Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting mibs@nice.org.uk.

Published evidence

Six studies (1 randomised controlled trial, 1 case-control and 4 case series) involving a total of 202 people with refractory variceal bleeding are summarised in this briefing. Out of a total of 202 patients, 134 were allocated to have Danis stent.

Clinically relevant outcomes that were reported include: acute bleeding control, absence of rebleeding, absence of serious adverse events and survival rates.

Table 1 summarises the clinical evidence as well as its strengths and limitations.

Overall assessment of the evidence

Overall, the evidence suggests that Danis stent is effective at controlling acute refractory oesophageal bleeding, as a bridging therapy to further treatment. Results from a randomised controlled study (Escorsell et al. 2016) suggest that, compared with balloon tamponade, Danis stent provides more frequent therapy success, a higher control of bleeding and is associated with a lower incidence of device-related severe adverse events.

Most of the studies evaluating Danis stent were case series, which lacked a comparator and included a relatively small number of patients, making it difficult to draw conclusions about the efficacy of Danis stent compared with standard care. In the 2 studies including comparators (Escorsell et al. 2016 and Maiwall et al. 2018), the therapies used are not recommended by NICE's guideline on acute upper gastrointestinal bleeding in over 16s. Although most of the studies were done in Europe (except 1 study in India), there is a lack of robust UK evidence with only 1 small case series done in an NHS centre. The use of endoscopy for stent deployment appeared inconsistent between studies. Although the company claims that in emergency situations, the device can be inserted without endoscopic image guidance, further research may be needed to support these claims. There was also notable variability between studies in the duration of stent implantation. In some of the studies, the stent stayed in place for up to 14 days (Dechêne et al. 2012; Wright et al. 2010; Zehetner et al. 2008), which is considerably longer than is recommended in the instructions for use (remove within 7 days of implantation). Evidence on the use of Danis stent in non-specialist centres may help to understand whether the technology is suitable for wider adoption by endoscopy units outside of larger specialist centres.

Table 1 Summary of selected studies

Escorsell et al. (2016)

Study size, design and location

A randomised controlled trial involving 28 people admitted to hospital with a diagnosis of cirrhosis and refractory AVB or massive variceal bleeding between March 2009 and January 2013.

Spain in 9 teaching hospitals.

Intervention and comparator(s)

Danis stent (n=13).

Sengstaken-Blakemore Tube (n=15).

Key outcomes

Compared with the balloon tamponade group, the Danis stent group had a more frequent success of therapy (66% compared with 20%; p=0.025), defined by absence of digestive bleeding, absence of SAEs and 15-day survival. There was higher control of bleeding in the Danis stent group at 15 days (85% compared with 47%; p=0.037), but no significant difference was seen at 6 weeks (54% compared with 47%; p=0.25). There were no significant differences between the 2 groups in both 15-day (69% compared with 47%; p=0.39) and 6-week survival (54% compared with 40%; p=0.46). Fewer SAEs happened in the Danis stent group, however this was not statistically significant (15% versus 47%; p=0.077). There was no reduction in duration of ICU stay (8 days versus 8 days; p=0.93) or hospital stay (14 days versus 14 days; p=0.55) between the 2 groups.

Strengths and limitations

A multicentre randomised controlled trial which was independent and independently funded. Use of intention-to-treat analysis. A power calculation was used to determine minimal sample size needed (n=46), however the study used interim analysis results which was 60% of desired sample size. The study was done in Spain and may not be generalisable to the NHS. No female patients were included the Danis stent group. People who had previously had balloon tamponade treatment (n=23) were excluded, however they could be a key target population for the Danis stent. More patients in the balloon tamponade group had earlier TIPS which could have affected survival results.

Wright et al. (2010)

Study size, design and location

A single-centre case series involving 10 people with cirrhosis and variceal haemorrhage, with contraindications to TIPS insertion or balloon tamponade, between March 2007 to July 2008.

UK, tertiary referral liver centre.

Intervention and comparator(s)

Danis Stent.

No comparator.

Key outcomes

Stent insertion was successful on first attempt in 8 out of 10 patients. 1 patient had successful insertion on second attempt, 1 patient had unsuccessful insertion because of the gastric balloon not inflating. Out of 9 patients actively bleeding at time of stent insertion, immediate control of bleeding was achieved in 7 patients (78%), with the remaining 2 patients discovered to have gastric varices. 6 of 9 stented patients (67%) survived the acute bleed, with stent removal at a median of 9 days (range 6 to 14 days). Failure to control the acute bleed was seen in 3 patients, and all 3 patients died because of either multi-organ failure or severe blood loss. Rebleeding 60 days after stent removal was seen in 1 patient. Overall survival rate at 42 days was 50%. Proximal oesophageal ulceration caused by stent insertion was seen in 1 patient.

Strengths and limitations

UK study generalisable to the NHS. Included patients in whom previous balloon tamponade therapy had failed. Study included 3 different care settings (ICU, A&E and endoscopy unit). A relatively small sample size in a single centre with no comparator group. 2 patients had gastric varices which cannot be treated with Danis stent. No statistical analysis done. A short follow-up of 42 days with no long-term outcomes reported. Source of funding was not clear.

Zehetner et al. (2008)

Study size, design and location

A single-centre case series involving 34 patients with acute oesophageal variceal bleeding not controlled with standard therapy between January 2003 to August 2007.

Austria.

Intervention and comparator(s)

Danis stent (n=34).

No comparator.

Key outcomes

Stent placement was successful and uncomplicated for all 34 patients. 33 patients had variceal bleeding only (1 patient had a non-variceal bleeding source), and immediate bleeding control was achieved in these patients. No bleeding recurrence was seen during stent implantation (median: 5 days; range 1 to 14 days). Complications included stent migration in 7 patients (21%) and slight distal oesophageal ulceration in 1 patient (3%) during extraction of the stent. Mortality rates at 30 and 60 days were 26.5% (9 patients) and 29.5% (10 patients), respectively. 2 patients died of liver failure within 24 hours of Danis stent insertion. 7 patients died of liver failure or multi-organ failure after stent removal.

Strengths and limitations

Patients with previous balloon tamponade treatment were included. Study was a single-centre case series with no comparator. No statistical analysis done. There was a short follow-up period of 60 days. Not done in UK so may have limited generalisability to NHS.

Maiwall et al. (2018)

Study size, design and location

A retrospective case-control single-centre study of 88 patients who had acute-on-chronic liver failure with refractory variceal bleeds from 2014 to 2016.

India.

Intervention and comparator(s)

Danis stent (n=35).

Continued treatment with repeat endotherapy and vasoactive drugs or balloon tamponade or both (n=53).

Key outcomes

Control of initial bleeding was significantly greater in Danis stent group compared with controls in pre-match (89% versus 37%; p<0.001) and PRS-matched cohorts (73% versus 32%; p=0.007). 5 patients (14%) in Danis stent group re-bled after initial haemostasis. There was a statistically significant reduction in deaths secondary to bleeding in Danis stent group in pre-match (14% versus 64%; p=0.001) and PRS-matched cohorts (6% versus 56%; p=0.001). 15‑day mortality was significantly reduced in the Danis stent group in pre-match (p=0.004, HR 2.56, 95% CI 1.35 to 4.83) and PRS-matched cohorts (p=0.017, HR 8.1, 95% CI 1.02 to 64.4). However, 6-week mortality was only significantly reduced in PRS-matched cohort (p=0.05, HR 8.1, 95% CI 1.02 to 64.4).

Strengths and limitations

An independent study with a good sample size and comparator group. Done in India so may have limited generalisability to the NHS. Statistically significant differences in patients' baseline characteristics (for example, Child-Turcotte-Pugh score, albumin and AST levels) were seen, however this was controlled using PRS. Selection bias may have happened from endoscopists choosing the therapy based on experience and preference. Study included patients with acute-on-chronic liver failure only, excluding other patients who could be a key target population.

Pfisterer et al. (2019)

Study size, design and location

A retrospective multicentre observational study involving 34 patients with cirrhosis and refractory oesophageal variceal bleeding between January 2009 to December 2016.

Austria 4 tertiary care centres.

Intervention and comparator(s)

Danis stent: Ella CS (n=34).

No comparator.

Key outcomes

Danis stent controlled acute refractory variceal bleeding in 27 patients (79.4%). In the remaining 7 (20.6%) patients, other treatments were applied (EBL, n=3; SEMS renewed, n=2; Linton BT, n=2). Among these patients, only 1 achieved successful long-term bleeding control. Early rebleeding within 6 weeks happened in 6 out of 34 patients (17.6%). 5-day mortality was 20.6% (7 patients). 13 patients (38.2%) died with Danis stent in place. In the 20 patients who had Danis stent successfully removed, (overall median dwell time of 3 days), 7 patients (35%) experienced rebleeding and 5 patients (14.7%) died. Overall 16 patients died because of bleeding (47.1%). Median survival after Danis stent placement was 2.1 months. Complications of Danis stent included stent dislocations in 13 patients (38.2%) and oesophageal ulcers in 4 patients (11.8%).

Strengths and limitations

A multicentre study with a long follow-up (1 year) for observations of adverse events and patient outcomes. Observational, retrospective design with no comparator is relatively low-quality evidence. Not done in UK so may not be generalisable to NHS. 3 patients had additional gastric varices which cannot be controlled with Danis stent. No patients had early TIPS procedures which could have affected mortality rates.

Dechêne et al. (2012)

Study size, design and location

A single-centre case series of 8 patients with cirrhosis and oesophageal variceal bleeding refractory to standard therapy between August 2007 to March 2011.

Germany.

Intervention and comparator(s)

Danis stent: Ella CS (n=8).

No comparator.

Key outcomes

In a total of 8 cirrhotic patients, 9 bleeding episodes were seen (1 patient treated twice in 7-month period). In all patients, immediate bleeding control with Danis stent was achieved. In the 3 patients who progressed to TIPS or liver transplant after stabilisation, all survived without any rebleeding 60 days after stent removal. The remaining 6 patients who only progressed to pharmacological treatment all died, with 3 patients experiencing rebleeding.

Strengths and limitations

An independent study. Reasonably similar baseline characteristics between patients helps to limit potential confounders. Single-centre case series with small sample size limits reliability. No comparison to standard care. Done in Germany and so may not be generalisable to the NHS. A short follow-up period of 60 days.

Abbreviations: AST, aspartate aminotransferase; AVB, acute bleeding from oesophageal varices; CI, confidence interval; EBL, endoscopic band ligation; HR, hazard ratio; ICU, intensive care unit; PRS, propensity risk score; SAE, serious adverse events; SEMS, self-expanding oesophageal metal stents; TIPS, transjugular intrahepatic portosystemic shunt.

Recent and ongoing studies