October 2001
P.C. Hayes, D.N. Redhead, N.D.C. Finlayson
Department of Gastroenterology
Centre for Liver and Digestive Disorders and
Department of Radiology
Royal Infirmary,
Edinburgh
SMJ 2001;46: 131-132
It is over 10 years since the first description of transjugular intrahepatic portosystemic stent shunts (TIPSS) to treat patients with portal hypertension. Since then there have been several suggested indications for the procedure some of which can be supported by published evidence. We review these indications in the light of current evidence.
TIPSS procedure
A TIPSS is essentially a side-to-side H-graft portosystemic shunt and therefore much information derived from surgical shunts is probably applicable to TIPSS. There are a number of differences, however, between these methods of making a portosystemic shunt.
First, and perhaps most important, a TIPSS can be inserted without a surgical operation and usually without a general anaesthetic. This has allowed the technique to be used on patients with very advanced liver disease in whom shunt surgery would not be considered because of operative mortality. Second, the diameter of the TIPSS shunt is 8-10 mm which is narrower than most surgical shunts. A TIPSS should therefore precipitate less encephalopathy but be more liable to thrombosis and be less effective in normalising the portal pressure.
Acute variceal haemorrhage
Many patients with acute variceal haemorrhage will have bleeding controlled and further bleeding prevented by endoscopic band ligation or injection sclerotherapy. Ten to 20 percent of patients will, however, continue to bleed despite apparently adequate endoscopic treatment and TIPSS should be considered for these patients.
Some have suggested that a second attempt at endoscopic treatment should be made before considering TIPSS, but in practical terms this is not always appropriate If, for example, a patient has had a significant lifethreatening haemorrhage and in spite of endoscopic therapy continues to bleed, there must be a strong case for placing a TIPSS, if necessary with a Sengstaken tube in place to control bleeding rather than risking further massive bleeding.
The niceties of exactly when a TIPSS should be used are unlikely to be examined in controlled clinical trials and as with many things in medicine local expertise and experience are likely to be important in decision-making.
A TIPSS can be placed in 95% of cases of acute variceal bleeding and is almost always successful in controlling acute bleeding.1 The effectiveness of TIPSS in comparison with other methods of therapy for recurrent bleeding is based more on clinical experience than randomised controlled trials. In a retrospective analysis of TIPSS compared with oesophageal transection in uncontrolled variceal bleeding, TIPSS was much superior in stopping variceal bleeding and lowering mortality, probably because of the large number of such patients who are unfit or unsuitable for surgery.2 Two other studies support this observation.3,4
Patients with massive variceal bleeding who have poor liver function, renal failure, and/or encephalopathy5,6 have a very poor prognosis and it might be argued that failure of endoscopic therapy in such patients should not be followed by further treatment. However, this is a very difficult decision to take, particularly when TIPSS is available and when patients may not be well known to the doctors treating them.
Prevention of rebleeding
Once control of acute variceal haemorrhage has been achieved using endoscopic treatment, rebleeding occurs in most patients within 2-3 months unless further treatment is given. Endoscopic sclerotherapy was the standard management until recently, but because of reduced side-effects and better efficacy endoscopic band ligation is now the treatment of choice.
TIPSS has been compared with endoscopic treatment in eight randomised controlled trials, seven of which have compared TIPSS with endoscopic sclerotherapy7 and only one TIPSS and band ligation.8 The results of these studies all show that bleeding rates are lower with TIPSS by approximately 25%. There was no evidence that TIPSS could reduce mortality, but these were studies designed with rebleeding as the primary end-point and there was the usual problem of insufficient patients to demonstrate mortality differences reliably. Furthermore, in nearly all studies TIPSS was used to rescue patients who continued to bleed after endoscopic therapy.
The number of patients needing TIPSS rescue varied from 12-30%. TIPSS rescue may well have affected survival, and in one study where TIPSS rescue was not used (only in one patient), improvement both in rebleeding and survival was seen. In reality, therefore, these studies compared TIPSS with endoscopy supplemented by TIPSS rescue and consequently it is not possible to say whether or not TIPSS itself has any effect on survival.
TIPSS and beta-blockers
The other treatment with which TIPSS should be compared is beta-blockade therapy. Meta-analyses have shown that beta blockers are effective in reducing rebleeding and death compared with no treatment and are comparable to sclerotherapy for survival but less effective than sclerotherapy for rebleeding. If survival were to be the only end-point of interest, a comparison of TIPSS and beta-blockade would be extremely valuable. However, this would again raise the problems of TIPSS rescue should serious variceal bleeding occur in patients on beta-blockers.
Gastric varices
TIPSS for treating gastric varices is less contentious because endoscopic treatment is generally considered ineffective in controlling acute gastric variceal haemorrhage. TIPSS has been shown to be as effective in controlling gastric variceal haemorrhage as it is for oesophageal variceal haemorrhage.9,10 Gastric varices are relatively rare and large clinical trials would be difficult to set up, and consequently many would support the view that TIPS S should be used early in the management of acute fundal gastric variceal haemorrhage.
A comparison of TIPSS with endoscopic intravariceal injection of butylcyanoacrylic or thrombin is not available but would be of interest.
Other varices
Bleeding occurs occasionally from varices in other sites in the gastrointestinal tract, particularly the rectum, and from stomal varices. TIPSS should be used in this situation although trials of therapy are not available.
Primary prophylactic therapy
There have been no studies on the effectiveness of TIPSS in the primary prevention of variceal haemorrhage. Because of the invasiveness of the TIPSS technique, the relatively high incidence of shunt occlusion and the fact that prophylactic shunt surgery has been shown to increase mortality,11 there is little enthusiasm for TIPSS in primary prophylaxis. Furthermore, propranolol and more recently band ligation, have proved reasonably effective alternatives for this purpose.
Ascites
TIPSS, as well as reducing portal pressure, has beneficial effects on renal perfusion, renal plasma flow, sodium and water excretion by the kidneys and plasma neurohumoral concentrations.12 For these reasons, ascites would be expected to respond well to TIPSS.
There have been several observational series reporting response rates of 40-74% in the treatment of refractory ascites with TIPSS.13,14,15 However, the most important publications have been two randomised control trials which have given frustratingly different conclusions. The first showed that TIPSS was effective in controlling ascites, but was associated with a higher mortality in patients with poor liver function (Child’s C).16
Enthusiasm for TIPSS for the treatment of refractory ascites then waned, until a second trial reported that in comparison with large volume paracentesis TIPSS improved survival without liver transplantation in patients with refractory or recurrent ascites.17The difference between these two trials probably reflects patient selection.
Further clarification of which patients will benefit from TIPSS is required, probably in the form of further randomised trials. The question of quality of life is also something that should be factored into the equation as well as just survival.
Other indications
There have been other situations in which a TIPSS has been reported to be of some benefit. These include hepatic hydrothorax, hepatorenal syndrome and the Budd Chiari syndrome. As these conditions are relatively unusual, there are few case series and no randomised trials. However, experience in these areas is likely to grow over the next few years.
Shunt occlusion
One of the major problems with TIPSS currently is the high rate of shunt narrowing and eventual thrombosis unless patency is maintained by regular active intervention and angioplasty. This is invasive and expensive and shunt dysfunction can occur despite surveillance. The recent introduction of PTFE covered stents which may reduce shunt thrombosis is an exciting development, and preliminary reports suggest that dysfunction with these covered stents will be less of a problem.18
Resource allocation
Although the indications for TIPSS are not yet settled, we believe enough is known to suggest that about 100-300 (2 - 6 per 100,000) are needed annually in Scotland. Most of these will be needed for variceal bleeding, and as transferring such patients over large distances is hazardous, the capacity to place TIPSS should be developed in all the major health boards. Such a service should, at least, be available seven days a week and probably around the clock to be fully effective as variceal bleeding would be the main indication.
Conclusions
The last decade has seen TIPSS established as a useful treatment for variceal bleeding. In acute bleeding, it is particularly useful for patients with oesophageal varices who rebleed after endoscopic therapy, and it is the treatment of first choice for those who bleed from varices in other sites such as the stomach. TIPSS is also a valuable secondary prophylaxis against variceal rebleeding where it may be the treatment of first choice or be used where variceal band ligation fails.
Shunt occlusion is common following TIPSS although its occurrence can be detected and treated by regular angiographic surveillance. Recently introduced covered TIPSS may reduce this problem in future.
TIPSS has no place in primary prophylaxis of variceal bleeding. TIPSS may also be useful in recurrent or resistant ascites, particularly in patients with relatively good liver function, but such patients should always first be considered for liver transplantation.