A unique type of fully covered metal stent for the management of post liver transplant anastomotic strictures

Background: We report our experience of treating anastomotic strictures using a novel type of fully covered metal stent (FCSEMS). This stent, known as the Kaffes Stent, is short-length with an antimigration waist and easily removable due to long retrieval wires deployed within the duodenum. Methods: 62 patients who had this stent inserted were prospectively analysed. Results: 81% of patients had long-term resolution of their stricture with a significant improvement in liver function tests. Complication rates were 15% with one patient requiring biliary reconstruction. Conclusions: This type of FCSEMS is effective and safe at resolving anastomotic strictures.


Background
Anastomotic strictures are isolated, short-length strictures affecting 4-9% of patients post-liver transplantation, typically associated with technical factors such as bile leaks, the length of the donor bile duct, suture placement and size discrepancy [1,2]. They typically develop 5-8 months after transplantation [3].
Stenting has historically been with plastic stents, on a 3-monthly basis for up to a year. Plastic stents are at risk of migration and stent occlusion, and their efficacy at resolving strictures is low, often leaving patients requiring multiple ERCPs or the need for a biliary reconstruction [4]. FCSEMS (fully covered self expanding metallic stents) have been shown to resolve benign strictures; however, stent migration can occur because the centre of the stent does not always overlie the stricture [5][6][7].
Kaffes stents (Taewoong Medical) are a novel type of FCSEMS that have a short-length, an antimigration waist, and long removal wires which lie within the duodenum for easy removal. The ends of the stents are larger in diameter than the mid-point. This unique design reduces stent migration by producing a radial force against the stricture, contributing to better outcomes. Previous randomised trials have illustrated their success with resolving anastomotic strictures when compared to plastic stenting [8].

Methods
Prospective data on Kaffes stents inserted from December 2016 through to February 2020 was analysed for stricture resolution, adverse event rates and improvements in both symptoms and liver function tests (LFTs). Ethics approval was not required as inserting this type of stent was standard care at our institution.

Inclusion criteria
This was a single centre study performed at King's College Hospital, London. All patients included had duct-to-duct biliary anastomoses and were referred for ERCP based on a radiologically confirmed anastomotic stricture on MR (magnetic resonance) imaging. Patients with complex hilar strictures were excluded.

Method for insertion
Two types of Kaffes stent were inserted -a 40 mm by 10 mm stent and a 40 mm by 8 mm stent. All Kaffes stents were inserted over a HydraJAG guidewire, 0.035-inch (Boston Scientific). Stents aimed to be removed 12 weeks after insertion with stent forceps and a cholangiogram was performed to assess whether the stricture had resolved or whether further stenting was required. Decisions to perform a sphincterotomy or balloon dilatations were made by the endoscopist at the time of the procedure but followed no pre-determined protocol (Fig. 1).

Definitions
Initial resolution of a stricture was defined as either where there was no stricture demonstrated on the cholangiogram (at the time of the second ERCP), or if there was still a stricture, whether a 12 mm extraction balloon was easily able to pass through. Recurrence was defined as where the stricture occurred within the follow-up period. Long-term stricture resolution was defined by the absence of recurrence within the follow-up period. Complications were found from discharge summaries. LFTs were analysed before stent insertion and after removal in patients with resolved strictures only.

Statistical analysis
Categorical and continuous variables were analysed using the Fisher's exact test and independent samples t-test respectively. A paired sample t-test was used to assess for improvement in LFTs. 2 × 2 contingency tables assessed significant differences between categorical variables www.graphpad.com/quickcalcs/contingency1/.

Results
Sixty-two patients had a Kaffes stent inserted, mean age of 53 (SD 11.9, range 13-72) years; 1 patient had a living-related donor right lobe graft; of the remainder 68% were DBD (donation after brain death) grafts. The mean CIT (cold ischaemic time) was 8.4 hours (± 2.4) for the DBD group and 8.9 (± 2.5) for the DCD (donation after circulatory death) group (Table 1). The aetiologies for liver disease are shown in Table 2. 13 (21%) patients had previous plastic stenting and 1 patient had had the traditional type of longer-length FCSEMS (Wallflex™, Boston Scientific) without stricture resolution. The mean time between the transplant and Kaffes insertion was 41 months (SD 72, range 3 days − 327 months). 38 patients had balloon dilatations of strictures prior to stent insertion. 29 had a sphincterotomy at the time of stent insertion, whilst another patient had had a previous sphincterotomy. Table 1 Highlights the main outcomes and specific factors of the Kaffes stent. DBD: Donation after brain death; CIT: Cold ischaemic time. * One patient died of frailty post-transplant 13 months after a Kaffes stent successfully resolved the stricture and another was re-transplanted for chronic rejection so for the purpose of this study, both were not included in the analysis for long-term stricture resolution.

Improvement in symptoms
Forty-six patients were asymptomatic with an anastomotic stricture on imaging (18 with cholestatic LFTs). 16 patients were symptomatic (6 jaundiced, 4 pruritic, 6 cholangitic). Of these, symptoms resolved in 13 (3 jaundiced, 4 pruritic, 6 cholangitic). Of the 3 patients who also continued to be jaundice after stent insertion, 2 had their stent removed within 4 weeks of insertion, the 3rd after 86 days; all 3 were found to have stricture resolution. Two of these patients were found to have papillary stenosis and improved after sphincterotomy; the other patient required re-transplantation due to chronic rejection. Overall, there was a significant improvement in patients' LFTs following stent insertion (see Table 3). patient had a wire-guided perforation of the bile duct without complication, 1 had the retrieval wires uncoil. There was no associated mortality.
Of the stents that were removed at the time of writing, all were removed successfully (mean of 114 days (SD 70), range 3-345 days), although as above, 1 stent needed 2 attempts because the removal wires uncoiled. Stent removal for a patient 345 days after insertion was delayed due to pregnancy; the stent was removed easily without complications.

Discussion
Our results confirm that the Kaffes stent is effective at resolving anastomotic strictures with no patients requiring biliary reconstruction. This one case was due to the inability to achieve wire guided cannulation of the stricture and therefore, unable to place a Kaffes stent. This effectiveness appears to be due to its unique design (Fig. 3D).
Our experience, however, has highlighted a 'learning curve' with using the Kaffes stent. One issue is the delivery of the deployment wires into the duodenum. We recommend the use of forceps, inserted through the scope, to pull the wires out. Care should be taken not to just withdraw the scope with the wires inside it, as this could dislodge the stent. Another issue is that the stents can collect sludge and debris, and this is evident at the time of stent removal. This finding was associated with 3 of our patients developing cholangitis after removal; a balloon trawl is advised to remove any remaining debris after the stent has been removed. One patient became pregnant after her Kaffes stent was inserted. It was therefore not removed for over a year but was successfully removed without complication. However, the longer the stents are left in for, the more likely they are likely to collect sludge and debris. The use of prophylactic ursodeoxcholic acid may be of some benefit although we have not adopted this approach at the time of writing. 12 weeks is therefore the recommended maximum length of time they should be left for. Finally, the Kaffes stent should not be deployed for anastomotic strictures too close to the hilum (< 2 cm) as they tend to migrate distally resulting in biliary obstruction.
There was no relationship between the patients who had sphincterotomies or balloon dilatations, and successful resolution of the anastomotic stricture using the Kaffes stent. The stent is deployed from an 8.5 Fr delivery system so unless the stricture is very tight, the stents were mostly able to cross the stricture without dilatation.

Conclusion
In conclusion, our preliminary observations indicate that the Kaffes stent is highly effective at treating anastomotic strictures. Complication rates are low and stent migration is not a feature. These findings will improve the quality of life for post-transplant patients avoiding repeated ERCPs and potentially biliary reconstruction. Further multi centre data is required to understand the efficacy of this stent in the management of anastomotic biliary structures post liver transplantation.   where it should sit over the stricture. Biochemistry immediately improved. The stent was removed 63 months later with complete resolution of the stricture as shown ( Figure 3C). Figure 3D shows the unique design of the Kaffes Stent with its short-stent length, antimigration waste and long retrieval wires.