The impact of modified esophageal stenting on oral intake in delayed spontaneous esophageal perforation

Abstract

Background

Spontaneous esophageal perforation, commonly referred to as Boerhaave syndrome, is a rare but life-threatening condition with a high mortality rate. Delayed diagnosis is common due to nonspecific symptoms, and optimal treatment remains challenging. This study evaluates the outcomes of a modified esophageal stenting technique in patients with a delayed diagnosis.

Methods

This retrospective study analyzed 18 patients treated for spontaneous esophageal perforation at a tertiary medical center between July 2012 and July 2022. Patients were divided into early (< 24 h) and delayed (≥ 24 h) diagnosis groups. A modified esophageal stenting approach, which combined fully covered self-expandable metal stent (FCSEMS) placement across the esophagogastric junction with minimally invasive drainage and external fixation, was applied in selected delayed cases. Outcomes were compared between the treatment strategies.

Results

Among the 11 patients in the delayed group, 5 underwent modified esophageal stenting. These patients had significantly lower Charlson Comorbidity Index scores and shorter operative durations compared to those receiving other treatments. Although the overall complication and mortality rates were higher in the delayed group (45.5% and 18.2%, respectively), patients treated with modified stenting were more likely to resume oral nutrition at discharge (p = 0.048), without an increase in hospital stay or postoperative complications.

Conclusions

Modified esophageal stenting is a promising, less invasive option for managing delayed spontaneous esophageal perforation, facilitating earlier recovery without compromising safety.

Introduction

Esophageal perforation is a rare but life-threatening condition with a high mortality rate. Spontaneous esophageal perforation, often referred to as Boerhaave syndrome, accounts for approximately 15–25% of all cases and involves transmural rupture due to a sudden increase in intraluminal pressure [1, 2]. Among the non-malignant etiologies of esophageal perforation, which include iatrogenic, traumatic, and foreign body-related causes, spontaneous esophageal perforation carries the gravest prognosis [3]. Early diagnosis and surgical intervention within 24 h are essential and have been shown to significantly improve outcomes [2, 4]. Mackler’s triad (vomiting, chest pain, and subcutaneous emphysema) is the classic presentation but is observed in fewer than 30% of patients, with most presenting with nonspecific symptoms [5] Consequently, diagnosis is delayed in over 30% of cases, which adversely impacts prognosis [2–7]

Management of delayed esophageal perforation is complex, and outcomes remain poor [8]. While primary repair is a viable option, it carries a high risk of leakage due to secondary tissue necrosis and fragility in delayed presentations [9]. Controlled fistula formation using T-tube drainage has been proposed as a reliable alternative treatment [10–11]. With the advancement of minimally invasive techniques, esophageal stenting combined with abscess drainage has shown favorable results [6, 12]. In cases of extensive disruption, esophageal exclusion, diversion, or esophagectomy may be considered as salvage procedures [13].

In this study, we present a 10-year review of spontaneous esophageal perforation, with particular emphasis on a modified esophageal stenting technique developed for delayed cases. This approach combines endoscopic stenting with minimally invasive surgical drainage, aiming to control contamination and promote healing while preserving esophageal continuity. We compare the outcomes of this strategy with those of conventional modalities, highlighting its role as a promising and less invasive alternative for managing delayed perforations.

Methods

Patients

This retrospective study included all patients referred to the thoracic surgery service for spontaneous esophageal perforation at Chang Gung Memorial Hospital, Linkou Branch, between July 2012 and July 2022. Delayed esophageal perforation was defined as a diagnosis made more than 24 h after the onset of initial symptoms. Patients were excluded if they had a recent history of upper endoscopy, esophageal instrumentation, intrathoracic or upper abdominal surgery, or if the perforation was associated with malignancy, achalasia, or a paraesophageal hernia. The study protocol was approved by the Institutional Review Board (No. 202500255B0, 202300960B0). The requirement for informed consent was waived due to the retrospective study design.

Evaluation

For patients with suspected esophageal rupture, computed tomography (CT) of the chest and abdomen was performed to evaluate the presence of air or fluid collections in the mediastinum, pleural cavity, or peritoneal cavity. Additionally, esophagography was also conducted to identify the location and severity of the perforation. Patient demographics, clinical presentations, and relevant clinical data were retrieved from medical records.

The Pittsburgh Perforation Severity Score (PSS) is a validated tool for stratifying the severity of esophageal perforation and predicting clinical outcomes, particularly in patients with Boerhaave syndrome [7, 14]. Therefore, the PSS was applied to each patient in this cohort. The scoring system is based on the following criteria:

• 1 point: Age > 75 years, tachycardia, leukocytosis, or pleural effusion.

• 2 points: Fever > 38.5 °C, non-contained leak, respiratory compromise, or time to diagnosis > 24 h.

• 3 points: Presence of cancer or hypotension.

Based on their total Pittsburgh Perforation Severity Score (PSS), patients were categorized into three risk groups: low risk (PSS < 2), intermediate risk (PSS 3–5), and high risk (PSS > 5).

Treatment

The management of esophageal perforation was individualized based on patient comorbidities, the severity of the disease, and the surgeon’s clinical judgment (Fig. 1). Primary repair was performed via thoracotomy, with or without reinforcement using a pleural patch or an intercostal muscle flap.

Fig. 1.

Treatment algorithm for spontaneous esophageal perforation. FCSEMS: Fully covered self-expandable metal stent; VATS: Video-assisted thoracoscopic surgery

A modified esophageal stenting technique was applied in selected patients with delayed presentation (> 24 h), in whom primary repair was considered unsafe due to tissue friability, extensive mediastinal contamination, or secondary tissue devitalization, while preservation of esophageal continuity remained feasible.

In this hybrid strategy, initial infection source control was achieved via video-assisted thoracoscopic surgery (VATS). The procedure involved comprehensive decortication and debridement, including evacuation of pleural effusion, pus, debris, and fibrinous material. The perforation site was systematically evaluated intraoperatively. When primary repair was deemed unsafe, a staged approach with planned esophageal stenting was adopted. Upon completion of VATS, chest tubes were routinely positioned adjacent to the perforation and in dependent regions to ensure adequate drainage. Additional interventions, including image-guided percutaneous drainage, were performed when clinically indicated.

Esophageal stent placement was subsequently undertaken as a second-stage procedure following initial stabilization and effective surgical drainage. Stenting was avoided in patients with extensive circumferential esophageal necrosis or long-segment perforations not amenable to complete coverage by a single stent. During the second stage, esophagoscopy was performed, and a fully covered self-expandable metal stent (FCSEMS; WallFlex, Boston Scientific, Natick, MA, USA) was deployed across the gastroesophageal junction. To prevent migration, the stent was secured using an extraluminal suturing technique via a mini-laparotomy approach with absorbable sutures, including polydioxanone (PDS; Ethicon US, LLC, Guaynabo, Puerto Rico) or polyglycolic acid (Dexon; Medtronic, Minneapolis, MN, USA). At least two sutures were placed to achieve adequate fixation (Fig. 2) [12].

Fig. 2.

Modified esophageal stenting technique. A fully covered self-expandable metal stent is placed to extend 2 cm beyond the gastroesophageal junction (indicated by the green dotted line and blue double-headed arrow). Extraluminal fixation is performed at the distal end of the esophageal stent using absorbable sutures (illustrated as red stitches)

Stent diameter was selected according to the esophageal diameter measured on preoperative esophagography to ensure adequate wall apposition for effective sealing while minimizing the risk of migration. Stent length was chosen to provide at least a 2-cm margin both proximal and distal to the perforation site to ensure complete coverage. A feeding jejunostomy was routinely established to facilitate early enteral nutrition, and gastric decompression was achieved using either a gastrostomy or a nasogastric tube.

For patients with extensive contamination or poor tissue condition in whom primary repair or esophageal stenting was not feasible, esophageal exclusion and diversion were considered as a damage-control strategy. This procedure included cervical esophagostomy and distal esophageal isolation at the esophagogastric junction via laparotomy [15]. Esophagectomy was reserved for patients with extensive tissue devitalization or secondary necrosis associated with delayed presentation, large perforations extending onto the gastric side, or persistent uncontrolled leakage despite prior interventions [16].

Outcomes and follow up

Following surgical intervention, all patients were admitted to the intensive care unit (ICU) for postoperative monitoring. Transfer to the general ward was arranged once hemodynamic stability was achieved and inotropic support was no longer required. Postoperative complications were assessed using the Clavien-Dindo classification system [17]. In this study, minor complications were defined as grades I to IIIa, and major complications as grades IIIb to V. Oral nutrition was initiated after esophagography confirmed the absence of esophageal leakage. Patients began with water intake, followed by a liquid diet, and progressed to a soft diet as tolerated.

For patients who underwent esophageal stenting, stent removal was routinely performed 8–12 weeks after initial placement via endoscopy under light sedation.

Statistical analysis

Categorical variables were presented as counts and percentages, while continuous variables were reported as medians with interquartile ranges. The Mann-Whitney U test was employed to compare differences between the early and delayed esophageal perforation groups, as well as to evaluate outcomes between the modified esophageal stenting group and other treatment modalities in the subgroup of patients with delayed perforation. Binary variables were compared using the Chi-square test or Fisher’s exact test, depending on data distribution, and expected frequencies.

Results

Patient characteristics

A total of 20 patients were diagnosed with spontaneous esophageal perforation during the study period. After excluding one patient with a history of prior surgery and another who received initial treatment at the other hospital, 18 patients were included in the final analysis (Fig. 3).

Fig. 3.

Flow chart of inclusion and exclusion of the studies

The most common presenting symptoms were vomiting (72.2%), chest pain (55.6%), fever (38.9%), epigastric pain (33.3%), and dyspnea (33.3%). Mackler’s triad was observed in only two patients (11.1%).

The most frequent findings on computed tomography were pneumomediastinum (83.3%) and pleural effusion or empyema (83.3%). Esophagography demonstrated esophageal leakage in 94.4% of patients, with 61.1% of leaks located in the distal esophagus and 33.3% at the esophagogastric junction (Table 1).

Table 1.

Clinical presentation and image findings for spontaneous esophageal perforation

Variables	Number (%)
Clinical presentation
Vomiting	13(72.2)
Chest pain	10(55.6)
Epigastric pain	6(33.3)
Fever	7(38.9)
Dyspnea	6(33.3)
Hematemesis	5(27.8)
Subcutaneous emphysema	3(16.7)
Computed tomography
Pneumomediastinum	15(83.3)
Pneumothorax	7(38.9)
Pneumoperitoneum	1(5.6)
Pneumopericardium	1(5.6)
Subcutaneous emphysema	3(16.7)
Mediastinitis	2(11.1)
Pleural effusion or empyema	15(83.3)
Esophagogram
Esophageal leak	17(94.4)
Distal esophagus	11(61.1)
Esophagogastric junction	6(33.3)

Categorical data shown as n (%)

Comparison between early and delayed groups

Among the 18 patients, 7 were classified into the early diagnosis group and 11 into the delayed diagnosis group. Patients in the delayed group had significantly higher PSS compared to those in the early group (p < 0.01). Primary repair was the predominant treatment modality in the early group, whereas the delayed group was associated with longer operative durations and greater intraoperative blood loss. Although patients in the delayed group experienced prolonged ICU stays, extended total hospitalizations, and delayed initiation of oral nutrition, these differences were not statistically significant (Table 2).

Table 2.

Characteristics of Patients with Spontaneous Esophageal Perforation

		Time to diagnosis
Variables	All (n = 18)	Early (n = 7)	Delayed (n = 11)	P Value
Age, years	63(55; 71.8)	65(62; 74)	59(48; 69)	0.20
Male	17(94.4)	7(100)	10(90.9)	1.00
Charlson comorbidity index	4(2; 5.3)	4(2; 6)	3(2; 5)	0.55
PSS score				< 0.01
Low (< 2) Intermediate (3–5) High (> 5)	0(0) 6(33.3) 17(66.7)	0(0) 6(85.7) 1(14.3)	0(0) 0(0) 11(100)
Treatment				0.001
Primary repair Esophageal stenting Exclusion and diversion Esophagectomy with delayed reconstruction	6(33.3) 5(27.8) 1(5.6) 6(33.3)	6(85.7) 0(0) 0(0) 1(14.3)	0(0) 5(45.5) 1(9.1) 5(45.5)
*ICU stay, days	3.5(1; 7.5)	1(1; 5)	5(2; 12.5)	0.13
*Hospital stay, days	30.5(22.3; 45.8)	24(22; 41)	32(25.5; 74)	0.20
Surgical duration, mins	347(213.3; 382.8)	196(153; 346)	367(343; 385)	0.016
Blood loss, mL	100(50; 482.5)	50(50; 100)	200(100; 1000)	0.017
Surgical complications				0.64
No/ Minor Major (3b)	11(61.1) 7(38.9)	5(71.4) 2(28.6)	6(54.5) 5(45.5)
Hospital mortality	2(11.1)	0(0)	2(18.2)	0.50
*Oral nutrition while discharge	9(56.3)	5(71.4)	4(44.4)	0.36
*Time to oral nutrition, days	15(9.9; 113.5)	11.9(8.0; 43)	34(14; 134)	0.063

Categorical data shown as n (%). Continuous data expressed as median (interquartile)

ICU: intensive care unit; PSS: Pittsburgh Perforation Severity Score

* Patients with in-hospital mortality were excluded

Subgroup analysis of the delayed group

Among the 11 patients in the delayed diagnosis group, 5 underwent modified esophageal stenting as the primary treatment strategy [12]. Patients who received esophageal stenting had significantly lower Charlson Comorbidity Index scores compared to those who received other treatments (p = 0.04). All patients in the delayed group had a PSS greater than 5, with a major complication rate of 45.5% and a mortality rate of 18.2%.

Compared to other treatment modalities, modified esophageal stenting was associated with a significantly shorter operative duration (343 min vs. 383.5 min, p = 0.045). Although patients in the stenting group had lower rates of major complications (20% vs. 66.7%, p = 0.24) and hospital mortality (0% vs. 33.3%, p = 0.46), these differences did not reach statistical significance. Additionally, there were no significant differences in ICU stay or total hospital stay between the two groups. Notably, a significantly higher proportion of patients in the modified esophageal stenting group resumed oral nutrition at discharge (p = 0.048) (Table 3). One patient in the stenting group was unable to resume oral intake due to underlying frailty and persistent swallowing dysfunction, rather than stent-related complications. All esophageal stents were successfully removed 8–12 weeks after initial placement via endoscopy, without technical difficulties.

Table 3.

Characteristics of patients with delayed spontaneous esophageal perforation

Variables	All (n = 11)	Treatment
		Esophageal stenting (n = 5)	Other (n = 6)	P Value
Age, years	59(48; 69)	49(48; 72)	65(55.5; 73.3)	0.31
Male	10(90.9)	4(80)	6(100)	0.46
Charlson comorbidity index	3(2; 5)	2(0; 3.5)	5(2.8; 7)	0.04
PSS score				NA
Low (< 2) Intermediate (3–5) High (> 5)	0(0) 0(0) 11(100)	0(0) 0(0) 5(100)	0(0) 0(0) 6(100)
*ICU stay, days	5(2; 12.5)	6(2; 23)	4.5(1.8; 14)	0.71
*Hospital stay, days	32(25.5; 74)	32(24; 77.5)	38(24.3; 70.5)	1.00
Surgical duration, mins	367(343; 385)	343(296; 371.5)	383.5(360; 480)	0.045
Blood loss, mL	200(100; 1000)	150(75; 465)	600(100; 1775)	0.23
Surgical complications				0.24
No/ Minor Major (3b)	6(54.5) 5(45.5)	4(80) 1(20)	2(33.3) 4(66.7)
Hospital mortality	2(18.2)	0(0)	2(33.3)	0.46
*Oral nutrition while discharge	4(44.4)	4(80)	0(0)	0.048
*Time to oral nutrition, days	34(14; 134)	20.8(12; 67)	167(134; NA)	0.053

Categorical data shown as n(%) . Continuous data expressed as medium (interquarile)

ICU: intensive care unit; NA : not applicable; PSS: Pittsburg Perforation Severity Score

*Patients with in-hospital were excluded

Discussion

Spontaneous esophageal perforation is a life-threatening condition associated with high mortality. Although primary repair within 24 h is linked to favorable outcomes, delayed diagnosis is common due to nonspecific clinical manifestations. In delayed presentations, ongoing mediastinal contamination and inflammation may lead to secondary tissue devitalization and necrotic changes, which differ from the primary pathophysiology of classic Boerhaave syndrome. Compared to other treatment strategies, patients with delayed esophageal perforation who undergo modified esophageal stenting experience shorter operative durations and a higher likelihood of resuming oral nutrition at discharge, without an associated increase in hospital stay or postoperative complications.

Timely diagnosis and appropriate treatment are crucial in the management of esophageal perforation. However, due to its rarity and nonspecific symptomatology, diagnosis remains challenging, with reported delayed diagnosis rates ranging from 33% to 42%. [2, 3, 5, 8, 18] Esophagography is the most commonly used diagnostic tool to identify thoracic or abdominal esophageal perforations, though 10–20% of cases may be missed [19]. In addition to esophagography, CT can evaluate the extent of contamination and assist in determining the need for surgical versus conservative treatment [19]. Compared to esophagography, CT has higher sensitivity (100.0% vs. 77.8%) but lower specificity (79.8% vs. 98.9%) [20]. Recently, CT with oral contrast has been shown to improve the timeliness and accuracy of esophageal perforation diagnosis [20, 21]. In this cohort, esophagography successfully identified esophageal perforation in 94.4% of cases. The most common CT findings were pneumomediastinum, pleural effusion, empyema, and pneumothorax.

Primary repair for spontaneous esophageal perforation within 24 h has been associated with favorable outcomes [1, 2, 4, 7, 9]. For patients unsuitable for primary repair, alternative treatment options include T-tube drainage, esophageal stenting, exclusion with diversion, or esophagectomy [6, 10–13]. In this cohort, patients with early diagnosis had lower PSS, and primary repair was the predominant surgical intervention. These patients also experienced shorter operative durations and reduced intraoperative blood loss. Although the early diagnosis group showed trends toward shorter ICU stays, shorter total hospitalizations, fewer postoperative complications, and earlier resumption of oral nutrition, these differences were not statistically significant. This may be explained by the fact that patients in the early group were older and had higher Charlson Comorbidity Index scores, which may have attenuated the potential benefits of early surgical intervention.

Esophageal stents, including FCSEMS and partially covered self-expandable metal stents (PCSEMS), are commonly used in both malignant and benign esophageal diseases [22, 23]. The European Society of Gastrointestinal Endoscopy (ESGE) recommends FCSEMS as the preferred option in benign conditions, with removal advised within a maximum of three months [22]. To ensure optimal anchorage, at least 2 cm of normal esophageal tissue proximal and distal to the lesion is suggested during stent placement [23]. However, FCSEMS use, benign pathology, and distal esophageal location are recognized risk factors for stent migration [24]. Given the benign and typically distal nature of the perforation in spontaneous esophageal perforation, we placed an FCSEMS across the esophagogastric junction to ensure sufficient landing distance. To minimize the risk of stent migration, we employed absorbable sutures for external fixation and scheduled stent removal within three months, consistent with ESGE recommendations [12].

The mortality rate associated with delayed esophageal perforation ranges from 40% to 60%, primarily due to the rapid progression of mediastinitis, sepsis, and subsequent organ dysfunction caused by the translocation of bacteria and digestive enzymes into the mediastinum [8]. A multicenter study reviewing 288 patients with esophageal perforation reported morbidity and mortality rates of 82.5% and 37.5%, respectively, among high-risk patients with PSS > 5 [14]. In our cohort, all patients in the delayed diagnosis group had high PSS, with a major complication rate of 45.5% and a mortality rate of 18.2%. Among these patients, those treated with the modified esophageal stenting technique had shorter operative durations and a higher likelihood of resuming oral nutrition at discharge, without an increase in surgical complications or hospital stay. These findings suggest that modified esophageal stenting might be a viable and less invasive alternative for managing high-risk patients with delayed esophageal perforation.

Limitations

This study has several limitations. First, the rarity of spontaneous esophageal perforation resulted in a limited sample size despite a 10-year study period, which may restrict the statistical power and generalizability of the findings. Second, the retrospective design introduces the potential for selection bias and unmeasured confounding. Third, multiple surgeons were involved over the study period, and variations in surgical expertise and clinical decision-making may have influenced treatment selection and outcomes.

In addition, the treatment strategy was not standardized, and the choice of intervention was based on clinical judgment, which may further contribute to selection bias. Finally, the small sample size limits the ability to detect statistically significant differences in clinical outcomes between treatment groups. Larger, prospective studies are warranted to validate these findings.

Conclusions

Spontaneous esophageal perforation is a rare and diagnostically challenging condition due to its nonspecific clinical presentation. Timely diagnosis and appropriate intervention are essential for improving patient outcomes.

In patients with delayed presentation, a modified esophageal stenting approach may be considered as a treatment option in carefully selected cases, particularly when primary repair is deemed unsafe but preservation of esophageal continuity remains feasible. In this study, this approach was associated with earlier resumption of oral nutrition without an observed increase in postoperative complications or hospital length of stay; however, these findings should be interpreted with caution given the limited sample size and retrospective design.

Author contributions

C.C. designed the study, developed the methodology. O. L. and P. C. wrote the main manuscript. O .L., Y. W. and S. H. collected and analyzed the data, P .C., W. C., C. F. W., H. F. contributed to the results section. O. L., P. C., C. Y. W. and C .C. contributed the discussion section. O. L., P. C., C. C. contributed to the table and figure legend. Y. L., Y. C. and C .C. revised the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.