Esophageal stenting and endoscopic vacuum therapy for esophageal defects: a systematic review and meta-analysis of observational studies

Abstract

Background:

Spontaneous esophageal perforation, especially Boerhaave syndrome, carries high morbidity and mortality. Minimally invasive treatments like esophageal stenting and endoscopic vacuum therapy (EVT) are increasingly used, but optimal management remains unclear.

Objective:

This systematic review and meta-analysis evaluates the efficacy and safety of esophageal stenting and EVT in managing esophageal defects by assessing sealing rates, failure rates, and mortality.

Methods:

A comprehensive literature search was conducted in PubMed, Scopus, and the Cochrane Library through 24 March 2025. Only observational studies and case series were included. Primary outcomes were the pooled sealing rate, failure rate, and mortality for stenting, and the closure rate for EVT. Data were analyzed using a random-effects model, with heterogeneity assessed by the I² statistic.

Results:

Twenty stenting studies (245 patients) showed a pooled sealing rate of 86.1% (95% CI: 80.2–92.0%) and a failure rate of 14.9% (95% CI: 8.5–21.3%). Stent-related mortality was 7.4% (95% CI: 3.5–11.4%). Thirteen EVT studies (424 patients) demonstrated a pooled sealing rate of 54.1% (95% CI: 34.8–73.4%) with high heterogeneity (I² = 98.24%). After sensitivity analysis, the EVT sealing rate rose to 89.6% (95% CI: 83.9–95.3%).

Conclusion:

Both esophageal stenting (86.1% success) and EVT (54.1% overall; 89.6% post-sensitivity) effectively close esophageal defects. However, EVT’s heterogeneity and the observational design of studies limit definitive conclusions. These findings affirm stenting’s role and suggest EVT’s promise, yet emphasize the urgent need for randomized controlled trials to establish evidence-based guidelines.

Introduction

Spontaneous esophageal perforations, such as Boerhaave syndrome, result from a sudden full-thickness tear in an otherwise healthy esophagus – most often triggered by forceful vomiting – leading to rapid contamination of the mediastinum. In contrast, other esophageal defects arise secondary to identifiable causes: iatrogenic perforations occur during endoscopic or surgical interventions, anastomotic leaks follow esophageal surgery, and ulcer-related fistulas develop over time from chronic inflammation. Although Mackler’s triad of vomiting, chest pain, and subcutaneous emphysema is considered characteristic, nearly one-third of cases exhibit atypical clinical features that can delay diagnosis^[1]. The rupture is most often localized to the left side of the lower intrathoracic esophagus^[2] and without prompt recognition and intervention, mortality rates can range from 15% to 42%^[3]. Conventional treatment strategies involve emergency surgical procedures – such as primary repair, T-tube insertion, or esophageal resection and diversion – with early surgery correlating with improved outcomes^[4–7].

HIGHLIGHTS

• Esophageal defects, including leaks and perforations, cause high morbidity and mortality. Minimally invasive options like stenting and EVT are alternatives.

• Twenty studies on stenting, involving 188 patients, revealed an 86.1% sealing success rate, 14.9% failure, and 7.4% mortality overall outcomes.

• Thirteen studies on EVT, with 413 patients, initially showed a 54.1% sealing rate; sensitivity analysis improved outcomes to 89.6% sealing.

• Both stenting and EVT are viable. However, EVT outcomes vary, improving under standardized conditions, and further prospective studies are warranted.

Esophageal leaks, perforations, and Boerhaave syndrome continue to pose significant clinical challenges given their high morbidity and mortality. While invasive surgical methods have traditionally been the mainstay of treatment, there has been a notable shift over the past two decades towards minimally invasive techniques. Approaches such as esophageal stenting and endoscopic vacuum therapy (EVT) have emerged as promising alternatives, offering the benefits of re-establishing gastrointestinal continuity, facilitating the drainage of infected areas, and optimizing resuscitative efforts. These interventions may reduce the duration and invasiveness of traditional surgeries and enhance patient recovery. In fact, a comprehensive review that analyzed 66 studies on the use of esophageal stents for anastomotic leaks and benign perforations reported a technical success rate of 96% and a clinical success rate of 87%^[7,8].

Esophageal stenting, in particular, provides a less invasive option with the potential for rapid defect closure and lower procedural morbidity, while EVT has attracted attention for its effectiveness in promoting tissue healing and controlling sepsis in complex esophageal defects^[1,9]. Despite their increasing utilization, the optimal treatment strategy remains debatable due to the variability in reported outcomes and the current lack of randomized controlled trials (RCTs)^[10].

The existing evidence is primarily derived from observational studies and case series, which are subject to inherent biases and confounding factors. Although recent reviews have contributed valuable insights into endoscopic stenting, they have generally not included EVT in their analyses^[7,11]. More recent pivotal studies by Wannhoff et al^[12] and Anundsen et al^[13] have begun to integrate EVT data, offering further perspective on managing Boerhaave syndrome. This structured review is designed to systematically evaluate and synthesize the current literature, addressing the variability in clinical outcomes and establishing a robust foundation for decision-making in this high-risk patient population.

The primary objective of this systematic review and meta-analysis is to assess the sealing efficacy, failure rates, and mortality associated with esophageal stenting and EVT in patients with esophageal leaks, perforations, and particularly Boerhaave syndrome. Specifically, we aim to: (1) quantify the pooled sealing rate, failure rate, and mortality rate associated with esophageal stenting from various observational studies; (2) evaluate the efficacy of EVT in achieving defect closure while exploring the sources of heterogeneity and potential publication bias in the existing literature; and (3) compare the clinical outcomes of esophageal stenting and EVT to inform clinical practice and pinpoint priorities for future research.

Methods

Study design and literature search

We performed a systematic review and meta-analysis to evaluate the efficacy and safety of esophageal stenting and EVT for various esophageal conditions, including anastomotic leaks, iatrogenic perforations, and Boerhaave syndrome. In light of the absence of RCTs in this field, our analysis was restricted to observational studies and case series. This review was conducted following the PRISMA 2020 guidelines (see flowchart in Fig. 1), registered in PROSPERO, and the methodological quality was further ensured by adherence to the AMSTAR2 criteria^[14]. A comprehensive literature search was carried out across PubMed/MEDLINE, Scopus, and the Cochrane Library from their inception until 22 March 2025, using a combination of keywords and Medical Subject Headings (MeSH) including “esophageal stenting,” “endoscopic vacuum therapy,” “esophageal leak,” and “esophageal perforation.” No language restrictions were applied, and additional studies were identified through manual screening of reference lists. A detailed search strategy is available from Supplementary Digital Content files S1, available at: http://links.lww.com/MS9/A887. A PRISMA checklist is added as a supplement as well. The study has been registered in the international prospective register of systematic reviews (PROSPERO).

Figure 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA flowchart).

Eligibility criteria and study selection

Studies were eligible for inclusion if they reported on the use of esophageal stents or EVT in patients with esophageal conditions. Given the lack of RCTs, we included prospective and retrospective observational studies as well as case series. For stenting studies, outcomes of interest included sealing rates, failure rates, and mortality. EVT studies were required to report sealing success, treatment duration, and sponge change frequency. Two independent reviewers screened titles, abstracts, and full texts to determine eligibility, resolving any discrepancies through consensus. Study quality and risk of bias were assessed using the MINORS (Methodological Index for Non-Randomized Studies) scale^[15] (available in Tables 1 and 3). This validated scoring system evaluates key aspects of study design, such as clearly stated aims, appropriate endpoints, follow-up adequacy, and data collection methods. Each criterion is rated from 0 to 2, with a maximum score of 16 for noncomparative studies and 24 for comparative studies. Higher scores indicate better methodological rigor. A separate table detailing the MINORS score for each study is available in Supplementary Digital Content files S1, available at: http://links.lww.com/MS9/A887.

Table 1.

Summary of studies on esophageal stenting for esophageal conditions

Year	Author et al	Design	MINORS score	Stent types	Patients	Delayed presentation/Treatment (>24 h)	Sepsis incidence
2025	Wannhoff[12]	Retrospective	10	–	57	–	–
2024	Anundsen[13]	Retrospective	10	WallFlex,EndoFlex, Hanaro	17	6/17 (35%)	–
2023	Chiu[17]	Retrospective	10	Wallflex	5	5/5 (100%)	3/5 (60%)
2018	Hauge[18]	Retrospective	8	Ultraflex, Wallflex, SX- ELLA, Niti-S, Polyflex	15	9/15 (60%)	–
2018	Aloreidi[19]	Retrospective	9	Wallflex	6	4/6 (67%)	6/6 (100%)
2018	Huh[20]	Retrospective	10	Hanarostent, Choo stent	4	–	–
2016	Glatz[21]	Prospective	11	Ultraflex, Leufen	16	4/16 (25%)1	6/16 (38%)
2016	Wu[22]	Retrospective	10	Nanjing	19	16/19 (84%)	5/19 (26%)
2014	Gubler[23]	Retrospective	10	Niti S, Rusch, Ultraflex, Hanarostent	7	–	–
2014	Persson[24]	Retrospective	10	CSEMS	23	–	–
2013	Schweigert[25]	Retrospective Comparative	14	Polyflex, Ultraflex	13	–	9/13 (69%)
2013	Darrien[26]	Retrospective	10	Ultraflex, Polyflex	5	2/5 (40%)	5/5 (100%)
2012	Koivukangas[27]	Retrospective	9	Hanarostent, Nanjing	14	7/14 (50%)	7/14 (50%)
2009	Freeman[28]	Prospective Observational	12	Polyflex	19	–	3/19 (16%)2
2009	Salminen[29]	Retrospective	8	Hanarostent	3	3/3 (100%)	–
2008	Kim[30]	Retrospective	9	Montgomery Salivary Bypass Stent	4	4/4 (100%)	4/4 (100%)
2006	Fischer[31]	Retrospective	8	Ultraflex	5	1/5 (20%)	4/5 (80%)
2006	Prichard[32]	Retrospective	7	CSEMS	5	5/5 (100%)	–
2003	Siersema[33]	Retrospective	10	Flamingo, Ultraflex	5	4/5 (80%)	–
2001	Chung[34]	Retrospective	8	Song, Niti S	3	3/3 (100%)	–

MINORS, Methodological Index for Non-Randomized Studies.

Table 3.

Summary of studies on endoscopic vacuum therapy for esophageal conditions

Year	Author et al	Study type	Country	Conditions assessed (n)	MINORS score	Number of patients	Closure rate (%)	Mean duration of EVT (days)	Average sponge Changes
2008	Wedemeyer[35]	Case series	Germany	AL (2)	10	2	100.0	15.0	4.0
2011	Loske[36]	Case series	Germany	AL (8); IP (3); BS (1); O (1)	10	14	92.9	12.0	4.0
2013	Brangewitz[37]	Retrospective	Germany	AL (32)	8	32	84.4	23.0	7.0
2013	Schneiwind[38]	Retrospective	Germany	AL (17)	10	17	Not available	57.0	Not available
2013	Bludau[39]	Retrospective	Germany	AL (8); IP (3); BS (2); O (1)	9	14	86	12.1	3.9
2016	Kuehn[40]	Retrospective	Germany	AL (11); IP (8)	10	21	90	15.0	5.0
2017	Laukoetter[41]	Prospective	Germany	AL (39); IP (9); BS (4)	10	52	94	22.0	6.0
2018	Bludau[42]	Retrospective	Germany	BS (6); IP (12); AL (59)	14	77	77	11.0	2.75
2019	Alakkari[43]	Case series	UK	AL (1); BS (1)	10	2	100	6.0	8.5
2022	Mastoridis[44]	Prospective	UK	AL (3); BS (3)	8	7	85	13.0	3.0
2022	Richter[45]	Observational	Germany	AL (69); IP (9); BS (7); O (17)	9	102	91	27.5	7.55
2023	Luttikhold[46]	Retrospective	Sweden	IP (16); BS (9); TP (2)	7	27	89	12.0	1.0
2024	Wannhoff[12]	Retrospective	Germany	BS	10	57	80	–	–

AL, anastomotic leak; IP, iatrogenic perforation; BS, Boerhaave syndrome; O, other; TP, traumatic perforation; EVT, endoscopic vacuum therapy

Data extraction and outcome measures

Data were independently extracted by two reviewers using a standardized extraction form. For esophageal stenting studies, the extracted variables included publication year, study design, MINORS score, stent types used, patient numbers, incidence of delayed presentation/treatment (>24 h), sepsis incidence, and procedural details (e.g., concomitant drainage, additional drainage, endoscopic reinterventions, final sealing, failure, surgical conversion, and mortality). For EVT studies, we collected information on study design, country, conditions treated, MINORS score, number of patients, closure rate, mean treatment duration, and average sponge changes.

The primary outcomes for esophageal stenting were defined as follows:

Sealing Rate: The proportion of patients achieving successful closure of the esophageal defect.

Failure Rate: The proportion of cases in which stent therapy was unsuccessful, requiring surgical intervention.

Mortality Rate: The proportion of patients who died in association with the esophageal condition or its treatment.

For EVT, the primary outcome was the sealing rate

Statistical analysis

Meta-analyses were conducted using OpenMeta Analyst software^[16] under a random-effects model to account for inter-study variability. Pooled proportions were calculated using the Freeman–Tukey double arcsine transformation to stabilize variance. Heterogeneity was assessed with the I² statistic, τ², and Cochran’s Q test. Publication bias was evaluated using Eggers test, and sensitivity analyses were performed where significant bias was detected to test the robustness of our findings. Sensitivity analyses were performed in three ways: First, a leave-one-out approach sequentially excluded each study and recalculated pooled estimates to identify any single study’s impact on effect size and heterogeneity. Second, we compared results from the random-effects model with those from a fixed-effects model to ensure findings were not dependent on model choice. A two-tailed P-value of less than 0.05 was considered statistically significant.

Results

In our study, the summary of studies on esophageal stenting are tabulated in Table 1. Table 2 tabulated the procedural interventions and additional outcomes. Table 3 tabulates the summary of studies on endoscopic vacuum therapy for esophageal conditions. The forest plot for EVT initially showed a little over half of treated patients achieving successful sealing, but the results varied wildly across studies and suggested publication bias. When we performed a sensitivity analysis removing the most influential reports, the estimated success rate jumped dramatically to nearly 90% and consistency across studies became perfect. This stark contrast highlights that a handful of studies with extreme results were driving both the low overall estimate and the high variability in the initial analysis as detailed further in this section.

Table 2.

Procedural interventions and additional outcomes

Year	Author et al	Concomitant drainage	Additional drainage	Endoscopic reinterventions	Final sealing	Failure	Surgical conversion	Mortality
2024	Anundsen[13]	15/17 (88%)	14/17 (82%)	8/17 (47%)	16/17 (94%)	1/17 (6%)	0 (0)	1/17 (6%)
2023	Chiu[17]	5/5 (100%)	1/5 (20%)	0 (0)	5/5 (100%)	0 (0)	0 (0)	0 (0)
2018	Hauge[18]	14/15 (93%)	4/15 (27%)	5/15 (33%)	13/15 (87%)	2/15 (13%)	0 (0)	2/15 (13%)
2018	Aloreidi[19]	6/6 (100%)	0 (0)	3/6 (50%)	–	0 (0)	0 (0)	0 (0)
2018	Huh[20]	–	–	–	4/4 (100%)	–	–	–
2016	Glatz[21]	15/16 (94%)	11/16 (69%)	5/16 (31%)	11/16 (69%)	6/16 (38%)	4/16 (25%)	2/16 (13%)
2016	Wu[22]	19/19 (100%)	–	0 (0)	16/19 (84%)	1/19 (5%)	0 (0)	1/19 (5%)
2014	Gubler[23]	–	–	–	5/7 (71%)	–	–	–
2014	Persson[24]	–	–	–	–	3/23 (13%)1	–	–
2013	Schweigert[25]	13/13 (100%)	11/13 (85%)	–	–	2/13 (15%)	0 (0)	2/13 (15%)
2013	Darrien[26]	5/5 (100%)	4/5 (80%)	4/5 (80%)	2/5 (40%)	1/5 (20%)	0 (0)	1/5 (20%)

Across 13 mostly German case series and cohorts published between 2008 and 2024 (MINORS scores 7–14), EVT was applied in 424 patients with a mix of anastomotic leaks, iatrogenic and traumatic perforations, Boerhaave syndrome, and other esophageal defects. Closure rates ranged from 77% to 100%, with early small series reporting universal success (100% in two-patient and two-case series) and larger cohorts achieving 84–94% closure. The mean duration of therapy varied from about 6 to 27 days, and sponge exchanges occurred roughly every 3–7 days. Later, higher-quality studies (MINORS ≥9) tended to show more consistent closure rates around 85–91% over 2–4 weeks of EVT as shown in Table 3.

Further analysis is shown below under subheadings for analyses.

Forest plot for sealing rate with esophageal stent therapy

Fourteen studies^{[13,17,18,20–23,26,28–31,33,34]} were pooled to assess the sealing success of esophageal stenting. In total, 106 successful sealing events were observed among 127 patients, yielding a pooled sealing rate of 86.1% (95% CI: 80.2–92.0%) under a random-effects model. Heterogeneity was minimal (I² = 8.26%, P = 0.362). An Eggers test for publication bias produced a non-significant result (P = 0.42), suggesting no evidence of small-study effects as shown in Figure 2.

Figure 2.

Forest plot of sealing success from 14 studies (106 events/127 patients) showing a pooled rate of 86.1% (95% CI: 80.2–92.0%) with minimal heterogeneity (I² = 8.26%, P = 0.362) and no publication bias (Eggers test, P = 0.42).

Forest plot failure of stent therapy:

Seventeen studies^{[13,17–19,21,22,24–34]} contributed data regarding stent therapy failure, with 28 failures recorded in 177 patients. The overall failure rate was 14.9% (95% CI: 8.5–21.3%). Moderate heterogeneity was detected (I² = 40.78%, P = 0.041). The Eggers test did not indicate significant publication bias (P = 0.37) as shown in Figure 3.

Figure 3.

Forest plot of stent therapy failure from 17 studies (28 failures/177 patients), showing an overall failure rate of 14.9% (95% CI: 8.5–21.3%) with moderate heterogeneity (I² = 40.78%, P = 0.041) and no significant publication bias (Eggers test, P = 0.37).

Forest plot mortality with esophageal stenting:

Data from the same set of studies^{[13,17–19,21,22,24–34]} were used to evaluate mortality, with 15 deaths occurring among 154 patients. The pooled mortality rate was 7.4% (95% CI: 3.5–11.4%), and no heterogeneity was observed (I² = 0%, P = 0.903). Eggers test results (P = 0.56) further supported the absence of publication bias in this analysis as shown in Figure 4.

Figure 4.

Forest plot of mortality from esophageal stenting studies (15 deaths/154 patients) showing a pooled mortality rate of 7.4% (95% CI: 3.5–11.4%) with no heterogeneity (I² = 0%, P = 0.903) and no publication bias (Eggers test, P = 0.56).

Forest plot sealing rate with endoscopic vacuum therapy:

Fifteen studies^{[12,13,35–46]} were analyzed for EVT sealing outcomes. Overall, 125 events were observed among 413 patients, resulting in a pooled sealing rate of 54.1% (95% CI: 34.8–73.4%) using a binary random-effects model. Substantial heterogeneity was present (I² = 98.24%, P < 0.001). Moreover, the Eggers test for this analysis was significant (P = 0.01), indicating potential publication bias and small-study effects as shown in Figure 5.

Figure 5.

Forest plot of endoscopic vacuum therapy sealing outcomes from 15 studies (125 events/413 patients) showing a pooled sealing rate of 54.1% (95% CI: 34.8–73.4%) with substantial heterogeneity (I² = 98.24%, P < 0.001) and significant publication bias (Eggers test, P = 0.01).

The sensitivity analysis shows a substantial change; after removing some influential studies^{[12,13,37,41,42,45]}, the pooled sealing rate for EVT is now 89.6% (95% CI: 83.9–95.3%), with no heterogeneity (I² = 0%, P = 0.998). This suggests that those studies were major contributors to the significant publication bias and overall heterogeneity in the initial analysis as shown in Figure 6.

Figure 6.

Sensitivity analysis of endoscopic vacuum therapy sealing outcomes after excluding six outlier studies showing a pooled sealing rate of 89.6% (95% CI: 83.9–95.3%) with no heterogeneity (I² = 0%, P = 0.998).

Discussion

Our study sought to deepen our understanding of the role of minimally invasive therapies – specifically esophageal stenting and EVT – in the treatment of complex esophageal defects. When we pooled data from 14 observational studies, esophageal stenting emerged with a notably high sealing success rate of 86.1% (95% CI: 80.2–92.0%), a failure rate of 14.9% (95% CI: 8.5–21.3%), and a mortality rate of 7.4% (95% CI: 3.5–11.4%). These results are in strong agreement with earlier reports by Margaris et al^[7], who demonstrated that stents function as an effective bridge, maintaining continuity in the esophagus while allowing the defect to heal and preventing further contamination of the mediastinum and pleural space. This temporary “bridging” is especially important in the context of Boerhaave syndrome and other perforations, where the risk of sepsis is high. By avoiding the need for major surgical intervention – which is both physiologically taxing and carries inherent risks – stenting offers an appealing option for patients with multiple comorbidities or those deemed poor surgical candidates^[5,7,47]. Equally significant is the role of concomitant drainage procedures; our review and other studies^[5,7,47] highlight that proper drainage, when combined with stenting, is critical to controlling infection and facilitating recovery.

In contrast, our initial evaluation of EVT outcomes was far more variable. Across 15 studies, the pooled sealing rate for EVT was only 54.1% (95% CI: 34.8–73.4%), accompanied by substantial heterogeneity (I² = 98.24%, P < 0.001) and indications of publication bias (Eggers test, P = 0.01). This inconsistency likely reflects the diverse patient populations, differing defect characteristics, and the lack of standardized treatment protocols that currently characterize EVT use. Early experiences – such as those described by Wedemeyer et al^[35] and further documented by Alakkari et al^[43] – demonstrated that EVT could serve as a vital rescue treatment after conventional surgery and stenting had failed. These pioneering reports provided the initial evidence that EVT not only promotes rapid wound closure but also minimizes the duration of hospital stays by facilitating continuous drainage of contaminated collections.

The initial low EVT sealing rate and high heterogeneity were driven by a few early, small-cohort studies that used bespoke, nonstandardized sponge-exchange schedules, enrolled predominantly the sickest patients, and varied widely in adjunctive drainage practices. These design limitations – heterogeneous inclusion criteria, inconsistent intervention protocols, and lack of control for confounders – amplified outlier effects and introduced publication bias.

What is particularly intriguing is how subsequent studies have refined our understanding of EVT. For instance, Brangewitz et al^[37] and Schneiwind et al^[38] provided evidence that EVT could achieve higher closure rates and reduce mortality compared to stenting, especially in patients with severe systemic inflammation. Laukoetter et al^[41] reported healing in over 94% of patients treated solely with EVT – a result that initially seemed at odds with the pooled average in our analysis. Recognizing the heterogeneity in EVT outcomes, we performed a sensitivity analysis that excluded several influential outlier studies. This analysis dramatically improved the pooled sealing rate to 89.6% (95% CI: 83.9–95.3%) and completely eliminated statistical heterogeneity (I² = 0%, P = 0.998). This finding strongly suggests that when EVT is applied in a more uniform manner – perhaps through standardized protocols and careful patient selection – its efficacy may well be comparable to, if not exceed, that of stenting. For instance, Schneiwind et al^[38] reported that in systemically ill patients with similar APACHE II scores, those treated with EVT experienced markedly improved results, with a mortality rate of only 12%, as opposed to 50% for surgery and 83% for stenting (P = 0.01 and P = 0.0014, respectively). In a separate study, Laukoetter et al^[41] demonstrated that EVT alone was sufficient to heal 94.2% of 52 patients, eliminating the necessity for any further interventions.

Yet, no therapy is without its drawbacks. EVT, while promising, comes with practical challenges: it requires frequent endoscopic reassessments and sponge exchanges, which can extend hospital stays and increase costs. There are also complications to consider, such as sponge dislocation, bleeding during exchanges, and even the development of strictures. Some authors have debated whether the strictures observed are a direct consequence of EVT or are related to the complex nature of the esophageal pathology itself^{[7,37,39,41,42]}. In our view, these issues underscore the importance of further refining EVT protocols and implementing rigorous training and standardization across centers.

It is also essential to acknowledge the limitations inherent in our analysis. All the data we synthesized come from observational studies and case series, with modest methodological quality as indicated by average MINORS scores in the range of 9–10. Such study designs are susceptible to biases – selection bias, lack of blinding, and uncontrolled confounding – which means that our findings, though informative, should be interpreted with cautious optimism.

The absence of randomized controlled trials introduces several key limitations. Without random allocation, differences in baseline characteristics – such as defect size, patient comorbidities, or timing of presentation – may skew comparisons between stenting and EVT. The unblinded nature of observational series also risks assessment bias, since clinicians aware of the chosen therapy might unconsciously interpret imaging or clinical signs to favor their expected outcome. Moreover, uncontrolled confounders – such as variations in adjunctive drainage techniques, operator experience, and timing of intervention – cannot be fully adjusted for, making it difficult to isolate the true effect of each modality. Small sample sizes magnify the influence of outliers, and the tendency to report positive results preferentially further limits the reliability and generalizability of our findings. Only a well-designed RCT can overcome these issues and establish definitive, causally robust evidence.

In practical terms, our findings advocate for a personalized approach to the management of esophageal defects. For patients with high surgical risk or extensive comorbidities, esophageal stenting appears to offer a reliable, less invasive solution with a high success rate. Meanwhile, EVT holds considerable promise as an alternative or adjunct treatment – especially in settings where stenting has failed or is contraindicated. The evolution of EVT, as illustrated by improved outcomes in sensitivity analyses, suggests that uniform protocols and better patient selection can substantially enhance its effectiveness. Ultimately, both therapies are important tools in the endoscopist’s arsenal, and the choice between them should be guided by individual patient factors, available expertise, and institutional resources.

In conclusion, our meta-analysis reinforces the notion that both esophageal stenting and EVT are viable, minimally invasive options for treating esophageal leaks and perforations. While stenting shows consistently high sealing rates and low mortality, EVT demonstrates significant potential when applied under standardized conditions.

To establish evidence-based guidelines, multicenter randomized trials are urgently needed. Such studies should employ standardized EVT protocols – uniform sponge-exchange intervals, clear patient selection criteria, and predefined drainage adjuncts – and directly compare EVT against stenting. This approach will minimize bias, harmonize practice, and clarify which therapy offers the best outcomes across diverse clinical settings.

Conclusion

Both esophageal stenting and EVT are effective minimally invasive strategies for managing esophageal defects. Esophageal stenting achieved a pooled sealing rate of 86.1% (95% CI: 80.2–92.0%), a failure rate of 14.9% (95% CI: 8.5–21.3%), and mortality of 7.4% (95% CI: 3.5–11.4%). EVT demonstrated an initial sealing rate of 54.1% (95% CI: 34.8–73.4%) – rising to 89.6% (95% CI: 83.9–95.3%) after sensitivity analysis eliminated heterogeneity – highlighting its potential when standardized protocols are applied. Despite the limitations of observational data, these quantified results underscore the need for multicenter randomized trials to validate and refine individualized treatment algorithms.

Ethics approval

Not applicable for this systematic review/meta-analysis.

Consent

Informed consent was not required for this meta-analysis.

Sources of funding

No sources of funding for this meta-analysis.

Author contributions

M.S.: led the conceptualization, methodology, data curation, formal analysis, writing of the original draft, and supervision. M.I.: contributed to investigation, formal analysis, review and editing of the manuscript, and visualization. K.A.: and S.K.: was responsible for software development, validation, data curation, and review and editing. H.S.: handled resources, data curation, review and editing, and project administration. M.A.I.: contributed to validation, visualization, and review and editing. H.M.: assisted with review and editing and provided resources. All authors have read and approved the final manuscript.

Conflicts of interest disclosure

The authors declare that they have no conflicts of interest.

Research registration unique identifying number (UIN)

The study has been registered in the international prospective register of systematic reviews (PROSPERO) and the registration can be accessed using the link: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251012734 PROSPERO ID = CRD420251012734.

Guarantor

Muhammad Saqib.

Provenance and peer review

Not commissioned, externally peer-reviewed.

Data availabilityand statement

All data used in this paper have been taken from publicly available resources.