Patient selection is paramount for TIPS creation. Predictors of poor outcome after TIPS include various scoring systems, emergent versus elective procedures, and comorbidities. The Child-Pugh score was originally used to predict mortality after TIPS and is still often used. Multiple studies have shown small differences in the ability of the Child-Pugh score and the model for end-stage liver disease (MELD) score to predict mortality post-TIPS (Salerno et al. 2002; Schepke et al. 2003; Ferral et al. 2004a, b). Due to the “ceiling effect” of the Child-Pugh score, the MELD score may be more easily applied (Schepke et al. 2003; Ferral 2005). The MELD score is specifically designed to predict mortality after elective TIPS and has been validated as a predictor of mortality (Malinchoc et al. 2000; Kamath et al. 2001; Angermayr et al. 2003; Montgomery et al. 2005; Pan et al. 2008). Studies have shown that patients undergoing elective TIPS with a MELD score < 18 have significantly lower mortality rates compared to patients with MELD scores ≥ 18 (Angermayr et al. 2003; Ferral et al. 2004 a,b). In the emergent setting or with high risk patients, a multidisciplinary approach in evaluation of risks and benefits of TIPS creation is recommended (Lopera 2005; Boyer et al. 2010).

Controlled trials have established the efficacy of TIPS for secondary prevention of variceal bleeding and refractory cirrhotic ascites (Rössle et al. 2000; Salerno et al. 2007; Zheng et al. 2008; Boyer et al. 2010; García-Pagán et al. 2010, 2013; Bai 2014). Additional indications for TIPS are based on uncontrolled series. These indications include refractory acutely bleeding varices, portal hypertensive gastropathy, gastric antral vascular ectasia, refractory hepatic hydrothorax, hepatorenal syndrome type 1 and 2, Budd-Chiari syndrome, veno-occlusive disease, and hepatopulmonary syndrome. Note that primary prevention of variceal bleeding is currently not an indication for TIPS (Boyer et al. 2010; Gaba et al. 2012; Copelan et al. 2014).

Due to increased cardiac preload after TIPS creation, absolute contraindications to the procedure include severe congestive heart failure, tricuspid regurgitation, and severe pulmonary hypertension. Systemic infection and sepsis are also absolute contraindications. Relative contraindications related to anatomic conditions increasing the technical difficulty of TIPS creation include presence of multiple hepatic cysts, primary or metastatic hepatic malignancy, and unrelieved biliary obstruction. Due to shunting of portal blood flow away from the liver, severe uncontrolled hepatic encephalopathy and rapidly progressive liver failure are also relative contraindications to TIPS creation. Other relative contraindications include uncorrectable severe coagulopathy and severe thrombocytopenia (Haskal et al. 2003; Boyer et al. 2010; Copelan et al. 2014). Portal vein thrombosis has historically been a relative contraindication to TIPS creation and liver transplantation; however, techniques for TIPS creation in patients with portal vein thrombosis are described and can potentially allow patients to achieve transplant candidacy (Blum et al. 1995; Bilbao et al. 2004; Habib et al. 2015; Salem et al. 2015).

The purpose of the preprocedure evaluation is to determine presence of an appropriate indication, to exclude contraindications, and to assess the risks and benefits of TIPS creation. Preprocedure evaluation includes obtaining a complete clinical history, physical examination including evaluation for hepatic encephalopathy, and obtaining pertinent laboratory data. Specifically, a complete blood count, liver function, renal function, and coagulation status are assessed. Based on the clinical and laboratory data, prognostic scores such as the MELD and Child-Pugh scores are applied to assess risk (Ferral 2005; Copelan et al. 2014).

Contraindications should be excluded. Doppler ultrasound, contrast-enhanced computed tomography, and contrast-enhanced magnetic resonance imaging are helpful tools for determining patency of the portal and hepatic vasculature and for assessing presence of hepatic cysts or other hepatic lesions. Echocardiography and cardiac consultation are not necessities but are recommended in patients with known history of cardiac disease or pulmonary hypertension (Ferral 2005). If there is an emergent life-threatening indication for TIPS, time for imaging and cardiac evaluation may not be available.

After careful patient evaluation, the risks of the procedure are assessed against the severity of patient’s portal hypertension complication and the likelihood of survival until transplantation. The risks and benefits of TIPS creation should then be discussed with the multidisciplinary team, the patient, and the patient’s family (Boyer et al. 2010).

Acute complications related to the technical aspects of the procedure include biliary fistula formation, arterial injury, nontarget organ injury, hemoperitoneum, shunt malposition, infection, contrast-induced nephropathy, radiation dermatitis, and shunt dysfunction. Biliary fistula formation has an incidence < 5 % and can result in hemobilia, cholangitis, sepsis, stent infection, or stent occlusion. Biliary fistula can be treated with biliary diversion, arterial embolization, and/or realignment of the hepatic track with a covered stent. Arterial puncture during TIPS creation is rare with a risk of symptomatic arterial injury of < 2 %. Arterial puncture can result in hemorrhage, pseudoaneurysm formation, vascular occlusion, or arterial portal shunt creation. The risk of hemoperitoneum and the risk of injury to the gallbladder, right kidney, duodenum, and colonic hepatic flexure can be reduced by preprocedure planning and analysis of cross-sectional imaging. Careful technique during and after stent deployment decreases the risk of shunt malposition and migration. Periprocedural infection or sepsis is uncommon, but reported, therefore preprocedure antibiotic prophylaxis is recommended (Gaba et al. 2011).

TIPS creation may require long procedure and fluoroscopy times, therefore careful monitoring and documentation of patient dose should be made. The risk of major radiation injury related to fluoroscopic procedures is low and estimated to be between 1:10,000 and 1:100,000 procedures (Padovani et al. 2005). Sequelae of radiation skin injury can be seen with skin doses > 2 Gy. The Society of Interventional Radiology recommends patient follow-up if peak skin dose is > 3,000 mGy, reference point air kerma is > 5,000 mGy, Kerma-area product is > 500 Gycm², and fluoroscopy time is > 60 min. Dermatologic consultation is recommended if skin changes are noted on follow-up (Stecker et al. 2009; Gaba et al. 2011).

Early shunt dysfunction or thrombosis is often technical related to stent shortening or migration. In the earlier experience of TIPS placement when bare metal stents were utilized, biliary-TIPS fistulae were a common cause of shunt dysfunction. Currently, the use of expanded polytetrafluoroethylene (ePTFE)-covered stent grafts has decreased shunt dysfunction with a multicenter trial showing a 1-year stent patency rate of 84 % (Charon et al. 2004; Cura et al. 2008). Other causes of shunt dysfunction include pseudo-intimal hyperplasia at the hepatic venous end of the shunt, hypercoagulopathy, and poor inflow related to portal vein dissection or shunting of flow through varices and mesocaval shunts (Cura et al. 2008).

Subacute and chronic complications of TIPS are related to physiologic changes and include hepatic encephalopathy, liver ischemia, progressive liver failure, pulmonary hypertension, and congestive heart failure. Shunt dysfunction can also be a subacute or delayed complication. The reported rate of hepatic encephalopathy after TIPS creation is between 14 % and 25 %, with a 5–10 % incidence of refractory encephalopathy (Haskal et al. 2003; Charon et al. 2004; Pomier-Layrargues et al. 2012). Due to shunting of portal flow to the systemic system, hepatic perfusion after portosystemic shunt creation depends on arterial reserve. If the arterial reserve is low or there is arterial injury, hepatic infarction or progressive liver failure may ensue. A negative correlation has been postulated between hepatic arterial reserve and the Child-Pugh score (Gaba et al. 2011). Transplantation can be used as salvage therapy for liver failure after TIPS creation.

Hyperdynamic circulatory states are reported after portosystemic shunt creation; however, the rate of clinically significant pulmonary hypertension or worsening cardiac function is likely low and not readily reported in the literature (Azoulay et al. 1994; Van der Linden et al. 1996; Sawhney and Wall 1998). Nevertheless, careful patient selection is important to avoid cardiopulmonary complications. A variety of percutaneous shunt reduction and occlusion techniques have been described to treat these physiologic complications when they are refractory to medical therapy (Madoff and Wallace 2005).

Periprocedural mortality has been reported at < 2 % with 30-day mortality reported between 7 % and 45 %. Mortality is correlated with severity of underlying liver disease and comorbidities. This correlation emphasizes the importance of preprocedure patient assessment (Sawhney and Wall 1998).

Postprocedure care entails observing for and treating acute and subacute complications of TIPS creation including bleeding and encephalopathy. Hepatic function and renal function are also monitored. Transient hemolysis, hyperbilirubinemia, and transaminitis can be seen post-TIPS creation (Sanyal et al. 1996; Pomier-Layrargues et al. 2012). In order to decrease the risk of cardiopulmonary complications and aide the decompressive effect of the TIPS on the portal system, some experts advocate overnight diuresis of > 1 l if the postshunt mean right atrial pressure has increased to > 10 mmHg (Valji 2006; Kandarpa and Machan 2011).

Ultrasound surveillance protocols for evaluation of TIPS dysfunction were originally created in the era of bare metal stent placement. Typically sonography was performed at 24–72 h after TIPS creation then at 1 month, 3 months, and 6 months postplacement and every 6 months thereafter. Frequent surveillance and subsequent portal venography and intervention resulted in a primary assisted patency rate for bare metal stent TIPS of approximately 85 %. With the development and current use of ePTFE-covered stent grafts for TIPS, the primary patency rate is reported between 81 % and 84 %. Due to the superior patency rate of the ePTFE stent grafts, frequent ultrasound surveillance is no longer necessary. Carr et al. showed that in the ePTFE-covered stent graft era, only 4 % of ultrasound examinations effected clinical management, and 83 % of these examinations were the baseline evaluation. Due to air bubbles in the ePTFE material after deployment, baseline ultrasound is recommended after postprocedure day 5 rather than within the immediate 24–72 h period (Carr et al. 2006).

Long-term clinical evaluation of patients is recommended. If there are clinical signs of shunt malfunction, the patient should be referred to the interventional radiologist for portography, direct pressure measurements, and possible intervention. Depending on the findings of portography, balloon dilation, stent placement, varix and mesocaval shunt embolization as well as pharmacomechanical thrombolysis can be performed to improve TIPS patency.

Surgical resection is the ideal first line therapy for hepatocellular carcinoma. Surgical resection provides the best potential for tumor control. Other locoregional therapies do not remove the cancer and may not yield complete necrosis. Resection also allows pathological evaluation of the specimen to understand the biological aggressiveness of the tumor. Conversely, surgical resection is potentially more costly and may have increased periprocedural risks. The technical difficulty and potential postoperative complications of liver transplantation after surgical resection are also increased. Liver resection is limited to patients with well-compensated liver disease, minimal to no portal hypertension and tumor location amenable to resection (Pompili et al. 2013).

Using minimally invasive techniques, nonthermal technologies, such as percutaneous ethanol injection (PEI) and irreversible electroporation (IRE), as well as thermal technologies, such as radiofrequency ablation (RFA) and microwave ablation (MWA), are used to achieve targeted tumor cell death. Typically under ultrasound or computed tomography guidance, one or more applicators (needles, electrodes, or antennae) are advanced from the skin into or adjacent to the targeted hepatic tumor. Percutaneous ablation allows for targeted tumor cell death with intended margin necrosis and with relative sparing of the remainder of the nontumor liver parenchyma. Historically, percutaneous ablation techniques were used for the treatment of liver tumors ≤ 3 cm; however, new technologies and techniques have enabled the use of percutaneous ablation for larger tumors. According to the BCLC classification, radiofrequency ablation is recommended for patients with very early (BCLC stage A) or early (BCLC stage B) stage hepatocellular carcinoma. In practice, this recommendation is often expanded to also include other percutaneous ablation modalities and to include the treatment of liver-dominant metastatic disease in nonsurgical candidates meeting tumor size and number criteria. In some transplant centers, needle track seeding after percutaneous ablation is a concern. In 2001, Llovet et al. showed a needle track seeding incidence of 12.5 % in a group of 32 patients. Subsequent studies have shown much lower needle track seeding rates with an estimated incidence of 0–2.8 %. In a study of 1314 patients, Livaraghi et al. showed RFA of hepatocellular carcinoma resulted in a track seeding rate of 0.9 %. Increased risk of tumor seeding has been associated with subcapsular tumor location, poorly differentiated tumors, elevated serum alpha-fetoprotein levels, and prior biopsy (Llovet et al. 2001; Jaskolka et al. 2005; Livraghi et al. 2005).

Contraindications of percutaneous ablation of hepatic malignancy include tumor location near the main biliary ducts, intrahepatic biliary dilation, and uncorrectable coagulopathy. Due to increased risk of complications, care must be taken in the treatment of patients with bilioenteric anastomosis and treatment of lesions near vital structures such as the bowel, stomach, or gallbladder. Hydrodissection or gas dissection can be used to decrease deleterious effects to adjacent organs.