Introduction
Acute variceal bleeding (AVB) is a common and lethal complication in cirrhosis patients, with short-term mortality ranging between 20% and 30%.1 The goal of management in AVB is to control bleeding, prevent rebleeding and bleeding-related mortality from infection and acute-on-chronic liver failure. Considering bleeding-related mortality as an important complication of AVB, the Baveno-VII consensus recommended 6 week mortality as the primary endpoint in studies evaluating treatment for AVB.2 Current guidelines recommend vasoactive drugs, prophylactic antibiotics, early endoscopy (within 12 hour), and restrictive blood transfusion to improve survival in AVB. Moreover, variceal eradication and beta-blocker combination therapy have been recommended as secondary prophylaxis to prevent rebleeding in these patients.3
Transjugular intrahepatic portosystemic shunt (TIPSS), the creation of a side-to-side shunt between the hepatic vein and portal vein, was introduced in the 1960s as a salvage therapy for AVB refractory to standard treatment.4 TIPSS was first performed in humans using expandable metallic stents in the 1980s.5 On one hand, by rapidly reducing the portal pressure, TIPSS effectively prevents rebleeding in AVB and other portal hypertension-related complications, including refractory ascites and hepatic hydrothorax.3 On the other hand, the diversion of portal blood flow may precipitate hepatic encephalopathy and liver decompensation. While the uncovered metallic stent used in the early days of TIPSS was associated with an increased risk of stent dysfunction and early thrombosis, this dreadful complication was mitigated by the use of expanded polytetrafluoroethylene (e-PTFE) stents to reduce TIPSS dysfunction.6
While TIPSS was originally designed as salvage therapy in AVB, the concept of using TIPSS as secondary prophylaxis for AVB was tested in a randomized trial by Jalan et al.7 in 1997. In that randomized controlled trial (RCT), pre-emptive TIPSS (p-TIPSS) insertion following endoscopic hemostasis was found to significantly reduce rebleeding risk in AVB patients, even though the overall survival was similar to that of patients who had received only endoscopic hemostasis.7 That trial was followed by a study by Monescillo et al.,8 which was the first to demonstrate the survival benefit of p-TIPSS when inserted in high-risk AVB patients with a hepatic venous pressure gradient (HVPG) ≥20 mmHg However, they used uncovered metal stents and sclerotherapy for variceal treatment, which is no longer considered standard of care. Moreover, the use of HVPG to risk-stratify patients for p-TIPSS was also limited by the availability of expertise to perform HVPG outside of specialized centers.
Pre-emptive TIPSS: current status
A landmark study by Garcia-Pagan et al.9 was pivotal in redefining the role of p-TIPSS in the management of AVB. Using simple clinical scores to identify Child-Turcott-Pugh (CTP) class C or CP-B patients with active bleeding during endoscopy and with a high risk of rebleeding, early placement of TIPSS within 72 hour successfully prevented rebleeding and decreased mortality in these patients.2 The benefits of p-TIPSS go beyond preventing rebleeding to preventing mortality and de-novo ascites, without significantly increasing the risk of hepatic encephalopathy.3 The benefits of early p-TIPSS were also described in an RCT by Lv et al.,10 independent of active bleeding during endoscopy. In contrast, the survival benefit of p-TIPSS was not observed in an RCT by Dunne et al.,11 yet a higher risk of hepatic encephalopathy following p-TIPSS was observed. Unfortunately, the survival benefit was inconclusive because this study was underpowered because of slow recruitment. To reconcile the controversies on the survival benefit of p-TIPSS, a recent meta-analysis and trial sequential analysis showed that the evidence from current RCTs is insufficient to support a 6 week survival benefit with p-TIPSS compared with the standard of care.12 A summary of all RCTs evaluating p-TIPSS is shown in Table 1.7–11,13
Table 1Summary of randomized trials evaluating pre-emptive TIPSS in the setting of acute variceal bleeding
| Author | Patient characteristics
| No. of patients (TIPSS vs. control) | Type of stent | Standard of care | Main findings | HE (%; TIPSS vs. control) |
|---|
| Definition of high risk | Predominant etiology |
|---|
| Jalan et al. 19977 | Not defined | ETOH | 31 vs. 27 | Expandable uncovered stent | EVL repeated weekly until eradication, and then at 3 and 6 months, and 6-monthly intervals thereafter | Variceal rebleeding frequency and severity reduced with TIPSS (p<0.0006). No significant difference in mortality rates or frequency of HE | 36% vs. 33% |
| Pomier-Layrargues et al. 200113 | CTP 7–12 | ETOH, cryptogenic | 41 vs. 39 | Not defined | EVL performed on days 1 and 10, then every 3–4 weeks until obliteration. 3-monthly surveillance thereafter | Variceal rebleeding significantly lower in TIPSS group at 2 years (p<0.001) but no difference in survival rate. No difference in probability of HE | 47% vs. 44% |
| Monescillo et al. 20048 | HVPG ≥ 20mmHg within 24hrs of AVB | ETOH, HCV | 26 vs. 26 | Uncovered stent | Sclerotherapy then NSBB, or EVL if NSBB contraindicated or not tolerated | No difference in 6-week mortality. Reduction in in-hospital and 1-year mortality with TIPSS (p=0.02, p=0.01 respectively). No increase in de no HE with TIPSS | 31% vs. 35% |
| García-Pagán et al. 20109 | CTP C≤13 or CTP B with AVB | ETOH, HCV | 31 vs. 32 | e-PTFE-covered stent | Optimization of NSBB and ISMN. Second EVL within 7–14 days then every 10–14 days until eradication | 6-week and 1-year survival higher in early TIPSS (NNT=3.3, NNT=4.0 respectively). Reduction in 1-year rebleeding (p<0.001). No difference in HE (p=0.13) | 25% vs. 39% |
| Lv et al. 201910 | CTP B or C<14 | HBV, HCV, ETOH | 86 vs. 46 | e-PTFE-covered stent | Optimization of NSBB. Second EVL within 7–14 days, then every 14 days until eradication | Transplant-free survival higher in TIPSS at 6-weeks and 1 year (p=0.02, p=0.046). No difference in HE (p=1.00) | 35% vs. 36% |
| Dunne et al. 202011 | CTP B and C≤13 | ETOH | 29 vs. 29 | e-PTFE-covered stent | Optimization of NSBB. Endoscopy at 2- to 4-week intervals until variceal eradication | No difference in 6-week or 1-year survival with early TIPSS. Trend toward reduced rebleeding with TIPSS (p=0.09). HE more common with TIPSS (p=0.04) | 41% vs. 17% |
Pre-emptive TIPSS: ongoing controversies
Ongoing controversies in defining the optimal role of p-TIPSS include patient selection, particularly in NASH cirrhosis patients with associated cardiovascular risk factors and in patients with acute-on-chronic liver failure, and the role of sarcopenia and renal impairment in the setting of p-TIPSS.
Patient selection
There are conflicting views on whether patients with Child-Pugh-Turcotte (CPT) class B disease and active bleeding or with MELD scores between 12 and 18 would benefit from pre-emptive TIPSS.14,15 Whether p-TIPSS remains beneficial when performed after 72 hour is debatable, because real-world evidence suggests that benefits may be observed in patients who underwent TIPSS16 after 72 hour of AVB. While active bleeding was defined as the presence of active bleeding during the insertion of endoscope, the reliability of the finding is also subject to the time that endoscopy was performed. The finding of active bleeding was associated with an increased risk of death among CTP class B patients in some studies,15 but not in others.10,17 Patient selection for pre-emptive TIPSS using the CTP score may be limited by subjective variables within the CTP score as it does not distinguish between the individual phenotype of liver dysfunction (liver synthetic dysfunction versus portal hypertension-related).18 Moreover, data on the efficacy of pre-emptive TIPSS in preventing rebleeding and death with gastric variceal bleeding is also scarce. A multicenter trial (GAVAPROSEC) is currently underway to compare the benefit of pre-emptive TIPSS and glue obliteration in preventing rebleeding and death in bleeding gastric varices. Lastly, the survival benefit of p-TIPSS was recently questioned,12 because the standard of care (carvedilol, variceal band ligation, and early access to endoscopy) has improved since the landmark trial conducted a decade ago.7–9 A large multicenter randomized trial (REACT-AVB) comparing p-TIPSS with the standard of care in patients with Child-Turcotte-Pugh scores 7–13 is underway in the United Kingdom.19 The trial will address fundamental questions on the survival benefits and the ideal target population for p-TIPSS in cirrhosis patients with AVB.
Pre-emptive TIPSS in NASH-related cirrhosis
Given the rising obesity pandemic, NASH cirrhosis will likely emerge as the driving cause of cirrhosis.20 However, the impact of p-TIPSS on NASH cirrhosis is not clear, as NASH-related cirrhosis has been under-represented in all the existing trials, in which the primary etiologies were alcoholic cirrhosis and chronic hepatitis B.7–11 There are several considerations when selecting NASH cirrhosis patients with AVB for p-TIPSS. First, cardiovascular complications, a key exclusion criterion for TIPSS, are prevalent in NASH cirrhosis.21 Cardiac evaluation is paramount because p-TIPSS may potentially unmask undiagnosed cardiovascular disease. TIPSS insertion shunts a significant volume of blood from the splanchnic into the systemic circulation, with consequential increases in cardiac output and right heart pressure. While the sudden rise in right heart pressure following TIPSS is usually transient, the development of cardiac decompensation following TIPSS can be detrimental.22,23 Patients with severe left ventricular dysfunction,24 severe aortic stenosis, or severe pulmonary hypertension should not proceed with p-TIPSS. Current guidelines recommend a 12-lead electrocardiogram and N-Terminal pro-B-type natriuretic peptide before TIPSS insertion.3,25,26 The fact that NASH cirrhosis patients experience portal hypertension-related at a lower HVPG may affect patient selection for p-TIPSS using HVPG.27 That is further confounded by the inter-observer variability of HVPG, particularly in cases with decompensated NASH cirrhosis.28 In summary, TIPSS can be associated with an increased risk of HE and renal dysfunction in NASH-related cirrhosis.29 Therefore insertion of p-TIPSS in NASH cirrhosis patients must consider co-existing cardiac and renal comorbidities to minimize potential TIPSS-related complications in those patients.
Acute-on-chronic liver failure
Acute-on-chronic liver failure (ACLF) is a clinical syndrome associated with multiorgan failure and high short-term mortality. Because the hemodynamic changes following TIPSS insertion may precipitate cardiac and liver failure, application of p-TIPSS in the setting of ACLF must be supported by strong evidence. In a multicenter prospective cohort study of 2,138 patients with AVB, the presence of ACLF retrospectively defined following European Association for the Study of the Liver–Chronic Liver Failure (EASL-CLIF) consortium criteria predicted mortality in patients with AVB.30 Among 380 patients (17.8%) who had ACLF, p-TIPSS was associated with a decreased risk of 6-week rebleeding (hazard ratio: 0.128; 95% confidence interval: 0.017–0.937; p=0.043), 42-day mortality (13.6% vs. 51.0%, p=0.002 and 1-year mortality (22.7% vs. 56.5%, p=0.002). However, before confirming the survival benefit of p-TIPSS in the setting of ACLF, there are several key considerations. First, the benefit of p-TIPSS was derived from only 22 ACLF patients (5.6% of the overall ACLF cohort), which raises concerns of the generalizability of the findings. Second, observational studies are subject to selection and indication bias. Subjects who underwent p-TIPSS had lower baseline MELD scores and less severe ACLF grade than those who did not undergo p-TIPSS. Third, there are differences in the definition, clinical phenotype, and primary etiology of ACLF between different ACLF consortia, which complicates patient selection for p-TIPSS in this setting.31 Between ACLF subjects with or without p-TIPSS, it is important to know the proportions of ACLF patients with intrinsic liver failure and with ongoing sepsis that may deter consideration of p-TIPSS. While p-TIPSS can reduce rebleeding, is it sufficient to change the trajectory of ACLF that results in a survival benefit in ACLF? Given the relative paucity of data, randomized trials are required to confirm the survival benefits on p-TIPSS in the setting of ACLF.
Sarcopenia
Sarcopenia is present in 40% of patients with decompensated cirrhosis and has been associated with an increased risk of hepatic encephalopathy and death in cirrhosis patients.32 Sarcopenia has been associated with an increased risk of HE following TIPSS,33 this association was not consistently demonstrated.34 On the other hand, TIPSS was associated with an improvement in sarcopenia.35 Retrospective analysis of 27 patients showed that the skeletal muscle index improved 6 months following TIPSS insertion. Current guidelines recommend sarcopenia assessment before TIPSS insertion,25 but sarcopenia should not be considered as contraindication for p-TIPSS. Further studies are required to understand the role of sarcopenia in patient selection for p-TIPSS.
Renal impairment
Most studies on p-TIPSS excluded patients with severe renal impairment7–11 because TIPSS was associated with a higher risk of hepatic encephalopathy36 and a lower natriuretic effect in these patients.37 The presence of acute kidney injury in the setting of AVB is generally not considered a contraindication for p-TIPSS. Among advanced cirrhosis patients with hepatorenal syndrome, TIPSS placement was associated with improvement in renal function, but a significantly higher risk of HE following TIPSS.38 Moreover, a recent meta-analysis highlighted significant heterogeneity among the included studies and the lack of high-quality studies evaluating TIPSS in patients with hepatorenal syndrome. Due to the increased risk of HE, p-TIPSS in the setting of AVB with hepatorenal syndrome or severe chronic renal impairment is currently considered experimental.
Post-TIPSS hepatic encephalopathy
HE is generally considered a relative contraindication in the setting of rescue TIPSS, as the risk of HE is outweighed by the survival benefit of TIPSS as a life-saving procedure. However, the concern of HE is reasonable in stable patients scheduled for elective p-TIPSS. Overall, the prevalence of post-TIPSS HE in RCTs evaluating p-TIPSS ranges from 25% to 47% (Table 1). Although a meta-analysis of randomized trials suggested a similar rate of HE following p-TIPSS,12 that information should be interpreted with caution, given that the studies generally excluded patients with recurrent HE or severe renal impairment, yet sarcopenia and covert HE were not systematically assessed in the p-TIPSS setting.
Pre-emptive TIPSS: the way forward
Despite the benefit to prevent rebleeding, p-TIPSS has not been widely adopted because of important logistic challenges, even in expert centers. The feasibility of pre-emptive TIPSS must be considered before adoption because p-TIPSS is a highly specialized and resource-intensive procedure with limited access in most centers in the real-world setting. The challenge in delivering pre-emptive TIPSS within 72 h was reflected in a nationwide French audit including 964 patients with AVB from 58 centers. Even though 35% of the patients fulfilled the criteria for p-TIPSS, only 6.8% underwent p-TIPSS.39 In a recent randomized trial by Dunne et al.,11 55% of patients in the p-TIPSS group either had TIPSS performed after 72 h or did not undergo TIPSS at all. The implementation of p-TIPSS also involves transferring critically-ill patients from low-volume centers, which may potentially overwhelm high-volume referral centers. That being said, rather than not performing p-TIPSS, the considerations should guide us in personalizing p-TIPSS in cirrhosis patients with AVB. How can we, then, optimize the outcomes of patients undergoing p-TIPSS? Numerous strategies have been proposed to mitigate the potential complications with p-TIPSS. First, HE is a common occurrence in 35–50% of patients following TIPSS, and is driven by increased ammonia load, infection, inflammation, or liver failure. One should recognize the risk factors of post-TIPSS HE such as advanced age, higher CPT scores, history of overt HE, and sarcopenia. Rifaximin may be considered as a primary prophylaxis for HE prevention before and after p-TIPSS. In a randomized trial including 197 patients undergoing TIPSS, rifaximin before TIPSS was associated with a decreased risk of overt HE (OR: 0.48) compared with placebo.40 That is in contrast with an earlier trial by Riggio et al.41 which showed that rifaximin did not reduce post-TIPSS HE The urgency to perform p-TIPSS within 72 hour of AVB may influenced the adoption of rifaximin as primary prophylaxis for HE following p-TIPSS. Currently, a European multicenter RCT (the PEARL trial) of combined treatment with rifaximin and lactulose for HE prophylaxis is currently ongoing. The results are highly anticipated.42
The existence and diameter of the portosystemic shunt (SPSS) is an independent predictor of post-TIPSS hepatic encephalopathy.43 A recent randomized trial by Lv et al.44 reported a lower risk of HE following TIPSS with concurrent embolization of a large spontaneous portosystemic shunt in patients undergoing TIPSS. Following embolization of large collaterals, the amount of blood shunting through TIPSS was reduced, as it was limited by the diameter of the TIPSS, which is often smaller than the co-existing large SPSS. The benefit of embolization of SPSS during TIPSS was demonstrated in an observational study by Leng et al. in which the risk of HE was similar in patients who underwent SPSS embolization and in those without SPSS.45 It makes sense that the risk of HE was decreased by creating a new, smaller portosystemic shunt, further randomized trials are needed to confirm safety from the perspective of variceal bleeding and ascites management. Meanwhile, a fully covered, small-diameter controlled-expansion stent can be used to use to minimize the risk of HE following p-TIPSS.46 Meanwhile, a smaller 8 mm covered stent can be considered as it has been shown equally effective in preventing rebleeding while reducing the risk of post-TIPSS HE by 47% compared with a 10 mm stent.47 Special attention should be paid to patients with sarcopenia, diabetes mellitus, or renal impairment, in whom the risk of HE is inherently greater.34,46–49 Patients should maintain sufficient caloric intake of 35–40 kcal/kg/body weight/day and a protein intake of 1.2–1.5 g/kg/body weight per day. Prolonged fasting should be avoided whenever possible. Finally, to reduce the risk of renal impairment, nephrotoxic drugs should be stopped before p-TIPSS, and the amount of iodinated contrast used during TIPSS should be minimized with the help of endovascular ultrasound and carbon dioxide venography.50
Conclusions
In summary, p-TIPSS is an important tool to reduce rebleeding in cirrhosis patients with AVB. While the benefits of reducing rebleeding and de-novo ascites are evident, controversies remain concerning patient selection, particularly among those with NASH cirrhosis and acute-on-chronic liver failure. Hepatologists must be familiar with the strength and limitations of p-TIPSS and be aware of the strategies to optimize outcomes in patients undergoing p-TIPSS. Future work should focus on improving access to p-TIPSS and individualizing p-TIPSS in the setting of AVB.
Abbreviations
- AVB:
acute variceal bleeding
- CTP:
Child-Turcotte-Pugh
- CSHP:
clinically significant portal hypertension
- HE:
hepatic encephalopathy
- HVPG:
hepatic venous pressure gradient
- NASH:
nonalcoholic steatohepatitis
- p-TIPSS:
pre-emptive transjugular intrahepatic portosystemic shunt
Declarations
Funding
YJW is supported by Nurturing Clinician Scientist Scheme, Duke-NUS Academic Medical Program.
Conflict of interest
YJW has been an editorial board member of Journal of Clinical and Translational Hepatology since 2020. The authors have no conflict of interests related to this publication.
Authors’ contributions
Study concept (YJW, JGA), drafting the manuscript (YJW, WLDH), critical review of the final manuscript (YJW, JGA).