Long-Term Efficacy of Repeated Endoscopic Incision Combined with Triamcinolone Injection in Patients with Recurrent Benign Esophageal Stricture

Abstract

Background and Aims

Endoscopic incision (EI) combined with triamcinolone (TAA) injection has emerged as an effective method for treating benign esophageal stricture (BES). However, a single EI and TAA injection seems to be insufficient for long-term relief of BES in some patients, especially those with recurrent BES. Here, we aimed to investigate the long-term efficacy and safety of repeated EIs and TAA injections in patients with recurrent BES.

Methods

Patients with recurrent BES who previously underwent endoscopic dilation or EI were enrolled in this retrospective study. All patients underwent multiple EIs and TAA injections. Patients were followed for at least 12 months. The primary outcome was the rate of re-stricture at 6 months after repeated EIs and TAA injections. The secondary outcomes included the symptom relief rate, the change in Stooler’s dysphagia score (0, normal swallowing; 1, semidry food intake only; 2, soft food intake only; 3, fluid only; and 4, no oral intake) and body mass index (BMI) at 6 months, and overall adverse events (AEs). The risk factors for 6-month re-stricture were identified using logistic regression analysis.

Results

A total of 73 patients were enrolled in this study, including 30 (41.1%) with esophagogastric anastomotic strictures, 15 (20.5%) with caustic injury-related strictures, 14 (19.2%) who developed strictures after ESD, 5 (6.8%) with esophagojejunal anastomotic strictures and 9 (12.3%) with other types of strictures. A total of 264 EIs and TAA injections were performed, with 3 (IQR: 2, 5) treatments for each patient. The 6-month re-stricture rate was 53.4%, and the 2-month, 3-month, and 12-month re-stricture rates were 43.8%, 48.0%, and 54.8%, respectively. The 2-month, 3-month, 6-month, and 12-month dysphagia relief rates were 75.3%, 65.8%, 56.2%, and 54.8%, respectively. At 6 months after treatments, the median Stooler’s dysphagia score had decreased from 4 (IQR: 4, 4) to 2 (IQR: 0, 4) (P < 0.001), and the median BMI had increased from 19.6 (IQR: 16.9, 21.8) kg/m² to 20.8 (IQR: 17.4, 22.6) kg/m² (P < 0.001). Bile reflux (OR 9.90, 95% CI: 0.99–98.43) and the number of EIs (OR 1.68, 95% CI: 1.17–2.41) were independently associated with 6-month re-stricture after repeated EIs and TAA injections. Four EI-related perforations were successfully treated with a covered metal stent (n = 3) and rescue surgery (n = 1).

Conclusions

Repeated EIs and TAA injections were effective and reasonably safe in managing recurrent BES. However, the long-term efficacy remains suboptimal, thus further investigations are needed.

Introduction

Benign esophageal stricture (BES) is still a clinical challenge for endoscopists. The common causes of BES include esophageal anastomotic surgery, endoscopic submucosal dissection (ESD), ingestion of caustic agents (strong acids, bases, and oxidants), and radiation injury [1]. BES is estimated to occur in 0.5 to 16% of patients after esophageal surgery [2], in 56-76% of patients after ESD for large-area early esophageal cancer, and in as many as 100% of patients with total circumferential disease [3, 4]. Dysphagia is the most common clinical symptom of BES, which diminishes quality of life and requires active treatment [5]. Traditional treatments for BES include balloon dilation (BD), bougienage, oral steroids, and submucosal injection of triamcinolone (TAA); however, only 60–70% of patients achieve long-term stricture relief [6], and even fewer patients with recurrent or refractory BES experience relief, making clinical treatment more challenging [7, 8].

Endoscopic incision (EI) has been recently recognized as a promising second-line treatment for BES [9–11]. Multiple studies have revealed that, compared with balloon dilation, EI has a lower stricture recurrence rate and a longer esophageal patency time [12, 13]. However, a single EI often cannot provide long-term relief of BES, and 13–58% of patients may require further intervention [14–16]. Previous studies revealed that repeated EIs may have better outcomes than a single EI does [12, 16]; however, the number of cases involving multiple EIs is limited, and the long-term effects of such treatment need further evaluation [16]. Additionally, the therapeutic effects of EIs on BESs of different etiologies vary [14–16]. Therefore, the outcomes of repeated EIs for BESs of various etiologies are worthy of further investigation.

In this study involving a relatively large sample of patients with recurrent BES, we explored the occurrence of re-stricture after repeated EIs and local injections of TAA, and investigated the risk factors for re-stricture.

Methods

Patients

This was a retrospective, single-center study conducted at the Air Force Medical Center. Seventy-three patients with recurrent BESs were treated with EIs and TAA injections between February 2019 and February 2023 at our institution. Although this was a retrospective study, our center prospectively collected treatment data and conducted active prospective follow-up of all consecutive patients with BES (including recurrent cases) to minimize missing critical data and patient loss to follow-up. Initial therapies (BD or EI) at local hospitals failed for most of the enrolled patients, resulting in a cohort of patients with complex or treatment-resistant cases. The inclusion criteria were as follows: ① Patients with recurrent or refractory BES who had previously undergone at least one endoscopic dilation or EI treatment procedure. Recurrent BES was defined as a stricture that recurred within 4 weeks after initial therapy. Refractory BES was defined as a stricture that underwent at least five sessions of treatment (interval of 2 weeks) but stricture retained [8, 17, 18]. ② followed for at least 12 months. The exclusion criteria were as follows: ① patients who had malignant esophageal strictures; ③ had active ulcers or fistulas; ④ had significant abnormalities in coagulation function or unstable hemodynamics.

EI Procedure

All patients underwent gastroscopy (Fig. 1a–c) and X-ray fluoroscopy (Fig. 1d) to evaluate the location, length, and diameter of the BES. Mediastinal computed tomography (CT) (Fig. 1e) or endoscopic ultrasound (Fig. 1f) was performed to determine the thickness of the scar tissue if necessary. The EIs were made under general anesthesia. First, a transparent cap (DH-28GR, FUJIFILM, Japan) was installed at the front end of the gastroscope (EG-600 WR, FUJIFILM, Japan), and the thickest side of the scar tissue, which is often stiffer and appears as a lighter color, was identified. Second, the scar tissue was cut with a Hook knife (KD-620UR, Olympus, Japan) from shallow to deep until the superficial muscle layer was exposed. Third, the Hook knife was inserted between the fibers of the scar tissue and the muscle to cut off the scar toward the center of the lumen (Fig. 2a). While incising the scar, the gastroscope was advanced until it passes through the stricture segment without resistance. (Fig. 2b, c). If the endoscope cannot pass through after the incision is made in the 1–2 direction, a radial incision method can be adopted (Fig. 2d). If the BES is severe and the lumen is poorly displayed, a guidewire (ACRO-35–480, COOK, USA) can be inserted through the stricture to better guide the direction of the incisions (Fig. 2e). The scar tissue and anastomotic nails could be removed during the incision. After the EI, all patients received a multipoint submucosal injection of TAA (Kunming Jida Pharmaceutical, China) through an injection needle (WS-2423PN, WILSON, China) (Fig. 2f). The dose varied depending on the length of the stricture (the effective injection dose was approximately 20 mg/cm).

Fig. 1.

Examination before treatment of esophageal stricture: a Reflux inflammatory esophageal stricture; b Esophageal stricture with active esophageal ulcer; c Fibrous stricture of the lower segment of the esophagus after cure of reflux esophagitis; d Esophagography showed esophageal stricture with dilation of proximal esophagus; e Mediastinal CT revealed uneven thickening of the esophageal wall; f Endoscopic ultrasonography showed uneven thickening of the esophageal wall

Fig. 2.

a The Hook knife was inserted between the scar fiber tissue and the muscle to cut off the scar; b Gastroscope could pass through the stricture after incision in one direction; c Gastroscope could pass through the stricture after incision in two direction; d Radial EI; e Severe stricture be incised with the guidance of a guide; f The submucosa surrounding the wound was given injection of TAA after EI

After the EI and the TAA injection, the wound was carefully examined to determine whether there was perforation or bleeding. For patients at risk for perforation (muscularis propria injury), clips were used and a gastric tube was implanted, and negative pressure aspiration therapy was performed. Esophageal covered metal stents (CZES-II, Huaian Sigma Industry Co. Ltd.) or clips (ROCC-D-26–230, Nanjing Micro-Tech Co. Ltd.) were used if obvious perforation or uncontrollable bleeding occurred.

The patients were closely observed and administered acid inhibitors after treatment, and antibiotics were given if the patients were suspected of having an intraoperative perforation. Fluid intake was allowed 24 h after treatment if the patients did not exhibit any signs of discomfort.

All patients underwent endoscopic evaluation every 4 weeks thereafter. If the endoscope was unable to pass through the stricture segment, we scheduled another EI and TAA injection. Otherwise, if no further treatment was required, we initiated the follow-up phase.

Outcomes and Data Collection

Outcomes

The primary outcome of our study was the re-stricture rate at 6 months, which was defined as the proportion of patients in whom the endoscope was unable to pass through the segment of stricture. The secondary outcomes included the following: ① re-stricture rates at 2 months, 3 months, and 12 months after treatment; ② dysphagia relief rates at 2 months, 3 months, 6 months, and 12 months after treatment; ③ Stooler’s dysphagia score and body mass index (BMI) at 6 months after treatment; ④ incidence of adverse events (AEs). Dysphagia severity was presented as a Stooler’s dysphagia score [19]: 0, normal swallowing; 1, semidry food only; 2, soft food only; 3, fluids only; and 4, no oral intake. Dysphagia relief was defined as a decrease in Stooler’s dysphagia score, which was evaluated at 2, 3, 6, and 12 months after the last treatment. AEs included bleeding that required blood transfusion and perforation that required stent insertion or surgical intervention, as described previously [16].

Data Collection

Demographic data, BES etiology, and prior BES treatment were retrospectively collected from electronic medical records. The following BES-related parameters, including the location, aperture, and length of the stricture, and the and presence of bile reflux, ulcers, or fistulas, were routinely recorded in our center. Among them, bile reflux is characterized by the presence of yellowish bile deposition at both the stricture site and within the gastric/jejunum lumen distal to the successfully traversed stricture during endoscopic examination. Follow-up data were collected until March 2024 or the date of death/loss to follow-up.

We meticulously monitored the status of the esophageal stricture and the presence of dysphagia at 2, 3, 6, and 12 months after the final EI session, utilizing methods such as gastroscopy, X-ray fluoroscopy, or mediastinal CT scans. Patients were strongly encouraged to reach out to the investigators (H.Y.C. or C.X.F.) at the first sign of dysphagia to ensure prompt attention and care.

Statistical Analysis

Parametric data are presented as the mean ± standard deviation (SD), median (IQR), or number (%). Categorical va riables were compared using the chi-square test or Fisher’s exact test, whereas continuous variables were compared using the Student’s t test or the Mann‒Whitney U test, depending on whether the data had normal or nonnormal distributions. The cumulative re-stricture rate was analyzed using the Kaplan‒Meier method and compared using the log-rank test. Univariate logistic regression was used to identify the factors influencing efficacy, and multivariate analysis included significant variables from the univariate analyses conducted using SPSS version 24.0 (SPSS Inc., Chicago, Illinois). Correlation analyses were performed via Spearman’s rho. Two-sided P values ≤ 0.05 indicated statistical significance.

Results

Patient Characteristics

Initially, 163 patients with recurrent BES were screened. After 14 patients with malignancy, 18 with active ulcers/fistulas, and 3 with unstable hemodynamics were excluded, 128 patients were deemed eligible for EI and TAA injection. Among them, 55 were excluded because of an insufficient follow-up duration (< 12 months), resulting in a final cohort of 73 patients. Among these patients, 54 experienced balloon dilation failure, and 19 experienced EI failure.

The patients’ characteristics are shown in Table 1. There were 55 males and 18 females, and the median age was 62 years. The causes of recurrent BES were categorized as follows: esophagogastric anastomosis (30, 41.1%), caustic injury (15, 20.6%), ESD (14, 19.2%), esophagojejunal anastomosis (5, 6.8%) and other rare causes (Such as spontaneous rupture of the esophagus, reflux esophagitis, esophageal injury in the neck, and chest surgery, etc) (9, 12.3%). There were 24 (32.9%), 16 (21.9%), 28 (38.4%), and 5 (6.8%) strictures in the upper, middle, lower, and multiple segments of the esophagus, respectively. Ten (13.7%) of the 73 patients had a history of fistula, and 3 (4.1%) had undergone radiotherapy at the stricture segment. Bile reflux was found endoscopically in 12 (16.4%) of the 73 patients. The median stricture length was 3.0 (IQR: 2.0, 5.0) cm, the median stricture diameter was 2.0 (IQR: 1.0, 3.0) mm, the median Stooler’s dysphagia score was 4 (IQR: 4, 4), and the median number of previous treatments was 1 (IQR: 1, 3).

Table 1.

Baseline characteristics of the patients (n = 73)

Category		N (%)/M (IQR)
Age (year)		62 (52, 67)
Male		55 (75.3%)
Etiology of stricture	Esophagogastric anastomosis	30 (41.1%)
	Caustic injury	15 (20.6%)
	ESD	14 (19.2%)
	Esophagojejunal anastomosis	5 (6.8%)
	Other types*	9 (12.3%)
Location of strictures	Upper	24 (32.9%)
	Middle	16 (21.9%)
	Lower	28 (38.4%)
	Multi	5 (6.8%)
Length of strictures (cm)		3.0 (2.0, 5.0)
Diameter of strictures (mm)		2.0 (1.0, 3.0)
Previous fistula		10 (13.7%)
Previous radiotherapy		3 (4.1%)
Concurrent bile reflux#		12 (16.4%)
Stooler’s dysphagia score before treatment		4 (4, 4)
Number of previous endoscopic treatments		1 (1, 3)

Values were presented as number (%) or median (interquartile range[IQR]); N number, M Median, ESD Endoscopic Submucosal Dissection

^*: Such as spontaneous rupture of the esophagus, reflux esophagitis, esophageal injury in the neck, and chest surgery, etc

^#:based on previous endoscopic examination

Clinical Outcomes and AEs

In the 73 patients, a total of 264 EIs and TAA injections were performed. The median cost per hospitalization was CNY 20,135 (IQR: 17,421, 25,384) (USD 2,960 (IQR: 2,562, 3,733). The exchange rate refers to the official central parity rate of the Bank of China in January 2024: 1 USD = 6.8 CNY), with a median of 3 sessions (IQR: 2, 5) per patient. The number of treatment sessions varied by etiology: esophagogastric anastomotic strictures required 2.5 sessions (IQR: 2, 4), caustic injury-related strictures required 4 sessions (IQR: 2, 6), post-ESD strictures required 3.5 sessions (IQR: 3, 5), esophagojejunal anastomotic strictures required 4 sessions (IQR: 2, 6), and strictures with other causes required 2 sessions (IQR: 2, 3). The median therapy duration was 30.0 (IQR: 21.4, 36.8) minutes, and the median follow-up duration was 18 months (IQR: 13, 27). During the EI, a guidewire was inserted to guide the incision direction in 60 (82.2%) patients. Furthermore, 6 (8.2%), 3 (4.1%), and 10 (13.7%) patients underwent clipping and placement of a metal stent or a gastric tube for the prevention of complications, respectively. Among all 274 surgeries, 167 (63.3%), 7 (2.65%), 3 (1.1%), and 11 (4.2%) involved the insertion or placement of guidewires, clips, metal stents, and gastric tubes, respectively (Table 2).

Table 2.

Parameters related to repeated EIs and TAA injections in BES patients

Combination treatment	Patients, N (%)	EIs, N (%)
Wire-guided EI	60 (82.2%)	167 (63.3%)
TAA injection	73 (100.0%)	264 (100.0%)
Gastric tube placement	10 (13.7%)	11 (4.2%)
Application of Clips	6 (8.2%)	7 (2.7%)
Covered metal stent placement	3 (4.1%)	3 (1.1%)

Values were presented as number (%); N Number, EI Endoscopic Incision, TAA Triamcinolone Acetonide

The 6-month re-stricture rate was 53.4% for all patients, and the 6-month re-stricture rates were 53.3%, 40.0%, 50.0%, 60.0%, and 77.8% for patients with strictures caused by esophagogastric anastomosis, caustic injury, ESD, esophagojejunal anastomosis, and other causes, respectively.

The 2-month, 3-month, and 12-month re-stricture rates were 43.8%, 48.0%, and 54.8%, respectively. The dysphagia relief rates at 2 months, 3 months, 6 months, and 12 months after treatment were 75.3%, 65.8%, 56.2%, and 54.8%, respectively. The re-stricture rates and dysphagia relief rates were not significantly different among the groups (Fig. 3). The 1-year recurrence-free survival (RFS) curve analysis also revealed no significant difference (Fig. 4).

Fig. 3.

Re-stricture rates and dysphagia relief rates of different types of recurrent BES. a Re-stricture rates among different types of recurrent BES. b Dysphagia relief rates among different types of recurrent BES

Fig. 4.

Re-stricture rates after EIs and TAA injections. a Re-stricture rates of all patients; b Re-stricture rates among different types of recurrent BES

For all patients, the median Stooler’s dysphagia score at 6 months after repeated EIs had decreased by − 1 (IQR: − 4, 0) (P < 0.001), − 1.5 (IQR: − 4, 0), − 3 (IQR: − 4, 0), − 1 (IQR: − 4, 0), 0 (IQR: − 3, 0), and 0 (IQR: − 1, 0) in patients with BES caused by esophagogastric anastomosis, caustic injury, prior ESD, esophagojejunal anastomosis and other types, respectively. Moreover, the median BMI at 6 months after treatment had increased by 0.7 (IQR: 0.02, 2.3) kg/m² (P < 0.001), 0.7 (IQR: − 0.009, 1.8), 1.7 (IQR: 0.02, 3.7), 1.2 (IQR: 0.6, 2.2), 0.3 (IQR: − 1.4, 0.5), and 0.8 (IQR: 0.04, 2.0) in patients with BES caused by esophagogastric anastomosis, caustic injury, prior ESD, esophagojejunal anastomosis and other types, respectively (Table 3).

Table 3.

Stooler’s dysphagia score and BMI of patients with different types of recurrent BES before and 6 month after repeated EIs and TAA injections

	EI Sessions	Stooler’s dysphagia score				BMI
Types of stricture	M (IQR)	Before EI, M (IQR)	Six-month after EI, M (IQR)	Stooler score reduction, M (IQR)	P value	Before EI, M (IQR) kg/m2	Six-month after EI, M (IQR) kg/m2	BMI difference, M (IQR) kg/m2	P value
All patients	3 (2, 5)	4 (4, 4)	2 (0, 4)	− 1 (− 4, 0)	< 0.001	19.6 (16.9, 21.8)	20.8 (17.4, 22.6)	0.7 (0.02, 2.3)	< 0.001
Esophagogastric anastomosis (n = 30)	2.5 (2, 4)	4 (4, 4)	2.5 (0,4)	− 1.5 (− 4, 0)	< 0.001	20.1 (17.6, 22.1)	20.7 (17.2, 22.5)	0.7 (− 0.009, 1.8)	0.12
Caustic injury (n = 15)	4 (2, 6)	4 (4, 4)	1 (0, 4)	− 3 (− 4, 0)	0.001	18.8 (15.6, 20.2)	21.2 (18.4, 23.2)	1.7 (0.02, 3.7)	0.02
ESD (n = 14)	3.5 (3, 5)	4 (3, 4)	1.5 (0, 4)	− 1 (− 4, 0)	0.04	21.7 (18.8, 25.2)	22.8 (20.8, 26.4)	1.2 (0.6, 2.2)	0.004
Esophagojejunal anastomosis (n = 5)	4 (2, 6)	4 (3, 4)	3 (0, 4)	0 (− 3, 0)	0.5	17.2 (16.5, 20.3)	17.4 (15.9, 20.3)	0.3 (− 1.4, 0.5)	1
Other types* (n = 9)	2 (2, 3)	4 (4, 4)	4 (3, 4)	0 (− 1, 0)	0.38	16.0 (15.0, 20.8)	18.4 (17.2, 21.6)	0.8 (0.04, 2.0)	0.16

Values were presented as median (interquartile range[IQR]); M median, EI Endoscopic Incision, ESD Endoscopic Submucosal Dissection, BMI Body Mass Index; *: Such as spontaneous rupture of the esophagus, reflux esophagitis, esophageal injury in the neck, and chest surgery, etc.

Univariate analysis (Table 4) revealed that bile reflux (P = 0.004), the time for first EI therapy (P = 0.046), the number of EIs (P < 0.001), and wire-guided EI (P = 0.002) influenced the occurrence of 6-month re-stricture. Furthermore, multivariate analysis revealed that bile reflux (P = 0.05) and the number of EIs (P= 0.005) were independent predictive factors for clinical resolution.

Table 4.

Univariate and multivariate analysis of risk factors for 6-month re-stricture after repeated EIs and TAA injections

		6-month re-stricture		Univariate analysis		Multivariate analysis
Category		No (n = 34)	Yes (n = 39)	OR (95%CI)	P value	OR (95%CI)	P value
Age (year)		59.5 (52.0, 67.0)	62.0 (52.0, 67.0)	1.00 (0.98, 1.03)	0.79
Sex	Female	9 (26.5%)	9 (23.1%)	0.83 (0.29, 2.42)	0.74
Types of stricture	Esophagogastric anastomosis	14 (41.2%)	16 (41.0%)	Reference	0.5
	Caustic injury	9 (26.5%)	6 (15.4%)	0.58 (0.17, 2.05)
	ESD	7 (20.6%)	7 (18.0%)	0.88 (0.25, 3.12)
	Esophagojejunal anastomosis	2 (5.9%)	3 (7.7%)	1.31 (0.19, 9.02)
	Other types*	2 (5.9%)	7 (18.0%)	3.06 (0.54, 17.23)
Location of strictures	Upper	9 (26.5%)	15 (38.5%)	Reference	0.62
	Middle	7 (20.6%)	9 (23.1%)	0.77 (0.21, 2.80)
	Lower	15 (44.1%)	13 (33.3%)	0.52 (0.17, 1.58)
	Multi	3 (8.8%)	2 (5.1%)	0.40 (0.06, 2.87)
Previous history
Perforation or fistula		3 (8.8%)	7 (18.0%)	2.26 (0.54, 9.54)	0.26
Radiotherapy		1 (2.9%)	2 (5.1%)	1.78 (0.16, 20.58)	0.64
Bile reflux		1 (2.9%)	11 (28.2%)	12.96 (1.58, 106.73)	0.004	9.90 (0.99, 98.43)	0.05
Length of strictures (cm)		2.5 (2, 5)	3 (2, 5)	0.99 (0.86, 1.13)	0.98
Diameter of strictures (mm)		2 (2, 3)	2 (1, 3)	0.74 (0.53, 1.03)	0.07
Time for first EI therapy(min)		25.0 (20.0, 45.0)	33.0 (23.0, 50.0)	1.02 (0.99, 1.05)	0.046	1.01 (0.98, 1.04)	0.42
Number of EI		2 (1, 3)	4 (3, 6)	1.85 (1.30, 2.62)	< 0.001	1.68 (1.17, 2.41)	0.005
Wire-guided EI		23 (67.6%)	37 (94.9%)	8.84 (1.80, 43.54)	0.002	4.38 (0.68, 28.07)	0.12
Gastric tube placement		2 (5.9%)	8 (20.5%)	4.13 (0.81, 21.00)	0.07
Application of Clips		2 (5.9%)	4 (10.3%)	1.83 (0.31, 10.67)	0.5
Covered metal stent placement		0 (0.0%)	3 (7.7%)	–	0.1

Values were presented as number (%) or median (IQR)

CI Confidence Interval, ESD Endoscopic Submucosal Dissection, EI Endoscopic Incision

^*: Such as spontaneous rupture of the esophagus, reflux esophagitis, esophageal injury in the neck, and chest surgery, etc.

There were 4 EI-related perforations. One perforation occurred during the first EI, two occurred during the fifth EI, and one occurred during the sixth EI. Three patients (the patients with esophagogastric anastomosis strictures) were successfully treated with a covered metal stent, and one patient (the patient who developed a stricture after ESD) was successfully treated with rescue surgery. One patient underwent rescue surgery because the gastric tube caused a perforation at the muscular injury site during insertion, which was not detected in time and resulted in empyema. None of the patients experienced delayed bleeding.

Discussion

EI have been recognized as an alternative method against BD for BES, and its efficacy for recurrent or refractory strictures has been reported [11]. However, a single EI may not provide long-term relief, and repeated EIs may lead to better outcomes than a single EI will. A concurrent TAA injection at the wound site can suppress the growth of postoperative fibrous scar tissue, thereby establishing a therapeutic window for neomucosa to regenerate and consequently cover the wound [20]. In our study, different types of BESs were classified, and the follow-up time was long. Nearly half of the patients (46.6%) remained recurrence-free for at least 6 months after repeated EIs and TAA injections. This study revealed that EI and TAA injection was safe, with a low incidence of AEs. Owing to the reduced vascularity in fibrotic strictures, intraprocedural bleeding is typically mild and self-limiting, and the main AE is perforation [21, 22]. Most perforations occur after multiple treatments and can be remedied endoscopically if identified in time. In addition, our study revealed that bile reflux was an independent risk factor for EI efficacy.

At present, there are few studies on repeated EIs as a means of therapy. Ma described 12 patients with refractory esophageal anastomotic stricture who were treated with a total of 38 EIs; the median number of EIs was 2.5 (IQR: 1, 9), and the success rate was 83% [23]. No complications occurred during the treatment or follow-up period. Another study investigating the value of EIs for refractory BESs revealed that the 6-month stricture-free rate after the first EI was 42.3%. However, in the repeated EIs group, the 6-month refractory stricture-free rates after the first and repeated EIs were 18.2% and 18.2%, respectively. No AEs were noted [16]. Our study revealed that after repeated EIs and TAA injections, the 6-month stricture-free rate was 46.6%. Although patients with recurrent BES were not completely differentiated from those with refractory BES, some patients were referred to our hospital for treatment with EIs because difficulty associated with esophageal fistula and severe reflux, as well as factors such as angulation and more severe and long strictures.

Patient-centered outcomes have been rarely reported in previous studies, raising substantial concerns. In one study, the authors reported that 88% of patients with refractory BES experienced dysphagia relief within 4 weeks after treatment with EIs [14], and in this study, 75.3% of patients experienced dysphagia relief within 2 months after treatment.

Previous studies focused mainly on postoperative anastomotic strictures [12, 24, 25] and have rarely involved other types of strictures. Our study revealed heterogeneity in treatment response across etiologies. For example, caustic injury-related strictures required more EIs (median: 4) than esophagogastric anastomotic strictures did (median: 2.5), possibly due to deeper tissue fibrosis and greater inflammatory activity in caustic cases. Similarly, patients with esophagojejunal anastomotic strictures had minimally improved Stooler’s dysphagia scores, likely because of anatomical complexity and bile reflux in such patients. These findings underscore the need for etiology-specific treatment protocols.

Our study revealed that bile reflux was an independent risk factor for stricture recurrence. The exact mechanisms are not understood. Studies revealed that inflammation, fibrosis, and scarring of the submucosa play important roles in stricture recurrence. Bile reflux may cause esophageal mucosal necrosis, inflammatory cell infiltration, mucosal erosion, and ulcer formation, which stimulate fibrosis during the healing stage and lead to stronger and more complex strictures that are refractory to further treatment [26]. Bile reflux was first identified as a new risk factor for decreased efficacy of repeated EIs and TAA injections, but the broad CI for bile reflux (OR = 9.90, 95% CI: 0.99–98.43) reflected uncertainty in the estimate, likely due to the small sample size, particularly in the low incidence of bile reflux events (16.4%, 12/73) in our cohort, which may have inflated the OR and reduced precision. While the P value met the conventional threshold for significance, we acknowledge that the wide CI requires careful clinical interpretation. Therefore, the value of bile reflux in predicting the efficacy of repeated EIs and TAA injection should be further investigated in larger or multicenter studies to improve stability.

We have observed that most BESs recur within 1–2 months, which suggests that EIs should be repeated 1 month after the previous treatment. In theory, maintaining the patency of the esophageal lumen can provide space for the mucosal epithelium to cover the wound, which is beneficial for stricture treatment [27], and whether regular repeated EIs and TAA injections can help reduce the rate of re-stricture still needs further exploration.

The depth of the incision during treatment is the key to success. If the incision is too shallow, the scar release is not complete, which affects the therapeutic effect. If the incision is too deep, AEs, such as perforation and bleeding, may occur. Therefore, this operation has certain technical difficulties and risks, suggesting that endoscopists must be skilled [28].

For patients whose strictures recur after repeated EIs and TAA injections, it may be necessary to consider an in vitro self-dilating balloon or endoscopic stent placement. In addition, new methods, such as tissue transplantation, tissue engineering, and the use of new biomaterials, have been explored.

The current study has several limitations. First, this was a small-scale retrospective single-center study that included only patients with BES during a specific period (between February 2019 and February 2023), and those lost to long-term follow-up were excluded, which inevitably introduces selection bias and limits the reliability of the results. Second, our study was conducted in a tertiary care center, which may introduce referral bias. Patients who are referred to our institution often present with recurrent or refractory strictures that have failed to respond to standard therapies, which typically increases the risk of re-stricture. Consequently, the observed re-stricture rate (53.4% at 6 months) and efficacy of repeated EIs and TAA injections may not reflect the outcomes of less severe cases or cases treated in nonspecialized settings. Future multicenter studies involving both community hospitals and tertiary centers are warranted to validate these findings across diverse patient populations. Third, while our study included patients with recurrent BESs, the distinction between “recurrent” and “refractory” strictures was incomplete, precluding proper stratification and analysis. Because refractory cases were not included in the exclusion criteria, the heterogeneity in stricture etiology and treatment history may have influenced the observed outcomes. In future studies, researchers should use standard definitions for refractory BES and evaluate the efficacy of repeated EIs in this subgroup separately. Fourth, the study design precluded the isolation of the individual effects of EI versus TAA injection. Future randomized controlled trials comparing EI monotherapy, TAA monotherapy, and their combination are needed to determine the relative contributions of each modality. Additionally, the occurrence, consequences and preventability of adverse events cannot be ascertained in a retrospective analysis.

In conclusion, repeated EIs and TAA injections were effective and reasonably safe in managing recurrent BESs in a tertiary care setting. However, the generalizability of these results to nonreferral populations or less complex cases requires further validation. We advocate for larger-scale prospective studies to establish standardized protocols applicable across different healthcare environments.

Author Contributions

Study conceptualization and design: S.B.N. Y.L.P., and C.X.F. Drafting of the manuscript: H.Y.C., Y.L., and M.N.X. Acquisition of data: R.G. and J.Z. Analysis or interpretation of data: C.X.F., M.H.L., and X.P.Z. Preparation of tables: B.R.L. and T.S. Preparation of figures: H.Y.C. and C.X.F. Critical revision of the manuscript: S.B.N. and Y.L.P. Administrative and material support: S.B.N.

Funding

This research was supported by grants from the Capital Medical Development Research Foundation of China (Z221100007422061), the National Key R&D Program of China (2022YFC2505100), and the Science and Technology Development Fund of Air Force Military Medical University (2021LC2201). The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.

Data Availability

Additional supporting data are available from the corresponding authors upon request. All requests for raw and analyzed data and materials will be reviewed by the corresponding authors to verify whether the request is subject to any intellectual property or confidentiality obligations.

Declarations

Conflict of interest

Each author declares that he or she has no commercial associations (e.g. consultancies, stock ownership, equity interest, patent/licensing arrangement, etc. from financial support, other institutions, or organizations) that might pose a conflict of interest in connection with the submitted article.

Ethical approval

This retrospective study involving human participants was conducted in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards, with approval from the Institutional Review Board (IRB) at the Air Force Special Medical Center Ethics Committee (2023–89-YJ01).