Safety and efficacy of multiband mucosectomy for Barrett’s esophagus: a systematic review with pooled analysis

Introduction

The incidence of esophageal adenocarcinoma (EAC) has been sharply increasing over the past several decades [1]. In the past, despite significant procedure-related morbidity and mortality, a radical surgical approach was the standard of care for Barrett’s esophagus (BE) with high-grade dysplasia (HGD) or early EAC (T1). However, in recent times, endoscopic management has become a viable approach for this subgroup of patients. Moreover, these minimally invasive endoscopic techniques have been demonstrated to achieve similar oncological outcomes yet confer less risk compared to traditional surgery [2]. The current consensus states that any visible lesion in the context of BE should be endoscopically resected, because of its higher likelihood of containing histologically significant pathology, namely HGD or early carcinoma [2]. Conversely, extended resection aiming at complete metaplasia eradication should not be preferred over ablation strategies owing to the higher risk of stricturing [3-5]. Current endoscopic methods of mucosal resection (EMR) include the lift-and-cut cap-assisted technique and, more recently, the multiband mucosectomy (MBM) technique. The cap-assisted procedure is more technically challenging. Furthermore, submucosal lifting is required for every resection, as well as the delicate re-positioning of the snare within the cap each time. Alternatively, MBM is performed with a device derived from a variceal ligator: the target lesion is suctioned into the ligation cap and a rubber band is released to create a pseudopolyp; then the lesion is resected with a hot snare without the prior requirement of submucosal lifting. MBM has already been demonstrated in previous studies to be as safe and effective as the cap-assisted technique, showing shorter procedure times and lower costs. Since its first description, several prospective and retrospective series on MBM have been published, but data on the efficacy and safety of MBM have never been systematically reviewed. We performed a pooled analysis to evaluate the safety and efficacy of MBM in patients with advanced BE.

Materials and methods

This systematic review with pooled analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [6]. Complete information about sources and search strategy, selection process, data extraction and quality assessment are reported in Appendix A.

Results

Study characteristics and quality

The literature search results in 964 articles (Fig. 1). After a review of titles and abstracts, 36 studies were considered as full text. Fourteen studies finally matched the selection criteria and were deemed eligible for the analysis [7-20]. The included studies were published between 2006 and 2019. All studies but 3 were performed in Europe; the remainder were from North America. Four studies [12,13,15,17] had a prospective design and the other 10 were retrospective [7-11,14,16,18-20]. Eight studies were single-center experiences [11,13,15,17]. The Newcastle-Ottawa Scale [21] for the methodology quality assessment of non-randomized studies resulted in a mean score of 5.4 (range 5-6) (Appendix B, Supplementary Table 1). Characteristics of the included studies are reported in Table 1.

Figure 1 PRISMA flow chart

We provide data on 1334 MBM procedures performed on 986 patients (mean age: 68.4±2.8 years), of whom 84.6% were male (9 studies [7,9-12,14-16,20]). The indications for endoscopic resection varied across the studies, with 62.9% of procedures being performed in order to resect focal lesions and the remaining 37.1% in the context of metaplasia/dysplasia eradication protocols. In 9.4%, patients had undergone previous endoscopic resections (data provided by 5 studies [8,14,16,19], 557 procedures). Visible lesion characteristics are provided in Appendix B, Supplementary Table 2. Considering the histological findings, 31.3% of the lesions were reported to be HGD and 52.2% were adenocarcinomas (7 studies [7,9,11,12,16,18,20]).

Primary outcomes: adverse events

The overall adverse event rate, pooled from the reports of all 14 studies, was 5.3% (I²=83.4%; see Fig. 2). Immediate and postprocedural bleeding occurred in 0.2% (I²=0%) and 0.7% (I²=0%) of procedures, respectively. During 99 out of 1334 (7.4%) procedures, mild bleeding not affecting the procedural completeness was reported. Perforations and strictures were reported at a rate of 0.3% (I²=0%, see Fig. 3) and 3.9% (I²=80.4%), respectively. Safety outcomes of each study are provided in Table 2.

Figure 2 Forest plot: overall adverse events rate CI, confidence interval

Figure 3 Forest plot: perforation rate CI, confidence interval

Discussion

MBM is the most commonly used technique for endoscopic resection in BE. This is the first pooled analysis confirming MBM to be safe for endoscopic resection of superficial Barrett’s neoplasia. Our results are relevant for the following reasons. First of all, serious procedure-related adverse events, such as perforations or severe bleedings, are quite uncommon and large studies assessing an accurate estimation of such complications are still lacking. Pooling data from more than 1300 procedures provides the most feasible and reliable estimation based on the current literature. Secondly, when pooling data in terms of safety, the low level of heterogeneity across the studies is even more relevant when one considers that American and European settings were involved in all the studies included in our analysis. This made our results generalizable where BE is most diffuse and our conclusions stronger.

According to our analysis, the most common event is intraprocedural bleeding, with a pooled rate of 7.4%. In all cases, however, it appears that such bleeding did not affect the completeness of the procedure. Some may consider this type of bleeding to be part and parcel of performing mucosal resection. As such, this degree of bleeding is probably underreported in the literature, particularly from retrospective series. As a matter of fact, the highest rate of such bleeding was reported in the recent 2018 international multicenter by Pouw et al. Intraprocedural bleeding occurred in 48 of 332 procedures (14%), but hemostasis was achieved in all cases using a variety of modalities including snare tip, grasper coagulation, argon plasma coagulation, and/or hemoclip placement.

More serious intraprocedural bleeding preventing completion of the resection occurred at a pooled rate of 0.2%. Clinically significant postprocedural bleeding was also a rare occurrence, at 0.7% (15 total events) across the 13 studies. In these instances, either supportive care or additional endoscopy was required to control hemorrhage from the resection site. Surgical or radiological intervention was not required in any cases.

As for the more feared complication of perforation, most series recorded no perforations as a result of MBM (11 studies) and the perforations reported in the remaining 2 studies yielded a pooled rate of 0.3% (4 total events). In the Pouw et al 2011 randomized study comparing cap-assisted resection and MBM using the Duette system (Cook Endoscopy, Limerick, Ireland), 1 perforation occurred in the MBM group (compared to 3 perforations in the cap-assisted group, P-value not significant). However, the defect in this case could not be closed endoscopically and the patient was referred for surgery. The other 3 perforations were reported in the recent 2018 multicenter MBM study in which Pouw et al evaluated the new Captivator device (Boston Scientific Corp, Marlborough, MA, USA). In 2 of them, perforation was recognized intraprocedurally and the defect was closed with hemoclips (1 with the addition of an endoloop around the clips). Both patients were observed in hospital without any further sequelae. In the third instance, the resection defect had been closed prophylactically with clips but subcutaneous emphysema was recognized postprocedurally and a contrast esophagram showed a small leak at the site of resection. This patient was also observed in hospital and recovered uneventfully without the need for any further intervention. It is possible that unfamiliarity with the new device was related to the higher number of perforations in this trial, but the numbers are too small to draw any firm conclusions. Additionally, 2 recent retrospective, single-center series, involving 40 and 76 patients, reported no differences between the Duette and Captivator systems after comparing endoscopic outcomes and histology features [14,16]. Finally, we consider it remarkable that no delayed perforations have been reported in the MBM literature. This has implications for the modality of patients’ management and follow up, probably meaning that, once the procedure has been successfully completed, there is no need for hospitalization as well as early postprocedural strict follow up. Likewise, no deaths have occurred.

Although these data are reassuring as regards the safety of MBM, they all came from tertiary referral centers. Hence, in order to achieve such safety outcomes, it is clear that any endoscopist who performs band mucosectomy must be very adept at managing adverse events such as intra/postprocedural bleeding, and must also be able to recognize and reliably close small esophageal perforations. This strengthens the argument that mucosal resection of Barrett’s neoplasia is best performed at high volume referral centers by experienced endoscopists.

Esophageal strictures can occur as a result of band mucosectomy. The frequency varies widely in the reported literature, but can approach 40-50% if extensive EMR is performed as a Barrett’s eradication protocol [4,5]. In this meta-analysis, the pooled rate for the development of a symptomatic stricture post-MBM performed for resecting visible lesions was 3.3%, with no heterogeneity across the studies. In all patients, endoscopic dilation allowed a successful management of strictures, with all patients being able to resume a regular diet. For comparison, a recent meta-analysis of the outcomes of endoscopic submucosal dissection for Barrett’s neoplasia reported a pooled esophageal stricture rate of 11.3% [5]. Nonetheless, the low pooled rate of strictures in this meta-analysis, compared to the historical literature on Barrett’s EMR, suggests that band mucosectomy may be less prone to subsequent esophageal stenosis and, in any event, no strictures refractory to endoscopic intervention were reported. This favorable outcome in terms of stricture developments may be related to the MBM technique, which normally does not achieve very deep tissue resection, preserving part of the submucosa.

This meta-analysis also confirmed the high clinical efficacy of MBM, as the pooled complete resection rate was 97.6% with a pooled R0 resection rate of 94.1%. However, these data should be interpreted with caution, since significant heterogeneity was reported.

This meta-analysis has several limitations. First, it did not address long-term follow-up outcomes or the rate of neoplasia recurrence; however, the role of EMR techniques in managing neoplastic BE has already been assessed by several high-quality studies [2]. Secondly, the patient populations varied across studies, and factors such as comorbidity burden and anticoagulant/antiplatelet use would be expected to affect safety outcomes. Moreover, the MBM technique is not standardized worldwide, and factors such as the length and surface area resected, as well as the number of pieces resected per session, could also impact safety and efficacy outcomes. Of course, the burden of such clinical and technical variables needs to be evaluated in future well-powered studies. However, being conclusively reassuring about the safety profile of the MBM approach, our study may be informative for designing such studies. Finally, the overall quality of the studies is limited, given that most were non-randomized and retrospective in nature. However, the low-to-moderate inter-study heterogeneity across the studies reported for the main outcomes offers reassurance as to the reliability of our estimates. This provides endoscopists with a hitherto missing tool, so they may better counsel patients about the safety of the most diffuse approach for treating neoplastic BE.

In conclusion, based on the available data, this meta-analysis confirms the safety and efficacy of MBM for the treatment of advanced Barrett’s neoplasia. Although infrequent, serious adverse events can occur; thus, any endoscopy center performing MBM must have the experience and expertise to manage complications.