Efficacy and safety of endoscopic cardia peripheral tissue scar formation (ECSF) for the treatment of refractory gastroesophageal reflux disease: A systematic review with meta-analysis

Abstract

Background:

Endoscopic treatment is increasingly used for refractory gastroesophageal reflux disease (rGERD). Unlike the mechanism of conventional surgical fundoplication, gastroesophageal junction ligation, anti-reflux mucosal intervention, and radiofrequency ablation have extremely similar anti-reflux mechanisms; hence, we collectively refer to them as endoscopic cardia peripheral tissue scar formation (ECSF). We conducted a systematic review and meta-analysis to assess the safety and efficacy of ECSF in treating rGERD.

Methods:

We performed a comprehensive search of several databases, including PubMed, Embase, Medline, China Knowledge Network, and Wanfang, to ensure a systematic approach for data collection between January 2011 and July 2023. Forest plots were used to summarize and combine the GERD-health-related quality of life (HRQL), gastroesophageal reflux questionnaire score, and DeMeester scores, acid exposure time, lower esophageal sphincter pressure, esophagitis, proton pump inhibitors use, and patient satisfaction.

Results:

This study comprised 37 studies, including 1732 patients. After ECSF, significant improvement in gastroesophageal reflux disease health-related quality of life score (mean difference [MD] = 18.27 95% CI: 14.81–21.74), gastroesophageal reflux questionnaire score (MD = 4.85 95% CI: 3.96–5.75), DeMeester score (MD = 42.34, 95% CI: 31.37–53.30), acid exposure time (MD = 7.98, 95% CI: 6.03–9.92), and lower esophageal sphincter pressure was observed (MD = −5.01, 95% CI: −8.39 to 1.62). The incidence of serious adverse effects after ECSF was 1.1% (95% CI: 0.9%–1.2%), and postoperatively, 67.4% (95% CI: 66.4%–68.2%) of patients could discontinue proton pump inhibitor-like drugs, and the treatment outcome was observed to be satisfactory in over 80% of the patients. Subgroup analyses of the various procedures showed that all 3 types improved several objective or subjective patient indicators.

Conclusions:

Based on the current meta-analysis, we conclude that rGERD can be safely and effectively treated with ECSF as an endoscopic procedure.

1. Introduction

Gastroesophageal reflux disease (GERD) is a disease in which gastroesophageal reflux causes damage to the esophageal mucosa, resulting in symptoms, and is classified as reflux esophagitis with esophageal mucosal damage and non-erosive-reflux-disease with symptoms only.^[1] The global prevalence of GERD is 13.3%,^[2] and it adversely impacts patients’ physical and mental health and quality of life.^[3] The primary treatments for GERD include lifestyle changes, medications, endoscopic interventions, and surgery to improve symptoms and esophageal mucosal damage.^[4,5]

Proton pump inhibitor (PPI) therapy is the most common form of treatment; however, it has a risk of adverse effects, including intestinal infections, pneumonia, kidney disease, and osteoporotic fractures with long-term use of PPIs.^[6,7] Moreover, 30% of patients with GERD whose symptoms are not effectively controlled despite double doses of PPI drugs are referred to as patients with refractory GERD (rGERD).^[7] Patients who do not respond to drug therapy should be evaluated for invasive anti-reflux options, the most typically performed being laparoscopic anti-reflux surgery (LARS). Although LARS is widely recognized as the preferred treatment option for PPI-unresponsive GERD, its failure rate ranges from 3% to 30%, and its reoperation rate ranges between 3% and 6%.^[8] Additionally, complications, including dysphagia and flatulence syndrome, are prevalent with LARS surgery, and the number of people choosing this procedure is decreasing annually because of its greater damage and higher incidence adverse effects.^[9] Consequently, there has been growing interest in endoscopic approaches for treating rGERD recently. Endoscopic treatment modalities for GERD mainly include gastroesophageal junction ligation, anti-reflux mucosal intervention (ARMI), and radiofrequency (RF) ablation. ARMI comprises 2 techniques, including anti-reflux mucosectomy (ARMS) and anti-reflux mucosal ablation (ARMA).^[10]

Unlike the traditional surgical LARS procedure, gastroesophageal junction ligation, ARMI, and RF ablation have similar anti-reflux mechanisms. All these endoscopic treatments elevate the pressure on the lower esophageal sphincter (LES) and its surrounding tissues by using various techniques to create scar tissue around the cardia, thus increasing the non-external force in the cardia and surrounding tissues to reach the anti-reflux goal. Because of their common anti-reflux mechanism, endoscopic cardia peripheral tissue scar formation (ECSF) has been introduced to describe the 3 types of endoscopic therapies.

This study aimed to conduct a meta-analysis to evaluate the safety and effectiveness of ECSF in treating patients with rGERD to offer a dependable theoretical foundation for the clinical management of GERD.

2. Materials and methods

2.1. Ethics

A meta-analysis protocol has been prepared in accordance with the PRISMA guidelines. The registration number is CRD42023427346. This meta-analysis did not require informed consent for participation.

2.2. Literature inclusion and exclusion criteria

The inclusion criteria were as follows: research type: included all randomized controlled trials, retrospective or prospective study; study participants: patients with confirmed PPIs-refractory GERD, whose symptoms (retrosternal heartburn and/or regurgitation) present at least 3 times per week not responding to a double dose of PPIs for 8 to 12 weeks, and no restrictions on race, age, sex, or occupation; study methods: Using ECSF as a treatment for patients with rGERD, with a follow-up period of over 1 month; interventions: study participants received any one of endoscopic fundoplication as follow: gastroesophageal junction ligation, ARMI, and RF; outcome indicators, including gastroesophageal reflux disease health-related quality of life (GERD-HRQL) score, gastroesophageal reflux questionnaire score (GERD-Q), PPIs medication use, time to acid reflux at pH < 4.2, lower esophageal pressure, esophagitis status, patient satisfaction and adverse reactions. All included studies contained one or more outcome indicators with complete data.

The exclusion criteria were as follows: patients from special populations (including patients with obesity, children, and patients with a history of esophageal or gastric surgery); studies with incomplete information; case reports, animal studies, preclinical studies, reviews, and meta-analyses; and studies with < 10 enrolled cases. Follow-up shorter than 1 month.

2.3. Search strategy

Databases such as PubMed, Embase, Web of Science, China Knowledge Network, Wanfang data were searched from January 2011 to July 2023, and the English search terms were ((ARMS) or (Antireflux Mucosectomy) or (ARMA) or (antireflux ablation) or (Peroral Endoscopic Cardial Constriction [PECC]) or (cardial constriction) or (ligation) or (RF) or (stretta) or (antireflux surgery) or (endoscopic treatment)) and ((GERD) or (gastroesophageal reflux disease) or (Esophageal Reflux)). The Chinese search terms were “gastroesophageal reflux” and (“radiofrequency” or “cardia narrowing” or “mucosal resection” or “fundoplication” or “endoscopic treatment”).

2.4. Literature selection and information retrieval

Two investigators independently conducted the literature screening process, with a third investigator deciding whether to include a study in case of controversy. The titles and abstracts of the identified studies were first read to exclude irrelevant literature, and subsequently further screened the articles based on the inclusion and exclusion criteria by reading the full text. For literature for which the full text was unavailable, attempts were made to obtain the necessary information by sending emails to the authors. The study was included if complete data were successfully obtained; otherwise, it was excluded.

Data extraction was performed using an Excel spreadsheet, and the extracted data from the included studies comprised the following: general characteristics: authors’ names, nationality, and publication date; general description: sample size, sex, age, BMI, follow-up time, treatment measures, study type; and relevant outcome indicators: GERD-HRQL score, GERD-Q score, DeMeester score, time to acid reflux at pH < 4.2, lower esophageal pressure, PPIs use, esophagitis, and patient satisfaction. In cases where the format of the data for the final outcome indicators in the study did not align with the requirements of this study, the data were transformed using the relevant calculation formula.^[11–13]

2.5. Literature quality assessment

The studies included in this meta-analysis included 20 prospective and 10 retrospective studies. Their quality was assessed using the Newcastle-Ottawa Scale.^[14] The Newcastle-Ottawa Scale has 3 main evaluation categories, which are subject selection, comparability, and exposure/outcome assessment. Each category contained several subitems that were evaluated, resulting in a total score of 9 points. In the entry of follow-up time, it was stipulated that a follow-up time >1 year was given 1 point; otherwise, no score was given. In the entry of data completeness between the exposed and non-exposed groups, it was stipulated that a lost follow-up rate of < 10% was scored 1 point; otherwise, no score was assigned.

Several of the included studies were randomized controlled trials. The studies that were included in the review were assessed for risk of bias using the evaluation tool recommended by the Cochrane Collaboration Network, which included randomization method, blinding, allocation concealment, completeness of outcome data, selective reporting of study results, and other biases, each of which was categorized based on its specific circumstances as “low risk,” “unclear,” or “high risk.”^[15]

2.6. Statistical analysis

The included studies were analyzed using Review Manager 5.4 analysis software, and forest plots and heterogeneity and publication bias tests were performed. The effect measures of relative risk (RR) and mean difference (MD) were used as indicators for statistical and measurement data, respectively, and the effect indicators were expressed as 95% confidence intervals (95% CI). The Q test and I² value were employed to assess the heterogeneity among studies. Fixed-effects model was applied if P ≥ .1 and I² ≤ 50%, indicating no significant heterogeneity among the studies. Random-effects model was employed when heterogeneity among studies was observed (P < .1 and I² > 50%). Sensitivity analysis was conducted by reapplying the fixed-effects model and comparing the outcomes with those derived from the random-effects model to evaluate the robustness of the results. Inverted funnel plots were constructed to assess publication bias for the outcome variables included in more than 10 studies.

3. Results

3.1. Study selection and characteristics

Based on the Chinese and English keywords, 15,951 documents were retrieved from the initial review. After multiple screenings, 37 studies were included. Among them, 7 were randomized controlled trials,^[16–22] and 30 were cohort studies (10 retrospective and 20 prospective studies)^[23–51] Overall, 1732 patients were included. Figure 1 illustrates the literature screening process, and Table 1 presents the fundamental characteristics of the included studies.

Figure 1.

Flow diagram.

Table 1.

Baseline characteristics of the included studies.

First author	Country	Patients	Publication yr	Sex (male/female)	Age (yr, x ± s)	BMI (kg/m2, x ± s)	Follow-up time	Interventions	Outcome measures	Type of study
Liu Sheng Zhen	China	60	2021	37/25	53.8 ± 9.9	23.5 ± 2.3	6 mo	C-BLART	②③④⑤⑥⑦⑨	Prospective study
Zhi Tong Li	China	68	2021	38/30	45.7 ± 11.7	–	6 mo	PECC	④⑤⑨	Retrospective study
Li Jian Liu	China	30	2022	11/19	51.13 ± 3. 82	–	2 mo	PECC	②③⑥	RCT
Hua Shi	China	23	2021	17/6	–	–	6 mo	PECC	①⑤⑥⑦⑨	RCT
Wenxi Jiang	China	16	2022	13/3	49.1 ± 10.4	22.9 ± 2.4	8 mo	PECC	⑤⑦⑧	Retrospective study
Haiqing Hu	China	13	2018	9/4	53.3 ± lO.5	–	6 mo	PECC	①②⑤⑥⑧	Prospective study
Waseem M	Egypt	75	2018	49/26	39.3 ± 5.1	–	12 mo	EBL	④⑤⑨	Prospective study
Honggang Li	China	24	2022	14/10	45.74 ± 12.75	–	2 mo	PECC	①	RCT
Shuai Tang	China	30	2020	20/10	47.66 ± 10.28	–	1 mo	PECC	①③	RCT
Dezhi He	China	28	2020	12/16	56.32 ± 8.19	–	12 mo	PECC	②⑥⑦	Retrospective study
Juelei Wang	China	15	2020	13/2	50.89 ± 13.69	–	2 mo	PECC	②④⑤⑥⑦⑧	Retrospective study
Yue Chang	China	50	2020	29/21	53.74 ± 9.79	–	12 mo	PECC	③⑤⑦	Prospective study
Chu Kuang Chou	China	23	2023	19/4	44.4 ± 11.7	23.83 ± 11.17	3 mo	ARMI	②④⑤⑥⑦⑨	Prospective study
Xinke Sui	China	39	2022	22/17	57.21 ± 12.881	23.59 ± 1.73	6 mo	ARMS	①④⑤⑦	Prospective study
Jian He	China	69	2022	44/25	–	–	6 mo	ARMS	①②③④⑤⑥	Prospective study
Dezhi He	China	20	2020	13/7	51.9 ± 10.11	–	12 mo	ARMS	②⑥⑦	Retrospective study
Inkyung Yoo	Korea	33	2019	11/22	51.3 ± 16.3	23.5 ± 4.1	6 mo	ARMS	②③⑤⑥⑦	Prospective study
Kazuya Sumi	Japan	109	2021	63/46	50.4 ± 15.7	–	12 mo	ARMS	②④⑤⑥⑦	Retrospective study
Gaurav Patil	India	62	2020	44/18	36 ± 9.9	–	12 mo	ARMS	④⑤⑥⑦⑨	Prospective study
Yan Wang	China	18	2023	10/8	59.39 ± 14.05	23.97 ± 2.97	2 yr	ARMS	④⑤⑦	RCT
Harry J Wong	USA	33	2020	22/11	55 ± 17	27.0 ± 4.8	6 mo	ARMS	①④⑤	Retrospective study
Xinyi Yang	China	18	2021	12/6	53.11 ± 7.62	23.82 ± 2.87	6 mo	ARMS	①②④⑤⑦⑨	Retrospective study
Xinyi Yang	China	21	2021	16/5	51.33 ± 12.39	24.01 ± 3.65	6 mo	ARMS	①②④⑤⑦⑨	Retrospective study
Sun Jun Gao	China	18	2023	12/6	54.66 ± 8.32	24.51 ± 1.16	6 mo	ARMS	②③④⑥⑦	RCT
Sun Jun Gao	China	18	2023	11/7	58.49 ± 6.70	24.68 ± 1.06	6 mo	ARMS	②③④⑥⑦	RCT
Laurent Monino	France	21	2020	11/10	56.87 ± 14.47	24.35 ± 4.58	3 mo	ARMS	④⑤	Prospective study
Zhi Hua Lan	China	40	2021	22/18	49.19 ± 5.87	–	6 mo	ARMS	②⑤	Retrospective study
Kalapala	India	29	2021	19/10	39.5 ± 12.3	–	6 mo	ARMA	①②⑤	Prospective study
Hernández Víctor	USA	108	2020	61/67	37.73 ± 11.88	29.62 ± 4.16	3 yr	ARMA	①②④⑤⑥⑨	Prospective study
Mayo Tanabe	Japan	12	2019	7/5	53.54 ± 14.07	–	2 mo	ARMA	①④⑤	Prospective study
Yan Wang	China	16	2023	11/5	54.31 ± 13.05	22.22 ± 2.48	2 yr	RF	④⑤⑦	RCT
XinkeSui	China	30	2022	12/18	52.04 ± 9.502	23.78 ± 1.31	6 mo	RF	①④⑤⑦	Prospective study
SuyuHe	China	28	2020	16/12	45.4 ± 9.6	22.1 ± 2.9	6 mo	RF	②③④⑤⑥	Prospective study
Peipei Liu	China	27	2019	20/7	49.56 ± 12.31	–	12 mo	RF	①②③④⑤⑥⑧⑨	Prospective study
Hai Feng Liu	China	90	2011	57/33	51 ± 13	–	12 mo	RF	①④⑤⑧⑨	Prospective study
Viswanath Y	England	50	2014	15/35	52.3 ± 13.9	–	2 yr	RF	①⑤④⑧	Prospective study
Lifeng Ma	China	86	2020	52/34	52.6 ± 6.2	27.2 ± 10.9	12 mo	RF	②③⑤⑥	Prospective study
Yue Chang	China	50	2020	27/23	53.21 ± 9.67	–	12 mo	RF	③⑤⑦	Prospective study
Di Lu	China	27	2022	16/11	55.7 ± 7.3	–	6 mo	RF	②③④⑤⑥⑦	Retrospective study
Fei Xu	China	50	2020	31/19	48.25 ± 7.58	22.5 ± 3.5	6 mo	RF	②③④⑤⑥	Prospective study
Zhi Hua Lan	China	40	2021	20/20	49.15 ± 5.85	–	6 mo	RF	②⑤	Retrospective study
Xin Lu	China	25	2019	19/6	53.3 ± 5.6	–	12 mo	RF	④⑤⑦	RCT
Li Jian Liu	China	30	2022	13/17	49. 56 ± 4. 83	–	2 mo	RF	②③⑥	RCT
Wei Tao Liang	China	60	2015	25/35	47.4 ± 10.7	–	12 mo	RF	④⑤	Prospective study

Outcome measures:①GERD-HRQL score; ②AET, time to acid reflux at pH < 4.2; ③lower esophageal pressure; ④PPIs use; ⑤adverse events; ⑥DeMeester score; ⑦GERD-Q score; ⑧patient satisfaction; ⑨esophagitis; studies without available data are not included in the table; “–” means not mentioned in the study.

ARMA = anti-reflux mucosal ablation, ARMI = anti-reflux mucosal intervention, ARMS = anti-reflux mucosectomy, BMI = body mass index, C-BLART = clip band ligation anti-reflux therapy, EBL = endoscopic band ligation, PECC = Peroral Endoscopic Cardial Constriction, RCT = randomized controlled trial, RF = radiofrequency.

3.2. Literature quality assessment

A total of 7 randomized controlled studies were included for analysis.^[16–22] Figure 2 displays the appraisal of the risk of bias for the studies included in the analysis. Moreover, 20 prospective and 10 retrospective studies were included,^[23–51] and the evaluation of study quality is shown in Table 2. Five studies had an overall score of 8, 8 studies had a score of 7, and 8 studies had a score of 6, with no very low-quality literature.

Figure 2.

Risk bias map for inclusion in RCT studies. RCT = randomized controlled trial.

Table 2.

Cohort study quality evaluation form.

Sources of literature	Subject selection	Comparability	Exposure/outcome	Total score
Liu Sheng Zhen et al	3	1	1	5
Zhi Tong Li et al	2	1	2	5
Wenxi Jiang et al	3	0	2	5
Haiqing Hu et al	2	0	3	5
Waseem M et al	3	2	3	7
Dezhi He et al	3	2	2	7
Juelei Wang et al	3	1	1	5
Yue Chang et al	2	2	3	8
Chu Kuang Chou et al	3	2	2	7
Sui Xinke et al	3	2	3	8
Jian He et al	3	1	2	6
Inkyung Yoo et al	3	1	2	6
Kazuya Sumi et al	3	2	2	7
Gaurav Patil et al	3	1	3	7
Harry J Wong et al	3	2	3	8
Xinyi Yang et al	2	2	2	6
Laurent Monino et al	3	2	2	7
Zhi Hua Lan et al	2	1	2	5
Kalapala et al	2	1	2	5
Hernández Víctor et al	3	2	2	7
Mayo Tanabe et al	3	1	2	6
Suyu He et al	2	2	2	6
Peipei Liu et al	3	1	3	7
Hai Feng Liu et al	3	2	3	8
Viswanath Y et al	2	1	3	6
Lifeng Ma et al	3	2	3	8
Di Lu et al	2	2	1	5
Fei Xu et al	3	2	1	6
Zhi Hua Lan et al	2	2	1	5
Wei Tao Liang et al	3	1	2	6

Newcastle-Ottawa Scale (NOS): 0–3 is classified as low quality, 4–6 as medium quality, 7–9 as high quality.

3.3. Meta-analysis results

3.3.1. Symptom improvement.

Improvement in GERD symptoms was measured using 2 reliable scoring systems.

For GERD-HRQL scores, 13 studies, comprising 15 study groups, were incorporated in the analysis, with a total of 611 cases. Using a random-effects model revealed substantial heterogeneity among the studies (P < .00001, I² = 97%). The results of the meta-analysis showed that the 3 forms of ECSF significantly reduced the GERD-HRQL scores in patients with GERD. As illustrated in Figure 3A, the MD was 18.27 (95% CI: 14.81–21.74, P < .00001). Replacement with a fixed effects model did not significantly change the results. The inverted funnel plot suggested no significant publication bias (Fig. 4A).

Figure 3.

Forest plot of the effect of ECSF on symptom. ECSF = endoscopic cardia peripheral tissue scar formation.

Figure 4.

Funnel plot of outcome indicators.

For GERD-Q scores, 16 studies involving 23 study groups comprising 508 patients were evaluated using this scoring system. Nevertheless, substantial heterogeneity was observed among the studies (P < .00001, I² = 95%) using a random-effects model. Meta-analysis showed that ECSF significantly reduced GERD-Q scores in patients with GERD, with a MD of 4.85 (95% CI: 3.96–5.75, P < .00001) between the groups, as shown in Figure 3B. Notably, replacement with a fixed effects model did not significantly change the results. The inverted funnel plot suggested no significant publication bias (Fig. 4B).

3.3.2. Endoscopic and pH monitoring.

For DeMeester scores, 18 studies with 21 study groups were eligible, with a total of 772 cases and significant heterogeneity among studies (P < .00001, I² = 98%), using a random-effects model. The results of the meta-analysis indicated that ECSF significantly reduced DeMeester scores in patients with GERD overall, with a MD of 42.37 (95% CI: 31.40–53.34, P < .00001) (Fig. 5A). Replacement with a fixed effects model did not significantly change the results. The inverted funnel plot suggested no significant publication bias (Fig. 4C).

Figure 5.

Forest plot of the effect of ECSF on endoscopic. ECSF = endoscopic cardia peripheral tissue scar formation.

For acid exposure time (AET), 18 studies comprising 23 study groups and 822 cases were deemed eligible for inclusion. However, substantial heterogeneity was observed among the studies (P < .00001, I² = 98%) using a random-effects model. The meta-analysis findings revealed a significant decrease in the percentage of AET in patients with GERD who received ECSF, with a MD of 7.98 (95% CI: 6.03–9.92, P < .00001) (Fig. 5B). Replacement with a fixed effects model did not significantly change the results. The inverted funnel plot suggested no significant publication bias (Fig. 4D).

For LES pressure, the meta-analysis demonstrated that the use of ECSF resulted in a significant increase in LES pressure among patients with GERD, as evidenced by a MD of −5.01 (95% CI: −8.39 to −1.62, P = .004) according to Figure 5C. Replacement with a fixed effects model did not significantly change the results. The inverted funnel plot suggested no significant publication bias (Fig. 4E).

3.3.3. Esophagitis.

Nine studies evaluated esophagitis before and after administering ECSF. There was significant heterogeneity between the studies (P < .00001, I² = 83%), employing a random-effects model. Meta-analysis results showed a reduction in the proportion of patients with GERD and esophagitis after ECSF treatment (77.9% vs 21.9%, RR = 3.36, 95% CI: 2.15–5.26, P < .00001) (Fig. 6A). Replacement with a fixed-effects model did not significantly change the results, and an inverted funnel plot analysis was not conducted because of the limited number of studies included.

Figure 6.

Funnel plot of the effect on other indicators.

3.3.4. PPIs usage.

Twenty-four studies comprising 27 study groups evaluated using PPIs before and after pancreatic tightening. A random-effects model revealed significant heterogeneity among the studies (P < .00001; I² = 68%). The outcomes of the meta-analysis demonstrated that ECSF significantly reduced the use of PPIs in patients with GERD (32.6% vs 97.6%, RR = 2.93, 95% CI: 2.51–3.42, P < .00001) (Fig. 6B). Replacement with a fixed effects model did not significantly change the results. The inverted funnel plot suggested no significant publication bias (Fig. 4F).

3.3.5. Patient satisfaction.

Eight studies documented patient satisfaction (number of satisfied patients) before and after the ECSF treatment. There was substantial variation among the studies (P < .00001, I² = 84%) analyzed using a random-effects model. The results of the meta-analysis showed a significant increase in patient satisfaction after ECSF treatment (16.5% vs 83.2%, RR = .14, 95% CI: 0.06–0.35, P < .0001) (Fig. 6C). Inverted funnel plot analysis was not conducted because of the limited number of studies included.

3.3.6. Adverse events.

Serious and nonserious adverse events (SAEs and Non-SAEs, respectively) were observed. SAEs are defined as events that necessitate extended hospitalization, additional medical intervention, or surgical procedures, life-threatening or death, mainly pneumothorax, gastrointestinal perforation, and severe inflammation, among other events. Non-SAEs refer to anticipated side effects and symptoms, predominantly gastrointestinal in nature, including epigastric pain, abdominal distension, bloating, diarrhea, and dysphagia. This study primarily recorded 4 items as follows: sore throat, chest pain, dysphagia, perioperative infection and bleeding.

Overall, 34 studies comprising 38 study groups reported the perioperative adverse events associated with ECSF (Table 3). The overall SAE rates were 1.1% (95% CI: 0.9%–1.2%) and 2.1% (95% CI: 1.8%–2.5%) in the ARMI group, and 0.5% (95% CI: 0.4%–0.7%) in the RF group. It is worth mentioning that none of the 348 patients in the 9 studies included in the ligation group experienced SAEs.

Table 3.

Occurrence of perioperative adverse events.

Study	Patient, n	Overall SAE, n(%)	Sore throat, n(%)	Thoracalgia, n(%)	Dysphagia, n(%)	Bleeding or infection, n(%)
Gastroesophageal Junction Ligation
Liu Sheng Zhen Etc.	60	0	–	–	3 (5.0)	–
Waseem M Etc.	75	0	–	–	19 (25.3)	–
Hua Shi Etc.	23	0	–	2 (8.7)	2 (8.7)	–
Juelei Wang Etc.	15	0	–	3 (20.0)	3 (20.0)	–
Haiqing Hu Etc.	13	0	–	1 (2.9)	2 (15.4)	–
Wenxi Jiang Etc.	16	0	–	10 (62.5)	–	2 (12.5)
Yue Chang Etc.	50	0	–	–	2 (4.0)	2 (4.0)
Dezhi He Etc.	28	0	–	–	–	–
Zhi Tong Li Etc.	68	0	–	25 (36.8)	28 (41.2)	2 (2.9)
Subtotal	348	0	–	41 (30.4)	59 (19.4)	6 (4.5)
ARMI
Chu Kuang Chou Etc.	23	0	0	0	1 (4.3)	2 (8.7)
Xinke Sui Etc.	39	0	–	–	1 (2.6)	–
Jian He Etc.	69	0	–	–	45 (65.2)	–
Dezhi He Etc.	20	0	–	–	–	–
Inkyung Yoo Etc.	33	0	–	–	2 (6.1)	–
Kazuya Sumi Etc.	109	3 (2.8)	–	–	13 (11.9)	–
Gaurav Patil Etc.	62	2 (3.2)	–	–	5 (8.1)	3 (4.8)
Yan Wang Etc.	18	0	–	–	5 (27.8)	–
Harry J Wong Etc.	33	3 (9.1)	–	–	–	1 (3.0)
Xinyi Yang Etc.	18	1 (5.6)	–	–	2 (11.1)	–
Xinyi Yang Etc.	21	5 (23.8)	–	–	7 (33.3)	1 (4.8)
Laurent Monino Etc.	21	0	–	–	3 (14.3)	4 (19.0)
Zhi Hua Lan Etc.	40	0	1 (2.5)	1 (2.5)	1 (2.5)	1 (2.5)
Kalapala Etc.	29	0	–	–	–	–
Hernández Víctor Etc.	108	0	–	13 (12.0)	14 (13.0)	4 (3.7)
Mayo Tanabe Etc.	12	0	–	–	1 (8.3)	–
Subtotal	655	14 (2.1)	1 (1.6)	14 (8.2)	100 (17.5)	16 (5.2)
RF
Xinke Sui Etc.	30	0	–	–	–	–
Suyu He Etc.	28	0	2 (7.1)	–	–	–
Peipei Liu Etc.	27	0	27 (100)	–	–	1 (3.7)
Hai Feng Liu Etc.	90	0	–	9 (10.0)	–	5 (5.6)
Viswanath Y Etc.	50	0	–	–	–	–
Lifeng Ma Etc.	86	0	–	–	2 (2.3)	–
Yue Chang Etc.	50	3 (6.0)	–	–	5 (10.0)	5 (10.0)
Yan Wang Etc.	18	0	–	–	0	–
Di Lu Etc.	27	0	–	–	–	–
Fei Xu Etc.	50	0	37 (74.0)	8 (16.0)	6 (12.0)	0
Zhi Hua Lan Etc.	40	0	1 (2.5)	0	0	0
Xin Lu Etc.	25	0	–	–	0	1 (4.0)
Wei Tao Liang Etc.	60	0	–	–	–	–
Subtotal	581	3 (0.5)	67 (46.2)	17 (9.4)	13 (4.8)	12 (4.3)
Total	1584	17 (1.1)	68 (32.7)	72 (14.8)	172 (15.0)	34 (4.7)

Studies without available data are not included in the table; “–” means not mentioned in the study.

ARMI = antireflux mucosal intervention, SAEs = serious adverse events, RF = radiofrequency ablation.

For Non-SAEs, among the studies that provided complete data, the highest incidence of dysphagia was 17.5% (95% CI: 15.0%–17.0%) in the ARMI group. Moreover, postoperative chest pain was the most frequent in the ligation group, with approximately one-third of patients experiencing postoperative chest pain with an incidence of 30.4% (95% CI: 27.5%–33.3%), while the incidence of dysphagia was 19.4% (95% CI: 17.8%–21.0%). Meanwhile, the RF group had the highest number of patients with postoperative sore throat, with an incidence of 46.2% (95% CI: 39.6%–52.9%). The predominant reason for this is that Liu et al reported a 100% incidence rate of postoperative sore throat in a 27-patient cohort after RF. Although Xu et al reported that the incidence of sore throat was 74%, other studies reported a much lower incidence.

AGREE classification has also been utilized to detail endoscopic adverse events (Table 4).

Table 4.

AGREE classification.

Study	Patient, n	No adverse event	Grade I	Grade II	Grade III and above
Gastroesophageal Junction Ligation
Liu Sheng Zhen Etc.	60	57	3	0	0
Waseem M Etc.	75	26	49	0	0
Hua Shi Etc.	23	19	4	0	0
Juelei Wang Etc.	15	9	6	0	0
Haiqing Hu Etc.	13	10	3	0	0
Wenxi Jiang Etc.	16	4	10	2	0
Yue Chang Etc.	50	45	2	3	0
Dezhi He Etc.	28	28	0	0	0
Zhi Tong Li Etc.	68	13	53	2	0
ARMI
Chu Kuang Chou Etc.	23	20	1	2	0
Xinke Sui Etc.	39	38	1	0	0
Jian He Etc.	69	24	45	0	0
Dezhi He Etc.	20	0	0	0	0
Inkyung Yoo Etc.	33	29	–	–	4
Kazuya Sumi Etc.	109	93	13	2	1
Gaurav Patil Etc.	62	0	52	3	7
Yan Wang Etc.	18	13	5	0	0
Harry J Wong Etc.	33	29	–	1	3
Xinyi Yang Etc.	39	29	4	0	6
Laurent Monino Etc.	21	14	4	0	3
Zhi Hua Lan Etc.	40	36	2	2	0
Kalapala Etc.	29	29	0	0	0
Hernández Víctor Etc.	108	76	10	4	14
Mayo Tanabe Etc.	18	17	1	0	0
RF
Xinke Sui Etc.	30	30	0	0	0
Suyu He Etc.	28	26	2	0	0
Peipei Liu Etc.	27	0	27	0	0
Hai Feng Liu Etc.	90	69	21	0	0
Viswanath Y Etc.	50	50	–	–	0
Lifeng Ma Etc.	86	76	10	0	0
Yue Chang Etc.	50	33	5	9	3
Yan Wang Etc.	18	18	–	0	0
Di Lu Etc.	27	27	–	–	0
Fei Xu Etc.	48	3	37	2	6
Zhi Hua Lan Etc.	40	39	1	0	0
Xin Lu Etc.	25	24	–	1	0
Wei Tao Liang Etc.	60	60	–	–	–

“–” means not mentioned in the study.

ARMI = anti-reflux mucosal intervention, RF = radiofrequency.

3.4. Subgroup analysis

Subgroup analysis of outcome indicators, including GERD-HRQL score, DeMeester score, AET, and PPI use, according to the type of ECSF treatment.

For GERD-HRQL scores, 4 studies that evaluated the impact of gastroesophageal junction ligation on GERD-HRQL scores, significant improvements were observed in GERD-HRQL scores after the procedure, with a MD of 19.69 (95% CI: 14.20–25.19, P < .00001). Seven studies reported that ARMI significantly reduced GERD-HRQL scores, with a MD of 18.16 (95% CI: 13.09–23.22, P < .00001). Additionally, 4 studies reported that RF ablation significantly reduced GERD-HRQL scores in patients with GERD, with a MD of 17.10 (95% CI: 9.81–24.38, P < .00001) (Fig. 7).

Figure 7.

Subgroup analysis on GERD-HRQL scores. GERD-HRQL = gastroesophageal reflux disease health-related quality of life.

For DeMeester score, 6 gastroesophageal junction ligation studies, 8 ARMI studies, and 6 RF ablation studies showed that these 3 procedures significantly reduced DeMeester scores in patients with GERD (ligation group: MD = 79.54, 95% CI: 58.22–100.87, P < .00001; mucosal intervention group: MD = 28.92, 95% CI: 12.05–45.79, P = .0008; and RF group: MD = 26.23, 95% CI: 16.55–35.90, P < .00001) (Fig. 8).

Figure 8.

Subgroup analysis on DeMeester scores.

For AET, all 3 forms of ECSF significantly reduced the AET in patients with rGERD (ligation group: MD = 11.87, 95% CI: 7.18–15.93.83, P < .00001; ARMI group: MD = 9.41, 95% CI: 6.03–12.80, P < .00001 and RF group: MD = 4.52, 95% CI: 3.34–5.70, P < .00001) (Fig. 9).

Figure 9.

Subgroup analysis on the duration of acid exposure.

All 3 forms of ECSF substantially reduced the proportion of patients using postoperative PPIs (ligation group: 35.8% vs 94.5%, RR = 2.85, 95% CI: 1.73–4.62, P < .0001; ARMI group: 35.5% vs 97.3%, RR = 2.63, 95% CI: 2.14–3.22, P < .00001; and RF group: 26.9% vs 99.9%, RR = 3.70, 95% CI: 2.65–5.16, P < .00001) (Fig. 10).

Figure 10.

Subgroup analysis on PPIs usage. PPIs = proton pump inhibitors.

3.5. Sensitivity analysis and publication bias

Sensitivity analyses were performed by removing each study from the meta-analysis, and the results were unchanged. Publication bias for each complication was assessed using funnel plots of the included studies, and visual inspection of the funnel plots showed that they had symmetry, suggesting no serious publication bias (Fig. 4).

4. Discussion

GERD is a common gastrointestinal disorder with a high prevalence in all age groups worldwide that affects the physical and mental health of numerous individuals. PPI is the most frequently used treatment modality; however, a significant number of patients experience ineffective PPI therapy or develop PPI dependence. With the advancement of endoscopic technology and the concerns of LARS causing recurrent postoperative complications, an increasing number of patients with rGERD prefer endoscopic treatment, owing to its less invasive nature.

The ARMI was originally conceptualized by Yeh et al^[10] It mainly includes ARMS and ARMA, which have similar mechanisms of action. ARMS is a technique first proposed by Inoue et al to inhibit acid reflux by removing the mucosa at the gastroesophageal junction through endoscopic mucosal dissection or endoscopic mucosal resection, with tissue contracture forming a scar during the repair process.^[52] Moreover, the ARMA technique is a modification of ARMS, which uses argon plasma coagulation or endoscopic mucosal dissection to ablate the gastric mucosa below the gastroesophageal junction.^[43] The goal is to form a scar around the gastroesophageal junction to create a tissue bulge that prevents reflux from entering the esophagus. A study found that the scar formed after mucosal ablation was more effective in improving reflux than excisional resection.^[42]

Gastroesophageal junction ligation reduced reflux by inducing scar formation. One of the main procedures is the PECC, which was first proposed and developed by Lihu et al^[53] PECC is mainly used to treat gastroesophageal reflux by ligating the mucosa and part of the muscle layer at the gastroesophageal junction and continuously contracting to form a scar, thus forming an anti-reflux barrier. Other procedures include endoscopic band ligation and clip band ligation anti-reflux therapy. Endoscopic band ligation for treating esophageal varices by inducing scarring was serendipitously found by Seleem et al^[27] to be effective in alleviating reflux symptoms in patients with GERD by applying ligatures in up to 4 quadrants at a time to induce scarring, thus reducing acid reflux. Nonetheless, clip band ligation anti-reflux therapy acts as an anti-reflux agent by placing 2 clips at the 6 and 12 o’clock positions of the cardia, where the gastroesophageal junction is constricted by a bulging mass and scar tissue formation; this method serves as an anti-reflux agent.^[24]

RF ablation is one of the first endoscopic treatments with proven efficacy. RF ablation causes coagulative necrosis and fibrosis of the LES and surrounding tissues through energy transfer, scar tissue formation, and increased LES pressure to achieve anti-reflux efficacy.^[54] RF ablation is mainly indicated for adults who are unsuitable for surgical treatment and are unable to take long-term PPI therapy. Additionally, RF ablation is effective in patients with rGERD, reflux hypersensitivity, or functional heartburn.^[55] The American Society of Gastrointestinal Endoscopic Surgeons strongly recommends endoscopic RF ablation for patients with GERD who are effectively treated with PPIs.^[56]

It can be observed that ARMI, gastroesophageal junction ligation and RF ablation have the same anti-reflux mechanism, though they use different instruments and methods. They form scars at the gastroesophageal junction through various means, and the scar tissue continues to contract, thus increasing the pressure on the LES, increasing the non-external force on the cardia and surrounding tissues, and ultimately achieving the therapeutic goal of anti-reflux. Because they have similar principles of action, we will refer to these 3 types of techniques collectively as ECSF.

In addition to the 3 therapies involving ECSF, transoral incisionless fundoplication (TIF) is a common alternative for patients with rGERD. TIF involves folding the gastroesophageal junction using various devices to create a new valve flap to increase LES pressure.^[57] The devices used in TIF include EsphyX, GERDx, and MUSE. However, TIF operates primarily through various types of devices, thereby creating an external force against reflux. Conversely, ECSF forms a scar at the gastroesophageal junction by various means, which contracts and pulls; thus, anti-reflux occurs. The anti-reflux mechanism of ECSF is formed under non-external forces, and because of the different mechanisms of action, we did not include TIF in our study on ECSF.

This study evaluated various treatment outcome indicators in patients with rGERD who underwent ECSF. Our study showed that ECSF was significantly effective in treating rGERD, improving quality of life, alleviating reflux, reducing symptoms such as acid reflux and heartburn, improving esophagitis, significantly decreasing PPIs use after the procedure, with approximately 70% of patients achieving discontinuation of PPIs-like medications, and having a low incidence of overall postoperative adverse effects, with over 80% of patients expressing satisfaction with the treatment; hence, ECSF appears to be a secure and effective therapeutic approach. Among them, all 3 treatment options (gastroesophageal junction ligation, ARMI, and RF) could significantly reduce GERD-HRQL and DeMeester scores, decrease pathological AETs, and substantially reduce the proportion of postoperative PPIs use by patients.

The frequency of severe perioperative adverse events, including perforation, bleeding, and infection, was associated with the operator experience with endoscopic treatment. In the ARMI group, which had the highest incidence of SAEs among the 3 treatment categories, the study by Yang^[38] reported 6 cases of SAEs. One-quarter of the patients experienced varying degrees of dysphagia, and 5 of them did not recover over time. The patients were treated with repeated Savary-Gilliard bougie dilation for esophageal stenosis. Moreover, the study revealed that although 180° and 270° ARMS could be equally effective for rGERD treatment, 180° ARMS might be recommended because of the lower incidence of new postprocedural dysphagia. For Non-SAEs, although sore throats occurred in approximately 46.2% (95% CI: 39.6%–52.9%) of patients after RF ablation, the majority of patients had self-limiting and mild symptoms that resolved on their own without the need for surgery. Particularly, Liu et al^[45] demonstrated that 27 patients had a 100% probability of developing sore throat after RF ablation treatment. In another study,^[50] 37 out of 50 patients with rGERD experienced discomfort, such as sore throat after RF ablation treatment, which had a large impact on the incidence of adverse reactions in the RF group. Moreover, the incidence of dysphagia with ECSF in this study was 15% (95% CI: 14.1–16.0), which is lower than the incidence of dysphagia after fundoplication reported in the previous publications.^[58] Overall, our study suggests that ECSF is a relatively safe class of procedure.

Several studies have also conducted systematic reviews and meta-analyses of endoscopic anti-reflux therapy.^[59] A study^[60] performed a meta-analysis of novel endoscopic treatments (ARMA, ARMS, band ligation (banding), novel full-thickness plication, and resection and plication) and indicated that all new endoscopic therapies had a clinical success rate of 63% (95% CI: 54%–72%) and a rate of adverse events of 8% (95% CI: 3%–14%). ARAM had the highest clinical success rate of 75% (95% CI: 45%–97%). However, procedures such as Nissen and Toupet are performed by artificially applying an external force to achieve a tightened cardia. In this study, we conducted a systematic review and meta-analysis by including 3 types of ECSF to further confirm the efficacy and safety of ECSF in patients with rGERD.

Considering that numerous previous studies have analyzed and compared the efficacy, advantages, and disadvantages of various endoscopic procedures, the significance of this meta-analysis is to analyze and compare the efficacy of 3 endoscopic therapies with similar principles and to make an overall evaluation. Compared to traditional laparoscopic surgical treatment, ECSF is equally reliable for patients with rGERD, with the advantages of a simpler operation, shorter hospitalization time, and fewer side effects. In the group of 16 trials that provided complete operative time data, the mean operative time for ECSF was 37.87 ± 13.20 minutes (95% CI: 30.83–44.90), with a mean operative time of 10.32 ± 2.80 minutes for the 96 patients treated with PECC. In contrast, 200 laparoscopic fundoplication requires a median operative time of 140 minutes,^[61] making ECSF easier to perform.

Moreover, each of the 3 procedures included in the ECSF tended to differ regarding patient selection. PECC is indicated for patients who are resistant to acid-suppressing drugs. ARMI is best suited for patients with extraesophageal GERD with mild to moderate morphological damage to the gastroesophageal junction.^[43] Moreover, RF is effective in patients with reflux hypersensitivity or functional heartburn.^[62] Because these procedures are all endoscopic therapies with similar principles, several other therapies can be used for sequential treatment when a certain procedure is ineffective. ECSF is an effective alternative for patients who have failed laparoscopic surgery and offers more options for patients with rGERD suffering from reflux.

ECSF is associated with a shorter operative time, lower trauma, and fewer adverse effects than surgical interventions. According to previous studies on surgical patient surveys,^[63] 83.2% of patients expressed satisfaction with the outcomes, which is comparable to the 78.4% satisfaction rate reported for the long-term results of laparoscopic Nissen fundoplication. The results of a recent prospective cohort study of nearly 400 surgical procedures showed that^[64] 4.4% of patients experienced serious adverse effects after surgery, much higher than the 1.1%(95% CI: 0.9%–1.2%) incidence of SAEs with ECSF treatment in the present study. Therefore, the ECSF is an effective alternative to surgery.

This study has several advantages, including conducting a methodical search of existing literature to establish specific criteria for selecting relevant studies, thoroughly removing duplicate studies, ensuring the inclusion of high-quality studies from around the world, extracting detailed study data, and thoroughly assessing the quality of studies through a rigorous evaluation process. Furthermore, we employed both objective and subjective measures to gauge the accuracy of the clinical outcomes, thereby enhancing the reliability and generalizability of the results. Studies with small sample sizes were also excluded to reduce bias.

This study has some limitations. First, the majority of studies included in the analysis were cohort studies, and there were relatively few high-quality randomized controlled trial studies. The quality of the included studies varied, and there may have been some bias. Some of the cohort studies used themselves as controls and lacked a control group. Second, some of the outcome indicators, such as reflux status, esophagitis status, and patient satisfaction, were included in a few studies, which may have led to bias in the analysis. Third, some of the included studies had a short follow-up period and insufficient long-term follow-up data; hence, more extended follow-up and data collection are required to verify the long-term efficacy. When accumulating follow-up data for endoscopic treatment, there is also a potential publication bias due to a series of unforeseen confounding factors regarding surgeon experience and expertise, statistical results of patient data, and follow-up. Fourth, there were limited direct comparative studies conducted for these treatments, and additional direct comparative studies are necessary to establish conclusive findings regarding the effectiveness of each endoscopic therapy. Finally, the diagnostic criteria for RGERD used in this study are outdated compared to the latest Lyon Consensus 2.0. Previous diagnostic criteria ignored objective evidence such as endoscopic findings and pH impedance monitoring, and were flawed in terms of accuracy. Moreover although the study participants were all patients with RGERD, the basic profile of the study participants varied among different studies. Some studies^[50] emphasized that the study participants were endoscopy-negative patients, while others^[21] required the inclusion of patients with moderate hiatal hernia of the esophagus, specifying a hernia sac of 3 to 5 cm.

This is the first systematic and comprehensive meta-analysis conducted on ECSF treatment, including gastroesophageal junction ligation, ARMI, and RF ablation. The findings of this study indicate that all 3 ECSF procedures demonstrated significant improvement in reflux and quality of life among patients, likely attributable to their shared mechanisms of action, and the results of the subgroup analysis supported this conclusion. Conclusively, ECSF administration appears to be a safe and effective therapy for patients with refractory GERD. Despite the limitations of the study, this meta-analysis was conducted to understand better the efficacy of 3 endoscopic cardiac peripheral tissue scar formation procedures in treating patients with GERD. This study does not recommend ECSF for all patients with GERD. For patients with PPI-rGERD, ECSF could be a viable option for discontinuing PPI therapy and enhancing the overall quality of life. However further research is required to directly compare its long-term effectiveness and safety with those of surgical interventions.

Author contributions

Conceptualization: Chaoyi Shi, Jun Zhang.

Data curation: Chaoyi Shi.

Formal analysis: Chaoyi Shi, Lina Ying.

Investigation: Liyang Cui.

Methodology: GeSang ZhuoMa, Zhenyu Zhang.

Project administration: Jun Zhang.

Resources: Jun Zhang.

Supervision: Liyang Cui, Ruifang Li.

Validation: Zhenyu Zhang, Ruifang Li.

Visualization: GeSang ZhuoMa, Lina Ying.

Writing – original draft: Chaoyi Shi, Jun Zhang.

Writing – review & editing: Jun Zhang.