3 Evidence

NICE commissioned an external assessment centre (EAC) to review the evidence submitted by the company. This section summarises that review. See the committee papers for full details of the evidence.

Clinical evidence

The main clinical evidence comprises 9 studies

3.1 The evidence assessed by the EAC included 9 full-text peer-reviewed studies including 247 people. Two of the studies were comparative: a randomised controlled trial and a retrospective case‑controlled study. The remaining 7 studies were non-comparative case series. For full details of the clinical evidence, see section 3 of the assessment report in supporting documentation.

The comparative evidence is relevant to the decision problem but has limitations

3.2 Both comparative studies compare the use of Danis stent with balloon tamponade. The studies report that using Danis stent improves control of bleeding at 5 days and 15 days. The randomised controlled trial (Escorsell et al. 2016) is the strongest evidence for Danis stent and reports a composite end point including control of bleeding and adverse events but it is underpowered for this result. The retrospective case-controlled study only included patients with acute-on-chronic liver failure and there were significant differences between the disease-severity scores of the patients in the control group compared with the interventional group. The EAC reported that both studies have a moderate risk of bias.

The randomised controlled trial (Escorsell et al. 2016) is not reflective of the UK care pathway

3.3 The randomised controlled trial was done in Spain and differences in the care pathway limit the generalisability of the findings to the UK setting. Expert advisers stated that the definitive procedure, transjugular intrahepatic portosystemic shunt (TIPS), was delivered at an earlier stage after presentation in this trial than it would be in the UK.

The EAC did a meta-analysis on the 7 non-comparative studies

3.4 The 7 non-comparative studies are low in quality. The studies have broadly similar populations and outcomes. Outcomes with low heterogeneity were included in the analysis. Immediate bleeding control was found to have been achieved in 88% of cases (95% confidence interval 0.38 to 0.9) based on the 7 case series, one of which (Wright et al. 2010) was done in the UK. Survival rate at 30 days was 68% from 3 studies.

Cost evidence

The company's comparison of the cost of Danis stent and balloon tamponade uses a cost-calculator model

3.5 The cost comparison has a 6‑week time horizon and is from an NHS and personal and social services perspective. The model estimates the cost associated with using Danis stent compared with balloon tamponade as bridging treatment for patients aged 16 or over with acute refractory oesophageal variceal bleeding in whom first-line therapy is unsuitable or has failed. The model captures the cost of the initial procedures, the likelihood of adverse events for both technologies and the cost and use of resources to remove the devices. The key model parameters are:

  • The proportion of patients that have either balloon tamponade or Danis stent as a bridging treatment and the proportion of patients that have either TIPS or band ligation as a definitive treatment.

  • Survival 6 weeks after treatment and relative risk of dying at 6 weeks with balloon tamponade compared with Danis stent.

  • Proportion of patients that have adverse events after treatment.

    For full details of the cost evidence, see section 4 of the assessment report in supporting documentation.

The EAC's updates to the cost model change the direction of the cost case

3.6 The EAC updated 5 cost parameters, including the cost of:

  • removing Danis stent (company, £1,257.00; EAC, £1,452.00)

  • re-bleed (company, £3,287.00; EAC, £4,978.75)

  • definitive TIPS treatment (company, £3,928.00; EAC, £4,965.56)

  • definitive band ligation (company, £1,114.00; EAC, £4,983.67)

  • severe hepatic encephalopathy (company, £400.52; EAC £400.56).

3.7 With the updated cost parameters, the EAC's base case shows that Danis stent is cost incurring by £923.00 per person. The company also presented 3 scenario analyses, all of which the EAC considered relevant. Scenario 1 modelled the cost of using each technology by cumulating the costs of the resources needed. Scenarios 2 and 3 explored uncertainty in the assumed impact of the bridge treatment on the choice of definitive treatment including (scenario 3) and excluding (scenario 2) hepatic encephalopathy costs.

Two additional scenario analyses include a second endoscopy for patients who have a balloon tamponade

3.8 Clinical experts highlighted that a second endoscopy at the time of balloon removal is needed for people that have balloon tamponade. The choice of definitive treatment is done on a case-by-case basis regardless of the bridging treatment used. Therefore scenarios 4 and 5 were done to include the cost of a second endoscopy:

  • Scenario 4 – an extension of scenario 2 but with the addition of a second endoscopy for people treated with balloon tamponade based on expert comments.

  • Scenario 5 – an extension of scenario 1 to explore the impact of reduced intensive care unit (ICU) bed days in the Danis stent group and a second endoscopy for people treated with balloon tamponade.

Updates to resource parameters in the micro-costed model are based on expert advice and data about hospital admissions

3.9 Estimates about resource use in the care pathway were updated based on expert advice and were reported as scenarios 5A and 5B. The key parameters changed were the:

  • proportion of patients that had Danis stent inserted in a theatre setting (increased)

  • ICU length of stay in the Danis stent group (reduced) and in the balloon tamponade group (increased)

  • cost of the balloon tamponade procedure (increased).

3.10 Costs were modelled for the proportion of patients that needed transferring from a secondary care setting to a tertiary care setting (scenario 5B only).

  • National Institute for Health and Care Excellence (NICE)