Severity-guided management of esophageal perforation: a retrospective cohort study supporting a pragmatic treatment algorithm

Abstract

Background

Esophageal perforation is a rare but life-threatening condition associated with high morbidity and mortality. Optimal management remains controversial, particularly regarding the indications for conservative, endoscopic, and surgical strategies.

Methods

We conducted a retrospective single-center cohort study including adult patients treated for esophageal perforation between January 2018 and November 2023. Patients were categorized into conservative management (CM) or surgical management (SM). Clinical presentation, etiology, treatment modalities, complications, and mortality were analyzed. A subgroup analysis compared esophageal preservation with esophagectomy within the surgical group.

Results

Fifty-eight patients were included (CM: n = 30; SM: n = 28). Patients in the SM group presented with significantly greater severity, including higher rates of mediastinitis (96.4% vs. 40%, p < 0.001), tachycardia (82.1% vs. 33.3%, p < 0.001), and vasopressor requirement (71.4% vs. 13.3%, p < 0.001). Severe complications were more frequent in the SM group (100% vs. 46.6%, p < 0.001), as was prolonged hospitalization (82.1% vs. 36.7%, p < 0.001). In-hospital mortality was similar between groups (21.4% vs. 20%, p = 0.89). Among surgically treated patients, esophagectomy was associated with higher mortality than esophageal preservation (50% vs. 10%, p = 0.02), likely reflecting greater baseline severity.

Conclusions

Management of esophageal perforation is primarily driven by initial clinical severity and the extent of contamination. Stable patients may be successfully managed with conservative or endoscopic approaches, whereas surgery is required in unstable patients or in the presence of extensive contamination. When surgery is indicated, esophageal preservation should be preferred whenever feasible.

Introduction

Esophageal perforation is defined as a full-thickness disruption of the esophageal wall, resulting in leakage of digestive secretions and saliva into surrounding tissues [1, 2]. Although rare, with 4765 cases reported in France between 2012 and 2021 [3], it is associated with a high mortality rate ranging from 20 to 30% [3–6]. The main etiologies include iatrogenic injury (46.5%), spontaneous rupture (37.8%), foreign body ingestion (6.3%), and caustic ingestion (1.8%) [7]. The incidence of iatrogenic perforations is increasing with the expansion of endoscopic procedures [8, 9]. Despite advances in diagnostic and therapeutic techniques, mortality remains substantial [6].

Prognosis is influenced by multiple factors, including patient-related characteristics (age, comorbidities) [4, 10–13], perforation-specific features (size, location, tissue viability) [4, 14–16], and, critically, the timing and adequacy of management [12, 17]. Early treatment, particularly within 24 h, has been shown to significantly reduce mortality, from approximately 20 to 7.4% [17, 18]. Prompt diagnosis is therefore essential but can be challenging due to nonspecific clinical presentations. Contrast-enhanced CT remains the cornerstone of diagnosis and may be complemented by diagnostic and/or therapeutic endoscopy [19–21]. Management requires a multidisciplinary approach involving gastroenterologists, surgeons, and intensivists.

The optimal therapeutic strategy remains debated. Conservative management (CM), including medical, endoscopic, and radiological interventions, and surgical management (SM) represent complementary rather than competing approaches, each suited to specific clinical scenarios. Appropriate patient selection is therefore crucial to optimize outcomes.

The aim of this retrospective single-center study was to compare morbidity and mortality between patients managed with CM and those undergoing SM, and to identify clinical factors influencing treatment selection. Based on these findings and existing literature, we also propose a pragmatic decision-making algorithm. Specifically, we sought to determine which patients benefit from early surgical intervention and which can be safely managed with esophageal preservation strategies.

Materials and methods

Study design and patient selection

This retrospective single-center cohort study was conducted at the University Hospital of Tours (France). Adult patients (≥ 18 years) managed for esophageal perforation between 8 January 2018 and 9 November 2023 were included.

Eligible patients were identified through a systematic search of the institutional database using International Classification of Diseases (ICD-10) codes K22.3 (esophageal perforation) and T81.2 (accidental puncture or laceration during a procedure). Patients were excluded if they were younger than 18 years, initially managed at another institution, died before initiation of treatment, or if the diagnosis was not confirmed. The patient selection process and treatment allocation are summarized in Fig. 1.

Fig. 1.

Flow diagram of patient selection and treatment allocation. Patients with esophageal perforation were identified using ICD-10 codes and included after confirmation by CT and/or endoscopy. After exclusion criteria, patients were categorized into conservative management (CM) or surgical management (SM). A subgroup analysis was performed within the SM group comparing esophageal preservation and esophagectomy

Diagnostic criteriade

The diagnosis of esophageal perforation was confirmed by contrast-enhanced CT and/or upper gastrointestinal endoscopy. CT findings considered diagnostic were contrast leakage, pneumomediastinum, subcutaneous emphysema, discontinuity of the esophageal wall, or esophageal enhancement defects (Fig. 2). Endoscopic visualization of a transmural defect confirmed the diagnosis (Fig. 3).

Fig. 2.

Axial CT scan of an esophageal perforation with indirect signs. Presence of pleural effusion (red arrow) with pneumomediastinum (green arrow) (A), and pneumothorax (white arrow) (B)

Fig. 3.

Endoscopic image of an esophageal perforation. The red star indicates the esophageal lumen, and the green triangle indicates the site of perforation

Initial management

All patients were admitted to the intensive care unit (ICU). Initial management included fasting, intravenous fluid resuscitation, broad-spectrum antibiotic therapy in accordance with institutional protocols, antifungal therapy when indicated, and vasopressor support (norepinephrine) in cases of hemodynamic instability. Further diagnostic evaluation was performed using contrast-enhanced CT and/or endoscopy according to clinical stability.

Treatment strategies

Management decisions were made on an individualized basis through multidisciplinary discussion involving intensivists, surgeons, and gastroenterologists. The treatment strategy was determined according to hemodynamic status, extent of contamination, perforation characteristics, and underlying esophageal pathology.

Patients were categorized into conservative management (CM), defined as medical treatment alone or combined with endoscopic and/or radiological interventions without surgery, and surgical management (SM), defined as immediate surgical intervention. Within the SM group, a subgroup analysis compared esophageal preservation procedures with esophagectomy.

Definitions

Treatment failure was defined as persistent or worsening sepsis, hemodynamic instability, or evidence of ongoing digestive leakage, including the presence of digestive fluid in chest drains in surgically treated patients.

Healing was defined as the resumption of oral intake in the absence of sepsis or inflammatory syndrome, with confirmation of esophageal integrity on contrast-enhanced CT.

Data collection and outcomes

Baseline variables included demographics (age, sex, body mass index), comorbidities, American Society of Anesthesiologists (ASA) score, World Health Organization (WHO) performance status, and relevant medical history. Perforation characteristics (etiology, location, associated organ injury) and treatment details were recorded.

The primary outcomes were in-hospital mortality and severe complications, defined as Clavien–Dindo grade III–IV [22]. Secondary outcomes included ICU length of stay, total hospital stay, and need for reintervention. Patients were followed during hospitalization and at days 9, 30, and 90.

Statistical analysis

Continuous variables were expressed as median (interquartile range [IQR]) and categorical variables as counts (percentages). Normality was assessed using the Kolmogorov–Smirnov test. Comparisons between CM and SM groups were performed using the chi-square or Fisher’s exact test for categorical variables and the Mann–Whitney U test for continuous variables. A multivariable logistic regression analysis was performed to identify factors associated with in-hospital mortality and severe complications (Clavien–Dindo grade III–IV). Variables included in the multivariable model were selected based on clinical relevance and a univariable screening threshold of p < 0.10. Results are reported as odds ratios (OR) with 95% confidence intervals (CI). Given the limited sample size, the number of variables included in the model was restricted to avoid overfitting. A subgroup analysis was conducted within the SM group to compare outcomes between esophageal preservation and esophagectomy. All statistical analyses were performed using SPSS version 30.0 (IBM Corp., Armonk, NY, USA). A two-sided p-value < 0.05 was considered statistically significant.

Ethical considerations

This study was conducted in accordance with French regulations for non-interventional studies (Jardé Law, Article L1121-4, modified by Ordinance No. 2016 − 800). Data collection complied with the requirements of the French Data Protection Authority (CNIL). The study was approved by the institutional data protection committee (CIL No. 2026_021).

Results

Study population

A total of 890 patients with organ perforation were identified during the study period, including 64 patients with esophageal perforation. After exclusion of patients younger than 18 years (n = 2), those initially managed at another center (n = 2), and those who died before treatment initiation (n = 2), a total of 58 patients were included in the final analysis.

Among them, 28 patients underwent surgical management (SM), while 30 patients received conservative management (CM) (Fig. 4).

Fig. 4.

Flow diagram of patient selection and treatment allocation. Patients with esophageal perforation were identified through the institutional database using ICD-10 codes. After applying exclusion criteria, patients with confirmed esophageal perforation were included in the analysis and categorized into conservative management (CM) or surgical management (SM). A subgroup analysis was performed within the SM group comparing esophageal preservation procedures and esophagectomy

Baseline characteristics

Baseline demographic and clinical characteristics were comparable between the SM and CM groups with respect to age, sex, body mass index, WHO performance status, and comorbidities (Table 1). In contrast, pre-existing esophageal disease was significantly more frequent in the CM group than in the SM group (83.3% vs. 46.4%, p = 0.003).

Table 1.

Baseline clinical and demographic characteristics of patients in the SM and CM groups

Data are expressed as median (IQR) or number of patients (%). BMI, body mass index; ASA, American Society of Anesthesiologists; WHO, World Health Organization; SM, surgical management; CM, conservative management

Clinical presentation

Patients in the SM group presented with more severe clinical conditions at admission. They were more frequently febrile (> 38.5 °C) (50% vs. 20%, p = 0.016), tachycardic (82.1% vs. 33.3%, p < 0.001), and required vasopressor support (71.4% vs. 13.3%, p < 0.001) and fluid resuscitation (85.7% vs. 26.7%, p < 0.001).

Mediastinitis was significantly more common in the SM group compared with the CM group (96.4% vs. 40%, p < 0.001). Biological parameters at admission did not differ significantly between groups (Table 2).

Table 2.

Perforation characteristics and clinical presentation in the SM and CM groups

Data are expressed as median (IQR) or number of patients (%). CRP, C-reactive protein; SM, surgical management; CM, conservative management

Perforation characteristics

The distribution of perforation location was as follows: cervical (3.4%), upper thoracic (12.1%), middle thoracic (22.4%), lower thoracic (46.6%), abdominal (10.3%), and complete esophageal necrosis (5.2%). There were no significant differences in anatomical localization between groups (Table 2).

The etiology of esophageal perforation differed significantly between groups (p = 0.0056). Spontaneous (Boerhaave) and caustic perforations were more frequent in the SM group, whereas iatrogenic perforations, particularly endoscopic, were more common in the CM group (Table 3).

Table 3.

Etiology of esophageal perforation in the SM and CM groups

Data are expressed as number of patients (%)

Diagnostic and therapeutic management

Contrast-enhanced CT was performed in 79.3% of patients overall, and upper gastrointestinal endoscopy in 51.7%. Endoscopy was more frequently performed in the CM group compared with the SM group (66.7% vs. 17.9%, p = 0.018). In the CM group, treatment consisted of endoscopic management (n = 16), medical therapy alone (n = 13), and radiological drainage (n = 1).

In the SM group, 20 patients underwent esophageal preservation procedures, including primary suture (n = 12), suture with flap (n = 1), diverticulum resection (n = 1), esophageal exclusion (n = 4), and surgical drainage (n = 2). Eight patients underwent esophagectomy. All surgical procedures were performed via thoracotomy.

The median time to treatment was 2 h (IQR 0–12), with no significant difference between groups (p = 0.58).

Outcomes

ICU and ward length of stay were comparable between groups. However, prolonged hospitalization (> 14 days) was more frequent in the SM group (82.1% vs. 36.7%, p < 0.001). Severe complications (Clavien–Dindo grade III–IV) were significantly more frequent in the SM group compared with the CM group (100% vs. 46.6%, p < 0.001), as was the need for salvage surgery (28.6% vs. 6.7%, p = 0.006). Acute kidney failure occurred more frequently in the SM group (21.4% vs. 3.3%, p = 0.035).

There was no significant difference in in-hospital mortality between the two groups (21.4% vs. 20%, p = 0.89) (Table 4).

Table 4.

Outcomes and complications in the SM and CM groups

Data are expressed as median (IQR) or number of patients (%). ICU, intensive care unit; SM, surgical management; CM, conservative management

Subgroup analysis (SM Group)

Among the 28 patients in the SM group, 8 underwent esophagectomy and 20 underwent esophageal preservation procedures. Baseline characteristics were comparable between groups.

Patients undergoing esophagectomy presented with more severe disease, including a higher rate of associated organ injury (50% vs. 5%, p = 0.005), greater need for vasopressor support (100% vs. 60%, p = 0.034), and increased fluid resuscitation requirements (100% vs. 40%, p = 0.004) (Table 5).

Table 5.

Comparison of esophagectomy and esophageal preservation in the SM group

Data are presented as number of patients (%)

There were no significant differences in complication rates, including fistula formation, acute kidney failure, venous thrombosis, or need for salvage surgery. However, mortality was significantly higher in the esophagectomy group compared with the esophageal preservation group (50% vs. 10%, p = 0.02). Multivariable logistic regression analysis did not identify independent predictors of in-hospital mortality or severe complications.

Discussion

In this retrospective single-center study of 58 patients with esophageal perforation, management strategy was primarily driven by initial clinical severity rather than etiology alone. Patients undergoing surgical management (SM) presented more frequently with mediastinitis and hemodynamic instability, whereas those managed conservatively (CM) were generally stable and more often had iatrogenic perforations. Despite higher morbidity and longer hospital stays in the SM group, overall mortality did not differ between strategies. Among surgically treated patients, esophageal preservation was feasible in most cases and was associated with lower mortality than esophagectomy, which was reserved for the most severe presentations. These findings support a severity-based, individualized approach to management.

Outcomes: mortality and morbidity

The overall mortality rate in our cohort was 21%, consistent with previously reported rates of 10–40% [3–7, 23, 24]. Prognosis in esophageal perforation is largely determined by early diagnosis and rapid control of contamination, with mediastinitis, septic shock, and delayed treatment consistently identified as major risk factors [3, 4, 11, 16, 25–33].

In our study, patients in the SM group had significantly higher rates of mediastinitis and hemodynamic instability, reflecting more advanced disease at presentation. This likely explains both the need for surgical intervention and the higher rate of severe complications observed in this group. Similarly, prolonged hospitalization and acute kidney injury were more frequent in SM patients, likely reflecting the systemic impact of sepsis and organ dysfunction rather than surgical treatment alone and consistent with the literature [34–36].

As in previous studies [26, 37–41], salvage procedures were required in both groups at comparable rates, although for different indications—persistent sepsis or necrosis in SM, and persistent fistula in CM. Importantly, secondary interventions were not associated with increased mortality in the CM group, supporting the safety of an initial conservative approach in selected patients.

Management strategies

Management of esophageal perforation depends on hemodynamic status, extent of contamination, etiology, and delay to treatment. Delayed diagnosis remains a key determinant of poor outcomes, particularly in spontaneous perforations [17].

Conservative and endoscopic management are increasingly used in stable patients, especially in early iatrogenic perforations. These approaches, combining antibiotics, drainage, and endoscopic techniques such as clips, stents, or endoscopic vacuum therapy, have demonstrated high success rates with low morbidity [37–39, 42, 43]. Endoscopic stenting can control leaks even in delayed cases, although it may increase the need for additional procedures [44]. Importantly, salvage surgery remains feasible after failed endoscopic treatment without adversely affecting outcomes [45].

Surgery remains essential in patients with hemodynamic instability or extensive contamination [38, 46, 47]. Primary repair is preferred when feasible and is associated with acceptable mortality (9–11% at 30 days) [1, 23, 24]. Esophagectomy is reserved for the most severe cases, including extensive necrosis or malignant perforation, and carries substantially higher morbidity and mortality (23% at 90 days) [13, 46]. In our cohort, the higher mortality observed after esophagectomy likely reflects the severity of the underlying condition rather than the procedure itself.

Certain etiologies, such as Boerhaave syndrome, malignant perforation and caustic injury, are frequently associated with extensive mediastinal contamination requiring surgical management [11, 13, 37, 38, 42, 43, 47]. Caustic injuries, in particular, may involve diffuse necrosis and necessitate complex resections, contributing to high mortality rates (up to 37%) [48–51].

Role of underlying esophageal disease

Underlying esophageal disease alone should not dictate management strategy. In our cohort, conditions such as achalasia and Zenker’s diverticulum were successfully managed conservatively following iatrogenic perforations, so as in previous study [52–54]. Esophagectomy should be considered only in advanced disease or in cases of extensive or complicated perforation [44, 55–58]. These findings suggest that initial management should primarily be guided by clinical stability and extent of contamination rather than underlying pathology.

Clinical implications

Our findings emphasize the importance of early multidisciplinary evaluation and rapid risk stratification. Hemodynamic instability, septic shock, or mediastinitis should prompt immediate surgical intervention, whereas stable patients with limited contamination may be safely managed conservatively.

Based on these findings and existing literature, we propose a pragmatic decision-making algorithm integrating physiological severity, etiology, and extent of contamination (Fig. 5). Such an approach may facilitate early therapeutic decisions and improve consistency of care. Prospective multicenter studies are needed to validate this strategy.

Fig. 5.

Proposed pragmatic decision framework for the management of esophageal perforation. EVT: Endoscopic Vacuum Therapy

Limitations and strengths

This study has several limitations. First, its retrospective single-center design introduces a risk of selection bias, particularly as treatment allocation was based on clinical severity rather than randomization, precluding any causal inference between treatment strategy and outcomes. Second, the relatively small sample size limits the statistical power of subgroup analyses, especially within the surgical cohort. Third, long-term functional and nutritional outcomes were not systematically assessed, restricting evaluation of recovery beyond the acute phase.

However, this study also has important strengths. It provides a detailed real-world analysis of esophageal perforation within a multidisciplinary setting, with comprehensive clinical, radiological, and therapeutic data. Importantly, it highlights the central role of initial clinical severity in guiding management decisions and offers a pragmatic, clinically applicable decision framework. The consistency of findings with existing literature further supports the external validity of our results.

Conclusion

Esophageal perforation is a rare but life-threatening condition requiring prompt multidisciplinary management. Treatment should be guided primarily by initial clinical severity and the extent of contamination. Conservative and endoscopic approaches are effective in stable patients with limited contamination, whereas surgery remains mandatory in unstable patients or in the presence of mediastinitis or extensive necrosis. When surgical intervention is required, esophageal preservation should be favored whenever feasible, as it is associated with improved outcomes.

Abbreviations

CM

Conservative management

SM

Surgical management

ICU

Intensive care unit

CT

Computed tomography

CRP

C-reactive protein

BMI

Body mass index

ASA

American Society of Anesthesiologists (score)

WHO

World Health Organization (performance status)

IQR

Interquartile range

EVT

Endoscopic vacuum therapy

Author contributions

Clara Poon contributed to the initial drafting of the manuscript and the preparation of tables and figures.Nicolas Michot contributed to manuscript revision, statistical analysis, and figure preparation.Urs Giger-Pabst contributed to manuscript revision and English language editing.Driffa Moussata, Ephrem Salamé, and Mehdi Ouaissi contributed to study supervision and coordination.All authors reviewed and approved the final version of the manuscript.

Funding

This study received no external funding.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with applicable regulations for retrospective studies. Formal ethics approval and informed consent were not required in accordance with national legislation. The study was approved by the institutional data protection committee (CIL No. 2026_021). Other clinical trial number: not apllicable.

Consent for publication

Not applicable.

Use of Artificial Intelligence and language tools

During the preparation of this manuscript, the authors used ChatGPT and DeepL Write to assist with language editing, text optimization, and structural refinement. All content was critically reviewed and approved by the authors, who take full responsibility for the final version of the manuscript.

Competing interests

The authors declare no competing interests.