Background
Benign esophageal strictures are often caused by esophageal reflux, the ingestion of caustic substances, esophageal surgery, and radiation therapy [
1‐
7]. These patients most commonly present with dysphagia [
3,
5]. Treatment of benign esophageal strictures with serial endoscopic dilatation using bougies or balloons has been established as a standard therapy [
2,
3,
6], and it is associated with an immediate 80-90% success rate of relieving dysphagia [
3,
6]. However, 30-60% of benign strictures will recur during long-term follow-up. Furthermore, in complex strictures that are usually longer (>2 cm), tortuous, angulated or have a severely narrowed diameter, the recurrence rate is considerably higher [
3,
5,
7]. In these patients, repeated sessions of dilatation without long-term clinical success are associated with increased the risk of developing complications, patient discomfort and recurrent dysphagia; thus, an alternative treatment strategy should be considered [
2,
3,
7]. Surgical procedures, including gastric pull-up and enteral replacement, are potentially curative, but they are associated with high rates of morbidity and mortality, and many patients are not be good surgical candidates or willing to undergo surgery [
1,
3,
6]. Theoretically, temporary placement of expandable esophageal stents permits a longer-lasting dilatation effect, maintains luminal patency and simultaneously stretches the stricture [
3‐
6]. Previous studies have proposed the use of temporary self-expandable metal stents (SEMS) with discouraging results due to the high rate of complications, including embedding of the uncovered metal portion of the stent in the esophageal wall, new stricture formation, necrosis and ulcerations from the stents themselves [
2‐
9].
To avoid complications of partially covered/uncovered stents, temporary placement of 3 different types of expandable stents have been used for the treatment of refractory benign esophageal strictures (RBES): self-expanding plastic stents (SEPSs), biodegradable stents and fully covered SEMSs. Several studies have evaluated the clinical effectiveness of SEPSs [
2,
4,
6,
10‐
17]. Although initial studies with SEPSs showed promising results pertaining to dysphagia relief in the great majority of patients [
10,
11], recent studies have shown less favorable outcomes, with a clinical success rate well below 50% [
2,
4,
6,
12‐
14]. Furthermore, SEPS are also associated with various complications, especially high stent migration rates [
2,
11,
13,
14]. Biodegradable stents have recently been developed and can serve as an alternative for SEPS. Saito et al. reported results from 2 series of patients who received poly-
l-lactic esophageal stents [
18,
19]. In one study Saito et al. observed a high migration rate (10/13 cases), although no symptoms of re-stenosis were observed within the follow-up period in all cases [
19]. Three recent studies have used a novel stent (Ella esophageal stent) composed of the biodegradable polymer polydiaxone [
4,
20,
21]. In these 3 studies authors observed low migration rates with the Ella stent (range: 0–22.2%), and encouraging clinical results with clinical success rates ranging from 33% to 60%. One study comparing the Ella stent with the Polyflex stent observed similar clinical performance and complication rates between the 2 stents as well as a significantly lower reintervention rate for the Ella stent group [
4]. Finally, an alternative to SEPSs and biodegradable stents in the treatment of RBES is offered by fully covered SEMSs [
1,
3,
22,
23]. However, a recent meta-analysis [
24] did not include studies with biodegradable stents, and only one study compared SEPSs and biodegradable stents [
4]. Thus, no study has included or compared the 3 types of stents available for the treatment of RBES, and no case series have addressed the outcomes of patients treated with a fully covered Wallflex stent.
This study compared 3 groups of consecutive patients with RBES who underwent temporary placement of SEPSs, biodegradable stents and fully covered SEMSs. Data were collected to analyze the technical success and clinical outcomes of the stents as evaluated by recurrent dysphagia, complications and reinterventions.
Discussion
According to our findings in a prospective, nonrandomized study of 3 consecutive cohorts of patients with RBES, temporary placement of a biodegradable stent or a fully covered SEMS may lead to a long-term dysphagia-free period in 30 and 40% of patients, respectively. The use of SEPS seems to be least preferable, as it is associated with frequent stent migration, more reinterventions and few cases (10%) of long-term improvement. A long stricture was the only significant factor associated with a higher recurrence rate after temporary stent placement.
Unfortunately, unsuccessful management of benign esophageal strictures by serial endoscopic dilatation using bougies or balloons occurs, and the management of these refractory strictures has been considered challenging [
4,
5,
7,
27]. Surgery can provide definitive treatment but has been associated with considerable mortality and morbidity, including the development of new anastomotic strictures [
2‐
9]. Partially covered metal stents, which are highly effective in the palliation of malignant esophageal strictures, are associated with high complication rates and are not recommended for RBES [
28]. Thus, in a group of patients who have few treatment options available, temporary stent placement can be considered, especially if the procedure is associated with a low complication rate. However, stent placement should be reserved for true refractory strictures, as defined earlier in the Methods section [
26]. In our study, all patients were first submitted to serial dilatations during a number of sessions ranging from at least 5 to 10 attempts (some patients were referred from community hospitals where the number of dilatations was greater than 20). Furthermore, all patients were further submitted to several sessions (3 to 6) of intralesional 4-quadrant steroid injections. As proposed by Siersema et al. [
4,
5,
7,
27], temporary stent placement was a late step in the treatment strategy of benign esophageal strictures.
There are several case series that have evaluated the outcomes of SEPS with conflicting results (Table
4). Some initial studies have shown good results with large numbers of patients (76.5-94%) being dysphagia-free at the end of follow-up and few complications [
10,
11,
15]. However, more recent studies have shown less favorable outcomes, tempering the initial enthusiasm, with success rates between 0 and 30% [
2,
6,
12‐
14].
Table 4
Case series evaluating the outcomes of self-expandable plastic stents for benign esophageal diseases
| 21 | 17 | 13/17 (76.5%) | Migration 9/17 (52.9%) |
| 15 | 15 | 12/15 (80%) | Migration 1/15 (6.7%) |
| 42 | 18 | 17/18 (94.4%) | Migration 1/18 (5.5%) |
| 30 | 22 | 5/83 interventions (6%) | Migration 53/83 stents placed (63.9%) |
| 40 | 40 | 12/40 (25%) | Migration 8/40 (22%) Perforation 2/40 (5%) Bleeding 3/40 (7.5%) Mortality 1/40 (2.5%) |
| 13 | 13 | 3/13 (23%) | Migration 4/13 (30.8%) |
| 5 | 5 | 0/5 (0%) | Migration 3/5 (60%) Perforation 1/5 (20%) Severe chest pain 1/5 (20%) |
| 8 | 8 | 1/8 (12.5%) | Migration 11/13 (85%) |
Regarding clinical success of SEPS, the results of our study are clearly in line with those of recent studies [
2,
6,
12‐
14]. At the end of the follow-up period, only 1 patient was dysphagia free. Migration was observed in 60% of our patients. Again, this is in accordance with some studies that have reported migration rates between 52.9 and 85% [
2,
11,
13,
14]. One explanation for our high migration rate could be the duration of the study (3 months); longer stenting times are associated with higher risks of migration [
2,
11]. In our study, half (n = 3) of the patients had a stent migration after 4 weeks. Alternatively, another explanation could be associated with the result that 8 patients had strictures located in the distal esophagus or in anastomotic areas, as these locations have been associated with higher migration rates for SEPS [
2]. Clinically significant tissue hyperplasia was not observed in our study as one might expect from a fully covered device, and this result is in line with the literature [
2,
4,
6,
10‐
16] and also with a recent pooled analysis where tissue in- and overgrowth was uncommon [
17].
Theoretically, a biodegradable stent is ideal for the treatment of RBES as it becomes embedded in the esophageal wall, reducing migration rates, dissolving spontaneously after placement and decreasing the need for reinterventions to remove the stent. Furthermore, a stent made from a material with good tissue compatibility should overcome/reduce the in- and overgrowth of reactive tissue. Three recent studies have reported preliminary results with the new Ella stent [
4,
20,
21]. An initial prospective study with 21 patients demonstrated a long-term improvement in 9/21 of the patients (42.9%) [
20]. Migration occurred in 2 patients (9.5%), and clinically significant tissue hyperplasia was observed in 1 patient (4.75%). Severe post-stenting pain developed in 3 patients, and minor bleeding was observed in 1 patient. In a prospective study comparing Polyflex stents with ELLA stents, the temporary placement of a biodegradable stent was curative in 6/18 patients (33%) [
4]. Major complications occurred in 4 patients (2 had severe retrosternal pain and 2 had hemorrhaging). Clinically significant tissue in- and overgrowth was observed in 2/18 patients (11%). A recent prospective case series of 10 patients who underwent Ella stent placement for benign esophagogastric anastomotic strictures showed a clinical success rate of 60% [
21]. No migration was observed, and no complications were observed. Signs of tissue hyperplasia were noticed in 4 patients; however, this was associated with reobstruction and symptoms of dysphagia in only 2 patients.
In our study, 30% of patients in the biodegradable stent group were dysphagia-free at the end of the follow-up period, which is in accordance with 2 previously described studies that have reported a clinical success between 33 and 42.9% [
4,
20]. Our migration rate was 20%; however, although this was similar to the rate reported by van Boeckel et al. [
4], this result was unexpected. We anticipated that the embedding of the uncovered stent into the esophagus would prevent migration. In our study, migration occurred in 2 initial patients, and both had long strictures. Eventually, it was found that length selection of the stent was not the best preventative measure of early migration. After the 2 episodes of migration, we began to perform larger balloon dilatation of the stent immediately after deployment to further embed the stent into the esophageal mucosa and reduce the risk of migration. The implementation of balloon dilatation of the Ella stent after deployment did decrease migration rates compared to before balloon dilatation. No other cases of migration were observed and we suggest that this measure can help to mitigate migration after Ella stent deployment. Clinically significant hyperplasia (ingrowth) was observed in 3/10 patients (1 bleeding, 2 reobstructions); this rate was slightly higher than the rates from the 3 studies previously described [
4,
20,
21]. The uncovering of the stent may facilitate hyperplasia along with the eventual chemical reaction of the esophageal mucosa with the polydioxanone. However, the low rate of in- and overgrowth found in the literature and in this study suggests that the biodegradable stent has good tissue compatibility.
Removable, fully covered SEMS are an alternative to SEPS and biodegradable stents. Their fully covered design is thought to induce less reactive hyperplasia, making these stents easier to remove and facilitating stenting over longer periods [
3,
23,
28]. Few published studies, mostly retrospective, have looked at the outcomes of the use of fully covered SEMSs, and they have reported variable data. One interventional radiology study evaluated the effectiveness of temporary stenting in 55 patients [
1]. Three different types of stents were left in place from 1 week to 6 months, and long-term symptom relief was reported in 31% (17/55) of patients. Migration and tissue hyperproliferation were observed in 25 and 31% of patients, respectively. Using the Alimaxx stent, Eloubeidi et al. reported their results from a study of 19 patients [
3]. Clinical success was achieved in 21% of patients, no tissue hyperplasia was reported and migration rates were 37%. A recent study retrospectively reviewed 24 patients with refractory post-esophagogastric anastomosis strictures [
23]. The stents were removed within 4–8 weeks after placement, and 17/24 patients (70.8%) were free from dysphagia at the end of the follow-up period (12 months) as evaluated by intention-to-treat analysis. Tissue hyperplasia was not observed, and migration was observed in only 1 patient (4.2%). Well-tolerated chest pain and reflux symptoms were the most frequent complications reported.
Our study is the first prospective case series to report the outcomes of treatment with a fully covered Wallflex stent. This stent is easy to implant and can be safely removed. However, complications do occur, and migration was the most common complication observed. Migration is frequent for fully covered stents because of the reduced anchoring capacity of these devices [
1,
3,
4,
28]. Migration occurred in 3 patients (30%), which is in accordance with the results of 2 previously cited reports (25 and 37%) [
1,
3]. Compared with the SEPS used in our study, which are also fully covered stents, the migration rate of the Wallflex stent was reduced by a factor of 2 (30% vs. 60%). We believe that this result was related to the larger diameters, stronger radial force and the design of the Wallflex stent, which has 2 flared ends. In contrast, the Polyflex stent has only 1 flared end. Another concern with temporary stent placement is tissue hyperplasia. The fact that these stents do not become embedded in the esophageal wall makes clinically significant tissue in- and overgrowth less likely to occur. In our study of 10 patients, no clinically significant tissue hyperplasia was observed, which is in accordance with other reports [
3,
23]. Only interventional radiology studies have reported significant tissue hyperplasia [
1,
22], which should be interpreted with caution because there was no endoscopic visualization of such tissue hyperproliferation. Furthermore, in our study, one patient was submitted to further stenting over 6 months, and, after that period, the stent was safely removed without signs of hyperplasia. Overall, 4 patients (40%) had long-term resolution of their symptoms after stent removal, which is comparable to the results of 4 earlier studies reporting different clinical successes: 31% [
1], 21% [
3], 48% [
22] and 70.8% [
23] (average: 42.5%).
Although not statistically significant, the results of biodegradable stents or fully covered SEMSs were superior to those of SEPSs in terms of several variables. For example, the duration that patients were dysphagia-free after temporary stenting with a biodegradable stent or a fully covered SEMS was almost 2-fold higher than after placement of a SEPS. While stents were in place, there was a significant difference in the dysphagia scores compared to baseline scores for all stents; all stents were equally effective. However, when comparing baseline dysphagia to post-treatment scores, a significant difference was found in the fully covered group and the biodegradable group but not in the SEPS group; this is a consequence of the differing clinical success rates of the 3 stents. Overall, in the SEPS group, only 1 patient experienced long-term improvement. An explanation for this could be the high rate of SEPS migration when compared with the other 2 stents used. Patients with fully covered or biodegradable stents had their stents in place for longer periods, which may indicate that longer stenting times correlate with higher success rates [
4,
10,
11]. The optimal time of stenting is unknown and has been a subject of debate as clear guidelines have yet to emerge. In our multivariate analysis, the length of the stricture was the only factor associated with higher recurrence rates after temporary stent placement. Thus, it is possible that longer strictures would require longer treatment as suggested elsewhere [
1,
7,
22,
29]. The optimal duration of temporary stenting can also be influenced by the underlying cause of injury or by the time since the injury to the esophagus occurred [
22,
29].
Stenting is safe. In our study, with the exception of migrations, the number of complications was low and, and the complications that occurred were well tolerated. Only 2/30 patients (6%) had significant complications (1 case of bleeding and 1 case of severe pain), which is consistent with most published studies using different expandable stents [
2‐
4,
10‐
12,
14,
15,
17,
20,
21,
23,
24,
30]. The safety of the procedure in association with reasonable clinical success in a group of patients that is very difficult to manage makes temporary stent placement an appealing option.
Poisson regression showed that fully covered SEMSs and biodegradable stents resulted in a significantly lower number of reinterventions. More reinterventions were required in the SEPS group, which was due to the higher migration rate along with the lower clinical success rate of SEPSs. One of the reinterventions in all groups was restenting. The clinical effectiveness of restenting is limited. While stents are in place, patients benefit from another dysphagia-free period, but after stent removal, the number of patients who experience long-term clinical success is low. This result is consistent with several studies [
1,
2,
12,
14,
22,
23]. In our study, restenting was performed in 9 patients, and only one was dysphagia-free after the restenting procedure was performed. In those patients whose primary stent migrated, the second stent generally migrated as well, regardless of whether it was the same or a different fully covered stent, which is in line with other studies in the literature [
1,
2,
14,
22]. After migration, restenting with fully covered devices is of poor value; therefore, we suggest restenting with a biodegradable stent which seems to be a better option, as it is associated with a lower migration rate due to the potential embedding of the biodegradable stent in the esophagus.
We suggest that further prospective randomized studies are needed to compare biodegradable stents with fully covered SEMS to determine clinical effectiveness, optimal duration of stenting, value of restenting for longer periods, cost and patient satisfaction.
The present study has several limitations. First, our nonrandomized design could have introduced sampling bias. However, there were no major differences in location, etiology or stricture length between the 3 groups studied. Another potential weakness is the different time to follow-up for the 3 groups. The SEPS group had a longer follow-up period, which could have influenced the final outcome because, theoretically, the longer the follow-up period the higher the probability of dysphagia recurrence. However, in our study, all patients had at least 8 months of follow-up after stent removal/degradation/migration, and most other studies showed that the benefits of temporary stenting decreased rapidly with time for non-responders [
1,
12,
20,
21,
23]. In our study, no patient had recurrent dysphagia after 9 weeks following stent removal/migration/degradation (Figure
3). Thus, different follow-up times were not likely to be associated with the poor results found in the SEPS group. Our study had a relatively low number of patients per group, which limited the study’s statistical strength. However, most of the studies using self-expanding stents have small population sizes due to the low incidence of RBES, and studies with a patient population large enough to have adequate power to detect minor outcome differences are generally impractical. Finally, this study was performed in tertiary referral centers, and we cannot exclude the possibility of obtaining a poorer outcome in community hospitals.
The strengths of our study are the prospective design with a relatively long-term follow-up. To our knowledge, this is the first study to compare the 3 types of self-expanding stents available for the treatment of RBES and the first case series to report the outcomes of patients treated with a fully covered Wallflex stent.
Authors’ contributions
Conception and design of the study, performing endoscopies, collection of data, analysis and interpretation of data, and drafting the manuscript: JC; performing endoscopies, collection of data, critical revision of the manuscript and approval of the final draft submitted: ML; performing endoscopies, collection of data, critical revision of the manuscript and approval of the final draft submitted: RRT; performing endoscopies, collection of data, critical revision of the manuscript and approval of the final draft submitted: PPM; performing endoscopies, critical revision of the manuscript and approval of the final draft submitted: CR; performing endoscopies, critical revision of the manuscript and approval of the final draft submitted: AC; critical revision of the manuscript and approval of the final draft submitted: BCN; critical revision of the statistical analysis and approval of the final draft submitted: MFS. All authors read and approved the final manuscript.