Endoscopic Suturing for Defect Closure in the Upper Gastrointestinal Tract: A Retrospective Cohort Study

Abstract

Background: The increasing availability of gastrointestinal (GI) interventions has raised the need to treat luminal defects. Endoscopic suturing (ES) is a minimally invasive technique that is used for a wide range of indications. This retrospective cohort study aimed to evaluate the performance of ES in treating upper GI defects. Materials and Methods: Data from a tertiary centre were collected for patients undergoing ES to treat upper GI defects. The primary outcome was long-term outcomes, defined as the successful sutures deployment. Secondary outcomes included technical success, immediate clinical success (confirmation of closure at the time of the procedure), recurrence, and complications. Descriptive statistics and x² test were used to calculate the rates of the outcomes and assess any link between independent variables and results. Results: Forty-two procedures were performed on 25 patients between 2018 and 2023. The mean age was 55 (±16.2) years, and 56% were female. The long-term clinical success rate was 69.6% (16/23), the technical success rate was 88.1% (37/42), and the immediate clinical success rate was 91.9% (34/37), with only two (4.8%) adverse events. The overall recurrence rate was 61.8% (21/34). Technical success was higher in the esophagus (92.3%), and stomach (100%) (p = 0.002), and immediate clinical success was more likely in patients with leaks (88.9%) or fistula (95.2%) compared to perforation (50%; p = 0.005). Conclusions: ES demonstrated high rates of technical and immediate clinical success for defect closure in the upper GI tract, with low rates of complications. The benefit is most prominently seen among patients with leaks and fistulas in the stomach and esophagus.

1. Introduction

The increasing prevalence of endoscopic interventions and novel surgical approaches has meant that there is a need for the treatment of the inevitable complications. Among them, perforations and fistulas in the upper gastrointestinal (GI) tract can occur following diagnostic and interventional endoscopic procedures as well as post-surgery, with a range that can raise to 10% [1,2]. They can also develop spontaneously because of peptic ulcer disease, foreign bodies, malignancy, or Boerhaave syndrome [3]. These defects are associated with high morbidity, prolonged admissions, and healthcare expenditure, and can be even life-threatening, especially when they evolve to mediastinitis, peritonitis, and overwhelming sepsis [4,5]. It is also understood that the surgical management of GI defects is associated with considerable morbidity and mortality and can lead to prolonged recovery times in the post-operative period [6].

When endoscopic intervention is considered, there are several options available for closure [7]. The theoretical benefit of all available devices that can anchor the GI tract layers promoted research on their applicability for defects by hypothesizing that bringing the margins closer could resolve defects. In this regard, through-the-scope (TTS) and over-the-scope clips (OTSC) and their combination with endoloops have been used. Furthermore, using fully covered stents to seal the defect’s area is also a reasonable solution, depending on the site of the defect. In this way, perforations, leaks, and fistulae have been managed using endoscopic measures. The European Society of Gastroenterology (ESGE) and the American Gastroenterological Association (AGA) recommend management with TTS clips for small perforations (<10 mm), OTSC for those >10 mm, and stenting for larger ones where applicable [4,8]. However, all the above-mentioned techniques are not always effective, and defect’s management remains challenging and affected by various factors, independent of the endoscopic technique [9]. Recently, the use of endoscopic vacuum therapy attached to a metal stent has been proposed and has demonstrated encouraging results in the treatment of esophageal leaks [10,11]. The variety of choices in the absence of clear evidence requires a multimodal approach involving endoscopists, surgeons, and radiologists in order to plan the optimal strategy and decide the best therapeutic plan, combining closure, drainage, and medical treatment, as indicated in [8,12] (Figure 1).

Endoscopic suturing (ES) is a minimally invasive technique for deploying full thickness sutures in the GI tract [13]. Although it was initially introduced in clinical practice as a method to perform bariatric interventions [14], the current range of indications has become broader, including defect closure following endoscopic full-thickness resection [15,16], post-operative leaks [17], fistula closure [18], perforation repair [19], stent fixation [20], closure following endoscopic resection [21,22], and treatment of large bleeding ulcers, in addition with other techniques, such as hemostatic powders [23,24]. (Figure 2, Figure 3 and Figure 4 ES can be used for large defects with irregular margins or complex defects where endoscopic clipping would not be feasible. Furthermore, the full thickness nature of the closure yields greater tissue approximation than other closure techniques.

The present study aimed to assess the efficacy and safety of ES in defect closure, regarding technical success, short- and long-term clinical outcomes, and adverse events.

2. Results

After applying the exclusion criteria, 25 patients were included to our cohort. A total of 42 ES procedures were performed between August 2018 and March 2023. The average age was 55 (±16.2) years and 14/25 (56%) were female.

In the majority of cases (14/25; 56%) the defect was located in the esophagus and the patients underwent 26 procedures. Seven patients (28–11 procedures) had suturing for a gastric pathology, whereas four cases (16%) involved the duodenum, with only one of them undergoing more than one procedure. The most common indication was fistula/leak, requiring 30/42 (71.4%) of the overall procedures. The remaining nine patients (12/42 procedures; 28.6%) were diagnosed with a perforation.

Table 1. Main characteristics of the sample.

	n = 25	(%)
Gender
Female	14	56
Male	11	44
Mean Age (±SD)	55.1	±16.2
Location
Esophagus	14	56
Stomach	7	28
Duodenum	4	16
Indication
Iatrogenic perforation	8	32
Leak	4	16
Anastomotic fistula	7	28
Post-PEG fistula	4	16
Congenital TOF	1	4
Spontaneous perforation	1	4
Number of procedures needed/patient	1.68	±0.8

depicts the main characteristics of the recruited patients.

2.1. Overall Outcomes

The long-term clinical success rate was 69.6% (16/23) among those with available 30 days follow-up data, including cases with repeated suturing. The mean number of applications to achieve long-term clinical success was 1.68 (+/−0.8).

The overall technical success rate for all the procedures was 88.1% (37/42). An immediate clinical success rate of 91.9% (34/37) was recorded for technically successful procedures; however, the overall recurrence rate was 61.8% (21/34).

There were two (4.8%) adverse events. Both occurred due to transient intraprocedural bleeding controlled endoscopically at the same time. In one of these cases, the patient had predisposing risk factors for bleeding, namely portal hypertensive gastropathy and florid gastric antral vascular ectasia (GAVE). There were no serious adverse events.

2.2. Outcomes According to Indication

In cases with fistula or post-operative leaks, the technical success rate was 96.7% (29/30), with an immediate clinical success rate of 96.6% (28/29). The long-term clinical success rate in this group was 62.5% (10/16), with one adverse event reported (bleeding, 3.3%, [1/30]). The only instance where immediate clinical success was not attained occurred in relation to an anastomotic leak post-sleeve gastrectomy. Interestingly, the rate of recurrence was 67.9% (19/28), which was experienced by 10 patients, with 7 experiencing more than one recurrence episode. The long-term clinical success rate in this subgroup was 40% (4/10). The mean number of applications to achieve long-term clinical success was 1.72 (±0.82).

Regarding perforations, technical success was achieved in 66.7% (8/12) of cases, with an immediate clinical success rate of 75% (6/8), with the two cases where immediate clinical success was not achieved being related to iatrogenic oesophageal perforation. In the first case, the defect had reduced in size by 90% and eventually closed with conservative management. The second case required repeat suturing 8 days later, closing the defect in its entirety. The long-term clinical success rate was 85.7% (6/7), and there was a single reported adverse event (bleeding, 3.3%, [1/30]). The rate of recurrence in these cases was 33.3% (2/6). More specifically, there were two patients in total who experienced recurrence of the index clinical issue, with both achieving long-term clinical success after repeated suturing. The mean number of applications to achieve long-term clinical success was 1.25 +/− 0.5.

When comparing the different scenarios, dealing with perforation was associated with significantly lower rates of technical (p = 0.024) and immediate clinical success (p = 0.005) compared to the other indications, although the recurrence rates and long-term results remained similar.

2.3. Outcomes According to Location

The technical success rate of suturing for oesophageal defects was 92.3% (24/26), with immediate and long-term clinical success rates of 87.5% (21/24) and 57.1% (8/14), respectively. Recurrence was recorded in eight patients after 15/21 (71.4%) procedures.

In the stomach, sutures application was completely successful (100%; 11/11) with the same immediate clinical success rate of 100% (11/11). For patients who completed their follow-up, the long-term clinical success rate was 85.7% (6/7), whereas the rate of recurrence was 45.5% (5/11).

Finally, only 40% (2/5) of cases with duodenal pathology experienced successful suturing, albeit with an optimal immediate and long-term clinical success rate of 100% (2/2). The rate of recurrence was 50% (1/2) and no adverse events were reported.

Table 2. Main results of suturing performance.

	n = 42	(%)	p Value
Technical success	37	88.1
Gender			0.72
Female	18	90.0
Male	19	86.4
Location			0.002
Esophagus	24	92.3
Stomach	11	100
Duodenum	2	40.0
Indication			0.024
Perforation	8	66.7
Leak	9	100
Fistula	20	95.2
Immediate clinical success	34	81
Gender			0.35
Female	15	75.0
Male	19	86.4
Location			0.018
Esophagus	21	80.8
Stomach	11	100
Duodenum	2	40
Indication			0.005
Perforation	6	50.0
Leak	8	88.9
Fistula	20	95.2
Long-term clinical success	16	69.6
Gender			0.38
Female	10	76.9
Male	6	60.0
Location			0.25
Esophagus	8	57.1
Stomach	6	85.7
Duodenum	2	100
Indication			0.44
Perforation	6	85.7
Leak	3	75.0
Fistula	7	58.3
Recurrence	21	61.8
Gender			0.02
Female	6	40.0
Male	15	78.9
Location			0.34
Esophagus	15	71.4
Stomach	5	45.5
Duodenum	1	50.
Indication			0.27
Perforation	2	33.3
Leak	5	62.5
Fistula	14	70.0
Adverse events	2	4.8
Gender			0.95
Female	1	5.0
Male	1	4.5
Location			0.69
Esophagus	1	3.8
Stomach	1	9.1
Duodenum	0	0.0
Indication			0.34
Perforation	1	8.3
Leak	1	11.1
Fistula	0	0.0

presents the main outcomes and the results of sub-analysis.

2.4. Multi-Variate Analysis

Binary logistics regression analysis for all outcomes (

Table 3. Multi-variate regression analysis assessing factors potentially affecting technical success.

	OR	95% C.I.		p-Value
Gender
Female	0.37	0.014	9.79	0.55
Location
Esophagus				0.71
Stomach	0			0.99
Duodenum	4.2	0.143	122,838	0.41
Indication
Perforation				0.68
Leak	0			0.99
Fistula	0.15	0.002	10,065	0.38

), including all available variables, did not reveal any significant co-variates affecting ES outcomes on defect closure.

3. Methods

3.1. Study Design

A single-centre, retrospective cohort study was conducted using the records of an internal electronic database of patients admitted to University College London Hospitals, United Kingdom, following Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Supplementary Table S1) [25]. Data over a five-year period were reviewed between August 2018 and March 2023. The study protocol conformed to the ethical guidelines of the last revision of the Declaration of Helsinki and complied with Good Clinical Practice guidelines [26,27]. Patients’ anonymity was ensured, and the collected data did not contain personal information by which anyone could recognize individual patients; given the retrospective nature of the study, patient consent and Institutional Review Board approval were not required.

3.2. Patients

All patients who had ES and were using the Apollo OverStitch™ (Apollo Endosurgery, Boston Scientific, Boston, MA, USA) device to treat luminal defect in the upper GI tract were eligible to participate in the study. More specifically, cases with anastomotic esophageal fistulas, congenital tracheoesophageal fistulas, gastrocutaneous fistulas, post-operative leaks, iatrogenic esophageal, gastric, and duodenal perforations or spontaneous esophageal perforation were assessed for inclusion in our study.

However, patients who underwent combined procedures for defect closure (e.g., suturing and stenting) or those without adequate follow-up were excluded from our cohort.

3.3. Data Collection

Patients fulfilling the eligibility criteria during the aforementioned period were recruited. The eligibility of the included cases was evaluated by N.A. Patients’ records were retrieved using a local software which enables the storage and retrieval of multiple medical reports including the history, clinical findings, laboratory and imaging examinations, endoscopic procedures, and outcome information of all patients admitted. Variables including gender, age, indication, number of applications, short- and long-term procedural outcomes, and adverse events were recorded. The type of defects encountered in the upper GI tract were subcategorised into three cohorts (fistulas, leaks, full-thickness perforation) to allow for comparative analysis. An Excel file (Microsoft^® Excel for Mac 2020, Microsoft Corporation, Redmond, WA, USA) was created and the variable values were extracted. All data have been stored in a secure server. In cases of conflict, a consensus was met by the intervention of a senior author (R.H.).

3.4. Procedure Details

All procedures were conducted under general anesthesia in an endoscopy suite or surgical theatre by an experienced endoscopist (RJH), with patients fasted for at least eight hours and anticoagulation/antiplatelets therapy managed according to the existing guidelines [28].

After completing a conventional diagnostic esophagogastroscopy and identifying the area of interest, the tissue around the defect was treated with argon plasma coagulation (APC) at 50 W prior to suturing at the endoscopist’s discretion, in order to facilitate tissue approximation post-defect closure [29].

The Apollo Overstitch (Overstitch or Overstitch Sx) device was subsequently mounted onto the endoscope and the sutures were deployed in order to close the defect, with or without the assistance of the Helix anchoring device. The number of sutures were at the endoscopist’s discretion depending on the size and the shape of the defect. Where available, fluoroscopy with contrast infusion at the area of the defect was used to assess for persistent leaks post-procedure.

3.5. Study Outcomes and Definitions

Regarding the defects treated, perforation was defined as a complete, transmural disruption of the GI wall, either traumatic or spontaneous; leak referred to the abnormal escape of GI contents (gas, fluid, or solid material) through a partial disruption or anastomotic failure in the wall of the GI tract, usually after surgery or trauma; and fistula was defined as an abnormal connection or communication between the GI tract and another organ, structure, or the skin.

The primary outcome was the long-term clinical success, which was defined by the resolution of the upper GI defect with the absence of ongoing clinical symptoms at the time of last follow-up, confirmed by endoscopy or imaging (e.g., fluoroscopy, CT).

Secondary outcomes included technical success, defined as the successful completion of all procedural steps of endoscopic suturing to close the defect. Secondly, the immediate clinical success, defined by endoscopic confirmation of closure or use of fluoroscopy at the time of the procedure to exclude extra-luminal contrast leaks. Recurrence during follow-up was defined by the recurrence of the upper GI defect and confirmed by endoscopy or imaging. In addition, we evaluated 30-day adverse events and reported these according to the American Society of Gastrointestinal Endoscopy’s lexicon for adverse events [30].

3.6. Statistical Analysis

Data analysis was performed using the Statistical Package for Social Science Software for Windows (IBM SPSS Statistics, Version 28.0. Armonk, NY, USA: IBM Corp). Continuous variables are presented as mean (±standard deviation, SD) and categorical variables are shown as percentages. Univariable models were used to investigate the individual associations between independent variables and recurrence, while in the multivariable regression, all variables were inserted to assess their relationship with the recurrence over time. Odds ratios (OR) and their 95% CIs were derived from each variable coefficient in the final model. Statistical significance was considered for p values ≤ 0.05 (two-tailed).

4. Discussion

This retrospective cohort study evaluated the application of ES on upper GI tract defects, presenting technical feasibility in 88.1% of the procedures. In cases with successful deployment of sutures, the rate of defect closure on the time of the procedure, was high (91.9%); however, these also showed high recurrence rates (61.8%). Dealing with perforations was the most challenging indication in terms of technical (p = 0.024) and immediate clinical success (p = 0.005) compared to fistulas and post-operative leaks; however, long-term outcomes and adverse events were irrelevant to the indication. The location of the defect impacted the outcomes, with a duodenal application of sutures being technically feasible in 40% of cases compared to 100% and 92.3% for gastric and oesophageal ones, respectively. In terms of adverse events, ES can be considered safe, as only two patients experienced non-severe bleeding, probably due to the existing tissue friability.

Endoscopic suturing represents an advanced procedure, thereby requiring a high level of baseline skills and training [31]. Even in the stomach, which provides space and allows wide movements, the learning curve depends on the endoscopists’ experience and the ability to deal with unexpected scenarios and complications. On the other hand, the narrow esophageal and duodenal lumen are challenging in terms of successful sutures deployment. Chon et al. [32]. retrospectively assessed ES in oesophageal and duodenal leaks and identified a technical success rate of 100%, although only two cases of duodenal defects were recruited. In our study, especially considering duodenal defects, only 40% of cases had proper ES, thus significantly differing from oesophageal and gastric ones. Although the esophagus is also a narrow tract, the technical success of ES was higher (92.3%), probably due to the absence of natural sphincters and angulations, compared to the duodenum (lower oesophageal sphincter, pylorus, duodenal angulation). These natural anatomical obstacles can make it difficult to access the area of interest and increase the risk of further complications as the device increases the overall size of the endoscopic system and protrudes from the endoscope tip. Alternative modalities that could effectively treat these defects could be OTSC in selected cases, or surgical interventions (laparoscopic or open). Therefore, in centres dealing with luminal defects, the availability of alternative closure techniques and devices is mandatory. Another significant issue is the cost-effectiveness of ES, compared with alternatives. Although there are no relative studies in the literature, it is self-evident that, when successful, endoscopic clipping could be a more affordable alternative, whereas the recently introduced X-tack seems to be superior to the traditional ES in terms of costs [33].

Unfortunately, there is no dedicated device offering optimal defect closure results for any type of defect and in any location. Given the high heterogeneity in the clinical scenarios and the broad range of variables affecting the defects closure, quality data comparing suturing with alternatives are lacking. In a retrospective comparison between ES and clipping of lower GI perforations, ES eliminated the need for surgical treatment; however, these findings cannot be generalized [34]. Regarding the other alternatives, conventional self-expandable metal stents seem inferior to over-the-scope clipping or endoscopic vacuum therapy, albeit the existing evidence is weak and further comparative studies are required [35,36,37]. Interestingly, Krishnan et al. [38] evaluated a novel suturing device (X-Tack, Apollo Endosurgery, Austin, TX, USA) for perforation or fistula closure, resulting in a clinical success of 92.4%, with the two failures been related to fistulas rather than perforations. A large multi-centre study, including a broad range of techniques (ES, clipping, vacuum therapy, stenting) concluded that a multi-modality approach is effective in 80.1% of cases, implying that endoscopists should be creative, flexible, and patient when treating defects, using all available tools [39].

Even after ideal sutures deployment or the application of any closing device, long-term clinical success remains challenging. Despite the fact that the rate of defect closure at the time of the procedure was high (91.9%) in our cohort, the recurrence rate remained high (61.8%), thus confirming the existing data in the literature. Therefore, selecting patients with a high likelihood of successful endoscopic treatment is essential. Early intervention, within 30 days, has been associated with an increased rate of successful closure [40]. Similarly, smaller defects, the absence of critical patient condition (hemodynamic instability, intensive care unit admission, respiratory failure), and previous gastrectomy were found to have a positive impact on defect treatment [39]. Fistula closure is per se more challenging, with rates of long-term clinical response being 25–30% [41]. These rates are further reduced in patients with malignancy-related leak/fistula or previous radiotherapy on the site of the defect. The identification of these prognostic factors that may affect the endoscopic closure outcome could be more important than the closure technique itself, thus guiding clinicians to design the appropriate therapeutic plan, likely combining more than one technique depending on the nature (fistula, perforation, leak), the size, the location, and the chronicity of the defect.

The selection of the most appropriate tissue apposition technique remains a challenge, as comparative data are limited and of low quality. Both ES and clipping have demonstrated high effectiveness in managing post-resection defects and perforations. However, addressing fistulae and leaks remains a significant challenge, necessitating focused research and potentially combined treatment strategies. Each of these modalities possesses distinct characteristics that should be carefully evaluated when selecting the most appropriate approach. Through the scope, clips offer the advantage of immediate deployment through the endoscope, enabling the prompt closure of perforations without requiring the endoscope to exit the operative field. In contrast, OTSCs and suturing devices are capable of sealing larger defects but require the removal of the endoscope to reload the device. Newer defect-specific technologies, such as the X-Tack system and vacuum therapy stents, have the potential to revolutionize the management of luminal defects. However, the current body of evidence is of low quality and exhibits considerable variability across studies. This highlights the need for well-designed research that defines clear outcomes while accounting for factors such as defect type, chronicity, location, and etiology. Such studies are crucial to determining the most effective closure techniques tailored to individual patient characteristics [42].

This study has also some limitations. The most significant one, also observed in all the existing studies in this field, is the retrospective non-comparative design due to the special nature of this clinical entity and the limitations of the available techniques and the number of cases. The only study comparing, although retrospectively, ES with clipping was on lower GI perforations. This drawback impacted the ability to collect variables which could affect the outcomes, such as the defect size, and the time of ES after the defect creation. Moreover, records on exact follow-up and exact time of recurrence were absent, thus not allowing further calculations, such as the Kaplan–Meier curve, to depict recurrencies. The rarity of this procedure affected our sample size, which did not allow us to perform extended and reliable sub-analyses. This is also reflected in the logistic regression analysis, where the 95%CIs are broad. Additionally, patient and lesions’ characteristics, technical issues, concerning suturing patterns, and the use of adjunctive devices or measures (e.g., APC) could not be reliably extracted due to potential heterogeneity in the reports. Finally, the short follow-up did not allow the evaluation of ES beyond 30 days.

Endoscopic suturing represents an established and safe technique in defect closure, especially in the esophagus and stomach, providing high rates of technical and immediate clinical success. However, the relatively high rates of recurrence and the heterogeneity among indications and defect characteristics require further high-quality studies to identify where ES can be more effective and guide a personalized approach.