CT (not shown).

Doppler Ultrasonography and Intrahepatic Portal Flow Direction:

Post-TIPS distal intrahepatic flow changes its direction from hepatopetal to hepatofugal in 53-78% of cases.

A change back to hepatopetal flow is a sign of dysfunction. This sign has a high specificity (81-100%)⁸ (see Figure 13). 

Fig. 13: A) Triplex Doppler US showing retrograde flow in portal branches (permeable TIPS). B) Anterograde flow observed 6 months later, due to TIPS dysfunction. Increased ascites was also noted.

Doppler Ultrasonography and Hepatic Vein Stenosis

Inverted hepatic vein flow is an indirect sign of stenosis, with a low sensitivity of 29% (see Figure 14).

Fig. 14: Patient with an occluded TIPS (cephalic and middle portions). A) Color Doppler US and B) Triplex Doppler of the middle hepatic vein, showing inverted flow and loss of normal waveforms.

Portal and Intra-TIPS velocities

Any degree of portal or intra-TIPS velocity reduction is an indirect sign of dysfunction.

Shunt velocities ≤90cm/s or ≥190cm/s, or a temporal increase or decrease in shunt velocities >50cm/s are direct signs of TIPS dysfunction² (see Figure 15). 

A main portal venous velocity <30 cm/s is an indirect sign of an altered intra-TIPS velocity. 

Portal velocity reductions ranging from 30-50% are also signs of TIPS dysfunction (see Figures 16, 17, 18, 19 and 20).

Fig. 15: (A,B). Baseline US with normal velocities. (C,D). Follow-up at 6 months showing diminished velocities. (E,F). Follow-up at 12 months showing occluded TIPS.

Fig. 16: A) Baseline US with normal portal velocity. B) US at 6 months with a 49% reduction in velocity as a sign of dysfunction.

Fig. 17: Thirty-one-year-old male with a history of polycythemia vera and hepatic vein thrombosis whom received a TIPS due to refractory ascites. Control ultrasound 3 months after placement. A) Intra-TIPS middle third spectrum with a velocity of 27cm/s (direct sign of dysfunction). B) Extra-hepatic portal vein spectrum with hepatopetal flow and a velocity of 20cm/s (indirect sign of dysfunction). C) Left portal branch spectrum: sustained hepatofugal flow

Fig. 18: Same patient as in Figure 17. A) Follow-up at 2 months. Venous phase CT: Occlusion in proximal third of TIPS secondary to thrombosis. B) Thrombolysis one day after diagnosis. C) TIPS-plasty using a 10mm diameter x 10cm length balloon, 2 days after occlusion.

Fig. 19: Same patient as in Figures 17 and 18. Metallic vascular stent (10x70mm) placed in cephalic portion of TIPS after balloon angioplasty. (Placing the stent up to the hepatic vein to vena cava junction in the initial procedure minimizes intimal hyperplasia and hepatic vein stenosis).

Fig. 20: Same patient as in Figures 17, 18 and 19. A) Follow-up CT (venous phase) one day after stent placement (permeable). Follow-up after 6 months with triplex Doppler ultrasound showing permeable TIPS and normal velocities in B) middle and C) cephalic portions.

What Velocity to use?

Multiple velocity thresholds have been proposed to distinguish normal from dysfunctioning TIPS; however, the velocity spectrum of patent shunts varies widely, from 50-300 cm/s.⁸

Benito et al. compared different proposed velocity thresholds obtained from Doppler ultrasounds with corresponding portograms to establish accurate criteria for TIPS dysfunction⁸ (see Figure 21).

Fig. 21: Different intra-TIPS velocity thresholds proposed by other authors with their reported sensitivities and specificities. Sensitivities and specificities obtained from the Benito, Bilbao, Hernández et al series appear in red, comparing MTITV to venography results.
References: Benito A, Bilbao J, Hernández T, Cuesta AM, Larrache J, González I, Vivas I. Doppler Ultrasound for TIPS: Does it Work? Abdom Imaging 2004;29(45):45-52

In their study they found that the best middle shunt velocity threshold was 98cm/s, with a sensitivity of 46% and a specificity of 79%, demostrating that intra-TIPS velocity is useless as an isolated parameter.⁸ 

The best portal velocity threshold as an indirect sign of TIPS dysfunction was 22cm/s, with a sensitivity of 55% and a specificity of 74%.⁸

Best Parameters Evaluating the Probability of Dysfunction

A decrease in maximum middle third intra-TIPS velocity >6% and a portal velocity <22 cm/s were the best parameters to evaluate the probability of dysfunction. Consequently, three "probability of TIPS dysfunction" categories were proposed using these criteria:

• Low
• Intermediate 
• High⁸ (see Figure 22).

Fig. 22: Sensitivity, specificity, PPV, and NPV percentages relative to venography for the three proposed categories based on a portal velocity <22 cm/s and a decrease >6% in maximum middle third intra-TIPS velocity.
References: Benito A, Bilbao J, Hernández T, Cuesta AM, Larrache J, González I, Vivas I. Doppler Ultrasound for TIPS: Does it Work? Abdom Imaging 2004;29(45):45-52

How to Avoid Ultrasonographic Pitfalls

Diminished velocities in the TIPS' portal end

The portal end frequently projects towards the portal vein lumen (blood flows freely since it is not occluded by the liver). These velocities reflect portal velocity. Examine the intrahepatic portion of the portal end (see Figure 23).¹⁰

Fig. 23: A) Portal vein triplex Doppler US before TIPS, with a velocity of 36.1cm/s. B) Portal vein triplex Doppler US after TIPS (intra-TIPS caudal segment), with a velocity of 65cm/s. Measure at this level to not underestimate velocities.

Intraluminal echoes on greyscale are not always thrombi    

The intra-TIPS echogenicity may be increased - particularly in prominently curved segments - resulting in the false image of a thrombus. Check other windows, use color, and measure segment velocities (see Figure 24).¹⁰

Incomplete intraluminal color saturation

The absence of color in an intraluminal segment may mimic a filling defect (thrombosis or intimal hyperplasia). Check color settings and flow repetition frequencies; measure velocities before, within, & after the segment (see Figure 25).¹⁰

Fig. 25: A) TIPS color Doppler US showing areas of turbulence without saturation at the cephalic end. B) Color Doppler from a different angle showing adequate saturation and permeability. C) Triplex Doppler US demonstrating flow in middle and cephalic portions.

Prognostic factors after TIPS placement

Child-Pugh Class C, variceal hemorrhage requiring emergent TIPS placement, serum billirubin >3mg/dL, ALT >100 IU/L and pre-TIPS encephalopathy not related to bleeding are associated with an increased mortality risk after TIPS placement. ¹

A MELD score higher than 18 predicts higher mortality at 3 months after TIPS placement. It has been found to be superior to the Child-Pugh score.¹¹