Gastro-oesophageal varices (swollen veins in the gullet and stomach) are present in 50% of patients with liver cirrhosis and are its most serious complication as their rupture results in potentially life threatening bleeding. Bleeding from these veins occurs in up to one third of patients with varices. This is associated with 20% mortality at 6 weeks.

In the event of bleeding from these veins the current UK guidelines recommend certain drugs followed by early endoscopic treatment with variceal band ligation (rubber bands placed over the veins to stop them bleeding). The use of a shunt inside the liver ("TIPSS" transjugular intrahepatic portosystemic shunt) is largely reserved for cases of uncontrolled bleeding from these swollen veins. A recent randomised multicenter study carried out by Garcia Pagan and colleagues reported improved survival with early TIPSS in patients with bleeding from these swollen veins in advanced liver disease. From these guideline international guidelines now recommend consideration of early TIPSS for all high-risk patients presenting with variceal bleeding. This practice clearly has significant cost implications. To validate the findings a further randomised control trial is needed.

This is a multi-center parallel-group randomized controlled trial. Both hospitals taking part in the trial will have a TIPSS service. Patients who consent to enter the trial will be randomized to either: (1) Endoscopic treatment (standard care) or (2) early TIPSS.

Potential participants will be all patients with a diagnosis of liver cirrhosis presenting with an acute variceal bleed to a participating hospital who do not fulfill an exclusion criteria. All causes of cirrhosis will be included.

Participants will be reviewed during their regular hepatology clinic appointments at their respective hospitals on 3 occasions over a one-year period.

Background and Rationale for the Study

Gastro-oesophageal varices are present in 50% of patients with cirrhosis [2] and are the most serious complication of portal hypertension as their rupture results in potentially life threatening variceal haemorrhage with overall mortality rates historically reported as 30-50% [3]. Although mortality can be up to 40% at 6 weeks, it can be up to 70% at one year [2]. With the generally improved management of critically ill cirrhotic patients, together with vasoactive therapy and new endoscopic techniques for managing variceal haemorrhage, overall mortality has reduced, with one centre in Europe showing a reduction from 42% in 1980 to 14% in 2000 [4]. 60-80% of patients who bleed from varices will re-bleed if not treated [5-8], and the risk of re-bleeding is greatest in the first 10 days [1,9], during which 50% of those who are going to re-bleed, do so. The risk of re-bleeding gradually falls over the first month when an additional 10% re-bleed [7, 8]; the risk after the first six weeks then plateaus out.

Despite the advent of endoscopic therapies and early pharmacological therapies, re-bleeding rates are still higher early on, with factors predictive of early re-bleeding /treatment failure at 5 days including: active bleeding at index endoscopy; severity of liver disease (Child-Pugh class); severity of bleed; and severity of portal hypertension [1, 10]. Hepatic Venous Pressure Gradient (HVPG) is one of the best predictors of identifying those who will re-bleed. After an index variceal bleed, a reduction of HVPG to less than 12mmHg or by at least 20%, reduces the risk of re-bleeding from 46-65% to 0.13% [11]. HVPG measurement is usually limited to specialist centres.

The early use of TIPSS has been explored in two studies [1,12]. The first [12] categorised patients presenting with a variceal bleed to be either high or low risk of re-bleeding determined by their HVPG. Those with an HVPG greater or equal to 20mmHg went on to receive an early TIPSS and had an improved outcome compared with those treated medically. However, the results must be interpreted with caution, as the medical management was not current standard of care.

A further study investigating the use of early TIPSS for variceal haemorrhage was carried out by Garcia-Pagan and colleagues [1]. This landmark study selected patients for early TIPSS (within 3 days) determined by Child-Pugh score. Participants, who had Child's B or C cirrhosis and on-going bleeding, progressed to early use of TIPSS (with an e-PFTE-covered stent). The results of this trial demonstrated a significant reduction in the failure to control bleeding and re-bleeding with no increase in the risk of hepatic encephalopathy. The study also demonstrated a significant survival benefit with early TIPSS at all time points. There were, however, several features of this study that raised concern.

Firstly recruitment was prolonged (3 years) to recruit 63 patients via 9 centres, with a high exclusion rate (296 patients excluded). The second concern featured the inclusion of patients with ongoing bleeding following index endoscopy. This might arguably be termed a rescue TIPSS and although no studies have been done in this area it is intuitive to suggest that survival would be improved if haemostasis has not been achieved. Thirdly, survival at 1 year with early TIPSS was extremely high (86% vs. 61% in the medical management group). For patients with Child's C cirrhosis presenting with variceal bleeding this survival rate is remarkable.

These two studies [1, 12] of the early use of TIPSS for variceal haemorrhage are the first in portal hypertension to show a mortality benefit. As such these demand action. Either the management of variceal bleeding in patients with advanced cirrhosis must change or, given the concerns regarding the study design and the significant cost implications, these results need to be further validated.

STUDY DESIGN This is a multi-centre, open-label, parallel-group, randomised control trial. Both hospitals taking part in the trial will have a 24-hour TIPSS service. Potential participants will be all patients with a diagnosis of liver cirrhosis presenting with variceal haemorrhage to a participating hospital who do not fulfil an exclusion criteria. Potential participants will be identified by the gastroenterology team responsible for the patient care. Patients and their relatives will be approached by a nominated registrar, or consultant, identified in the ethics application or delegation log. Patients will be followed up for one year and will be reviewed on three occasions. Patients who consent to enter the trial will be randomized to either: (1) EVBL ("standard care") arm or (2) early TIPSS arm.

FOLLOW UP During their regular Hepatology follow-ups, patients will be reviewed at 6 weeks, 6 months, and 12 months, which is the end of the study. TIPSS patency will be checked at 6 months and 1 year as per current standard protocol. All patients will be followed up until death or the end of the study, whichever is first.

IDENTIFYING PARTICIPANTS Potential participants will be all patients with a diagnosis of liver cirrhosis presenting with an acute variceal haemorrhage to a participating hospital who do not fulfil an exclusion criteria. All aetiologies of cirrhosis (including cryptogenic and where the aetiology is yet to be established) will be included. Potential participants will be identified by the gastroenterology team responsible for the patient. The consultant responsible for the care of each patient will determine whether the identified patients are suitable and may be approached.

OBTAINING CONSENT If patients are alert and able to give informed consent then they will be issued with a patient information sheet and consent sought from the patient. Patients and their next of kin will have sufficient time usually 3 hours or more, from the index endoscopy to consent to inclusion in the study.

Patients with liver disease may have fluctuating conscious level and may lack capacity due to hepatic encephalopathy. Potential participants who lack capacity will be included in the study. Here capacity will be assessed by the consenting clinician at the time of taking consent. A further consent will be sought from those patients who regain capacity.

RANDOMISATION This study will involve simple 1:1 randomisation. Randomisation will be done using a web-based randomisation programme that will be available 24 hours a day to both study centres.

STUDY ASSESSMENT Participants will attend the gastroenterology/hepatology clinics at their respective hospital on three occasions over a one-year period. If participants are in hospital at the time of a scheduled visit then they will either be assessed whilst inpatient or at the programmed visit as an outpatient. The programme visits are: 6 Weeks, 6 Months and 1 year.

SAMPLE SIZE CALCULATION Given that we are validating results of the Garcia Pagan study, we have powered the study using the results that they observed. We wish to find a difference in survival between the two trial arms. Garcia Pagan observed 14% and 39% deaths in the two trial arms. We calculate that we would need 48 patients per group. This is from a 2-sided log-rank test, with alpha=0.05, and a power of 80%, allowing 4 extra patients per group to allow for drop out and non-compliance.

Inclusion Criteria:

• An acute oesophageal variceal bleed with haemostasis following initial endoscopy.
• A diagnosis of liver cirrhosis
• Childs-Pugh score ≥8

Exclusion Criteria:

• Inability to control bleeding at index endoscopy (this is a "rescue TIPSS")
• Previous portosystemic shunt or TIPSS
• Bleeding from isolated gastric or ectopic varices
• Known portal vein thrombosis precluding TIPSS
• Active cancer including hepatocellular carcinoma
• Age less than 18 or more than 75
• Clinically significant encephalopathy causing recurrent hospital admissions
• Pregnant at time of index endoscopy