Management and outcomes of acquired benign tracheoesophageal fistula: a single-centre experience of 58 cases

Introduction

Tracheoesophageal fistula (TEF) refers to an uncommon pathologic connection between the oesophagus and the trachea. The most common aetiologic factor of acquired benign TEF is pressure necrosis and ischaemia from prolonged intubation, which are often accompanied by an indwelling nasogastric tube (1). Other less common causes are trauma and complications of oesophageal operations (2). The majority of patients experience coughing after swallowing, which contributes to malnutrition and further exacerbates their condition (3). TEF can also coexist with tracheal stenosis, making it even more challenging (4). Spontaneous closure is rare, and this condition is life-threatening; therefore, intervention is always needed (5). Various surgical treatments are possible; however, single-stage two-layer oesophageal closure with tracheal resection is the gold standard (3,6). Nonsurgical approaches have also been described, although they are more effective for treating malignant or small fistulas (1,3). Morbidities greater than 50% have been described (2,7); therefore, the identification of risk factors for surgical outcomes is essential. Because of the extensive preparation required and the complexity of the condition, surgical treatment is practised only in experienced centres. Given the rarity of TEF, there is a limited amount of data. We present a single-centre experience in one of the largest studies of 58 patients spanning between 1998 and 2023. This study aimed to provide additional valuable clinical data and identify risk factors related to the surgical treatment of TEF. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-399/rc).

Methods

Ethics statement

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of the National Research Institute of Chest Diseases, Warsaw, Poland (96/2021). Formal consent was waived because this was a retrospective study.

Patient population

From January 1998 to December 2023, patients who were diagnosed with TEF and who underwent oesophageal closure with tracheal segmental resection at the National Research Institute of Chest Diseases were included. Patients with congenital and malignant fistulas were excluded.

Patient demographic data, radiographic, bronchoscopic, and gastroscopic images, operation records, and postoperative results were obtained from the patient registry and medical records. Each patient was examined using computed tomography and/or roentgenogram. Before the operation, fiberoptic bronchoscopy or rigid bronchoscopy and gastroscopy were performed. During bronchoscopy, each patient’s tracheal lumen culture was collected, and if the culture was positive, the patient was treated with antibiotics until a negative culture was obtained. Additionally, the resected segment underwent microbiologic examination.

Patient outcomes were divided into short-term and long-term outcomes and into good, medium, and poor outcomes. Short-term outcomes were defined as those within 6 months, and long-term outcomes were defined as those over one year. A good outcome was defined as a lack of complications. A medium outcome was defined as having minor complications that did not affect quality of life, such as pneumonia, wound infection or evidence of narrowing in bronchoscopy. A poor outcome was defined as severe complications such as vocal cord paralysis, anastomosis dehiscence, recurrence of tracheal stenosis or TEF requiring stenting or reoperation.

Surgical technique

Intubation was performed using a bronchofiberscope, which allowed the balloon of the endotracheal tube to be located below the fistula. In patients with tracheostomies and obstructed proximal tracheal sections, ventilation in the first stage was performed by ventilating patients using a wire-reinforced tube inserted through the tracheostomy hole. The anterior wall of the trachea in the mediastinum was mobilized to the level of the tracheal bifurcation and cephalically mobilized to the level of the hyoid cartilage. The trachea was transversely cut below the stenosis and fistula. The trachea was subsequently cut and separated from the oesophagus at the site of the fistula, and the scarred proximal section of the stenosis was resected. From this point onward, ventilation was performed from the surgical field, and a wire-reinforced tube was inserted into the distal section of the trachea. After sufficient mobilization of the trachea was achieved to perform the anastomosis without excessive tension, two traction sutures were applied on the lateral walls of both ends at the point of connection (including the membranous and cartilaginous parts on both sides of the trachea). The membranous trachea was sutured with a running 4-0 polydioxanone (PDS, Ethicon Inc., Somerville, NJ, USA) suture, whereas the anterolateral ends were anastomosed with interrupted 3-0 Vicryl sutures (Ethicon Inc., Cornelia, GA, USA). When the membranous part of both ends of the trachea was anastomosed, the endotracheal tube was removed from the surgical field, and the patient was intubated via the larynx, as is typically performed. After separation of the trachea and oesophagus, the oesophagus was sutured in two layers with single 3-0 Vicryl sutures. Conversely, the trachea was resected along with the fistula and the narrowed section and/or the tracheostomy foramen. In some cases, the pedicled sternohyoid muscle flap was interposed between the oesophagus and the airway.

Statistical analysis

The study included 58 patients who underwent surgery between 1998 and 2023. The analysis of complications was performed using multivariate logistic regression with model selection on the basis of the Akaike information criterion. Most factors were categorical variables, and only a few were continuous variables (e.g., the length of resection). The differences between groups were examined by including the group as a factor in the logistic regression. The significant factors were determined using the Wald test, where a P<0.05 was considered statistically significant. All analyses were performed using R software version 4.4.

Results

Patient and surgical characteristics

The study included 58 patients, 31 men (53%) and 27 women (47%), with a mean age of 46.5 years (ranging from 19 to 81 years). The most common causes of intubation and/or tracheostomy leading to TEF were respiratory distress (69%), general anaesthesia associated with surgery (5%), trauma (5%), and sudden cardiac arrest (3%) (Table 1). The mean intubation time was 20.8 days (range: 8–63 days). A total of 41 (71%) patients had a history of tracheostomy, and its mean duration was 24 days (range: 18–63 days), whereas for patients without tracheostomy, the mean duration was 17.6 days (range: 8–23 days). One patient had tracheal stents while awaiting the procedure (2%). We collected cultures from the tracheal lumen before the operation, which were positive in 36 patients (62%) (Table 2).

In the majority of cases (76%), the surgical access was cervicotomy; eight patients required additional partial sternotomy (14%), and in one case, only sternotomy was needed. In 37 patients (64%), the TEFs were located in the lower part of the trachea, 16 in the middle part (28%), and 5 in the upper part (8%). The mean length of the TEF was 22 mm (5–45 mm), and the median length was 34 mm. Tracheal stenosis and resection length varied between 0 (2%) and 4–6 cm (9%). The most common lengths were 1–2 cm (31%), 2–3 cm (31%), and 3–4 cm (12%). Cricoid cartilage resection was required in 4 patients (7%) (Table 2).

Outcomes and complications

Owing to excessive tension at the anastomosis site, tracheostomy was performed in 43% of the patients. Two patients (4%) required straight stents, and 4 (7%) required a T-tube after the operation. The long-term mortality rate was 3%. The most common complication was vocal cord paralysis (14%). Dehiscence of the anastomosis led to recurrence of the fistula, and resection was observed in 4 patients (7%) (Table 3). The usual length of hospitalization was 15–30 days (33%), and the mean time from surgery to discharge was 17 days.

Forty-four patients (76%) had good short-term outcomes, 7 (12%) had medium outcomes, and 6 (10%) had poor outcomes. Good long-term outcomes were observed in 76% of the patients, medium long-term outcomes in 9%, and poor long-term outcomes in 9%. Two patients died (3%) (Table 3).

Multivariate analysis

According to the multivariate analysis, the risk factor for poor short-term surgical outcome and severe complications was the resection length greater than 3 cm [odds ratio (OR) 1.78; 95% confidence interval (CI) 1.12–3.67; P=0.004]. No studied factor was statistically significant for long-term results or recurrence of stenosis and resection (Table 4).

Discussion

TEF is a complex therapeutic challenge. TEF often coexists with tracheal stenosis, which itself is a problem that requires multistage treatment. Therefore, the treatment of both TEF and tracheal stenosis is conducted at selective institutions with experience. Our study with 58 patients is one of the largest thus far. Because the treatment of choice is surgical intervention, our research focused on management and surgical outcomes.

The reported incidence of TEF varies between 0.5% and 3.8% (5,8). The recurrence of TEF is not uncommon, and its rate can reach 11.1% (7).

When it is acquired and benign, the usual aetiologic factors are high endotracheal cuff pressure, prolonged intubation, an indwelling nasogastric tube, and trauma (1,9). These factors contribute to pressure, ischaemia, and necrosis of the tracheal wall. The ischaemic component and coexistent infection weaken the tracheal and oesophageal wall and make it even more difficult to perform definitive repair. Even though new low-pressure, high-volume cuffs have been developed, approximately 75% of acquired benign TEFs are iatrogenic. On the other hand, Shen et al. reported that the most common causes were complications of oesophageal surgery, granulomatous infection, trauma, and stent erosion (2). Comorbidities such as diabetes, the use of steroids, malnutrition, or coronavirus disease 2019 (COVID-19) may significantly increase the risk of favourable outcomes (1,9).

The prolonged intubation was the main contributor to TEF formation, independent of other risk factors. In the presented study, the mean time of intubation was 17.6 days, and in the case of tracheostomy, it was 24 days. The main cause of prolonged intubation was respiratory distress (70% of cases). The mean size of the fistula was 22 mm (median 34 mm) according to endoscopic and radiologic examination. In the majority of instances, tracheal stenosis measuring 1 to 3 cm accompanied TEF (62% of cases). In extreme cases, the length of the stenosis reached up to 6 cm (9% of cases). Given the nature of this condition, the oesophageal fluid constantly contaminates the airway lumen. During the preoperative period, obtaining the bronchial tree’s bacteriological purity is key. In our study, up to 62% of patients had positive tracheal lumen culture. Those patients required targeted antibiotic therapy and regular bronchoscopic suctioning, which resulted in a prolonged preoperative period.

The ideal course of action has been debated (10). Although stenting is currently the most viable option for treating malignant TEFs, it serves as a bridge to definitive treatment for patients with benign TEFs (3). Spontaneous closure has been noted but is relatively rare (9,11). Therefore, a surgical attempt at closure should be considered in the case of TEF. The technique is dependent upon the preference, location, and size of the TEF. Fistulas resulting from prolonged intubation are most likely associated with tracheal stenosis (8). Therefore, Grillo et al. suggested anterior cervicotomy with single-stage tracheal resection and two-layer oesophageal closure (12). It provides an excellent visual field and reach, which minimises morbidity and provides an option for simultaneous resection of stenoses (11). This approach is considered the gold standard, and it is practised at our institution (6). Surgical access was predominantly cervical (76%), followed by cervicotomy with partial sternotomy (14%), which is consistent with the literature (5,7). No thoracotomy was needed, in contrast with Shen et al., who reported thoracotomy as the main approach (2). The fistulas in their study were caused primarily by oesophageal surgery, which may explain this discrepancy. Whether a muscle flap between the oesophageal and tracheal suture lines is necessary is disputed. The literature provides many successful examples (7); however, Camargo et al. argued that it is dispensable (10). Additionally, the oversized flap may compress the trachea, resulting in pressure necrosis and secondary tracheal stenosis (11). In our study, a sternohyoid muscle flap was used in some cases, and the recurrence rate was similar to or lower than that reported in the literature (2,7).

The complication rates described in the literature vary greatly. On the one hand, Bibas et al. reported a morbidity of 55% (4). On the other hand, Marulli et al. and Macchiarini et al. noted complication rates of 20% and 22%, respectively (5,11). Among other complications, pneumonia, dehiscence of the anastomosis, and vocal cord paralysis are mentioned (4,5). We noted that the most common complications were vocal cord paralysis (14%) and dehiscence (7%). The palsy rate was higher than that reported in the literature (2,5), which may be because all of our patients required tracheal segmental resection and had larger fistulas on average than those previously reported. On the other hand, the recurrence rate of TEF in our study was similar to or lower than that reported in the largest studies (7,13). Patients with TEF risk aspiration of oesophageal fluid into the lungs, which may cause chronic pneumonia, atelectasis, obstruction, and distress (5). This further worsens the patient’s cardiopulmonary functions. Approximately 80% of patients experience Ono’s sign, which is coughing after swallowing substances. This leads to malnutrition, which is one of the greatest risk factors for unsatisfactory outcomes (3). One of the reasons for surgical failure is inflammation at the site of anastomosis. When performing tracheal segmental resection, improper healing, restenosis, and the need for resection can occur. The foreign bacterial flora in the trachea could contribute to pathological healing; therefore, we conducted preoperative culture collection. Although the influence of the microbiome on TEF in adults has not been previously described, the literature suggests its role in pathological tracheal healing (14). Mattos et al. noted that bacterial cultures were similar to those of children without TEF (15). In our study, bacterial culture was not a significant risk factor.

Data concerning the intubation time before TEF are scarce, although intubation times are generally prolonged. One study reported a mean intubation time of 23 days. In our study, the mean time was 20.8 days. Given that one of the indications for tracheostomy is prolonged ventilation, the intubation time was longer in patients with tracheostomy (24 days). Moreover, tracheostomy causes damage to the tracheal wall, and combined with pressure necrosis, it is not associated with a lower risk of TEF (1). Given that almost every patient in our database had a history of intubation, the majority of TEFs were located in the lower (64%) and middle (28%) parts of the trachea. An additional risk factor may be prolonged insertion of a gastric tube, the hard wall of which constitutes resistance to the compressive balloon of the endotracheal tube, causing long-term compression of the tissues located between both tubes and resulting in chronic ischaemia and necrosis. The location of the fistula was confirmed via computed tomography (CT) or roentgenogram (RTG) and during diagnostic bronchoscopy and gastroscopy. Although Kim et al. (1) reported that iatrogenic TEF occurs in the middle and distal trachea, the literature reports are inconsistent in this regard (4,7). Given the span of years, some of our data may be incomplete in determining the exact underlying cause of TEF. The size of the TEF is another important factor determining the therapeutic approach. Fibrin glue was successfully used by Scappaticci et al. in the closure of TEFs smaller than 5 mm (16). The use of atrial septal occlusion has also been described (1). However, we observed fistulas reaching 45 mm; therefore, each patient was treated surgically. Compared with those reported in the literature, the TEFs in our study were larger on average (2,4). An analysis conducted by Shen et al. revealed that increased size is not associated with increased morbidity, the need for reoperation, or the length of hospitalization. However, the concomitant circumferential damage to the trachea makes segmental resection with reconstruction more likely to be required (2).

Placement of a T-tube was required in 7% of the patients in this study, which is lower than that reported by Puma or Camargo (6,10). The T-tube can be removed after a long time, allowing the trachea the necessary time to regenerate. Moreover, the T-tube does not exert pressure on the trachea, which is the leading cause of TEF and stenosis (6). Multivariate analysis revealed that the only independent risk factor for poor short-term outcomes after anastomosis was resection of a tracheal segment exceeding 3 cm (OR 1.78; 95% CI: 1.12–3.67; P=0.004).

This study has several limitations. First, this was a retrospective, single-centre study that covered a large number of years (1998–2023), which had an impact on the lack of data. However, because it is the largest study published thus far, it has great clinical significance. There are many conservative and surgical treatments for TEF; however, we focused on only one of them to compare outcomes and risk factors. We did not include patients with malignant or congenital fistulas because they may influence the data, and we wanted to focus on one specific condition and its treatment.

Conclusions

Acquired benign TEF with tracheal stenosis is a complex and life-threatening condition that should be treated in experienced centres. Single-stage oesophageal occlusion with segmental tracheal resection is a highly effective and safe therapeutic method that is a definitive treatment in most cases. The T-tube placement after a procedure is required only in a minority of patients. The length of the resected tracheal segment is an independent prognostic factor for poor short-term outcomes and severe postoperative complications, including fistula recurrence.

Acknowledgments

The Abstract was presented at the 12th Congress of the Polish Society of Cardio-Thoracic Surgeons, Krakow, Poland, 26–28th April 2025.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-399/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-399/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-399/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-399/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional ethics board of the National Research Institute of Chest Diseases, Warsaw, Poland (96/2021), and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Kim HS, Khemasuwan D, Diaz-Mendoza J, et al. Management of tracheo-oesophageal fistula in adults. Eur Respir Rev 2020;29:200094. [Crossref] [PubMed]
2. Shen KR, Allen MS, Cassivi SD, et al. Surgical management of acquired nonmalignant tracheoesophageal and bronchoesophageal fistulae. Ann Thorac Surg 2010;90:914-8; discussion 919. [Crossref] [PubMed]
3. Zeng A, Liu X, Shaik MS, et al. Surgical strategies for benign acquired tracheoesophageal fistula. Eur J Cardiothorac Surg 2024;65:ezae047. [Crossref] [PubMed]
4. Bibas BJ, Guerreiro Cardoso PF, Minamoto H, et al. Surgical Management of Benign Acquired Tracheoesophageal Fistulas: A Ten-Year Experience. Ann Thorac Surg 2016;102:1081-7. [Crossref] [PubMed]
5. Marulli G, Loizzi M, Cardillo G, et al. Early and late outcome after surgical treatment of acquired non-malignant tracheo-oesophageal fistulae. Eur J Cardiothorac Surg 2013;43:e155-61. [Crossref] [PubMed]
6. Puma F, Vannucci J, Santoprete S, et al. Surgery and perioperative management for post-intubation tracheoesophageal fistula: case series analysis. J Thorac Dis 2017;9:278-86. [Crossref] [PubMed]
7. Muniappan A, Wain JC, Wright CD, et al. Surgical treatment of nonmalignant tracheoesophageal fistula: a thirty-five year experience. Ann Thorac Surg 2013;95:1141-6. [Crossref] [PubMed]
8. Fiala P, Cernohorský S, Cermák J, et al. Tracheal stenosis complicated with tracheoesophageal fistula. Eur J Cardiothorac Surg 2004;25:127-30. [Crossref] [PubMed]
9. Walker KN, Carlson KJ, Rubinstein BJ, et al. Tracheoesophageal Fistula as a Complication of Prolonged Ventilation in COVID-19: Description of Reconstruction and Review of the Literature. Ear Nose Throat J 2024;103:120S-4S. [Crossref] [PubMed]
10. Camargo JJ, Machuca TN, Camargo SM, et al. Surgical treatment of benign tracheo-oesophageal fistulas with tracheal resection and oesophageal primary closure: is the muscle flap really necessary? Eur J Cardiothorac Surg 2010;37:576-80. [Crossref] [PubMed]
11. Macchiarini P, Verhoye JP, Chapelier A, et al. Evaluation and outcome of different surgical techniques for postintubation tracheoesophageal fistulas. J Thorac Cardiovasc Surg 2000;119:268-76. [Crossref] [PubMed]
12. Grillo HC, Moncure AC, McEnany MT. Repair of inflammatory tracheoesophageal fistula. Ann Thorac Surg 1976;22:112-9. [Crossref] [PubMed]
13. Mathisen DJ, Grillo HC, Wain JC, et al. Management of acquired nonmalignant tracheoesophageal fistula. Ann Thorac Surg 1991;52:759-65. [Crossref] [PubMed]
14. Carpenter DJ, Hamdi OA, Finberg AM, et al. Laryngotracheal stenosis: Mechanistic review. Head Neck 2022;44:1948-60. [Crossref] [PubMed]
15. Mattos C, Phinizy P, Duffy KL, et al. Clinical predictors of laryngotracheoesophageal clefts and tracheoesophageal fistulae in children with dysphagia. Pediatr Pulmonol 2021;56:3792-5. [Crossref] [PubMed]
16. Scappaticci E, Ardissone F, Baldi S, et al. Closure of an iatrogenic tracheo-esophageal fistula with bronchoscopic gluing in a mechanically ventilated adult patient. Ann Thorac Surg 2004;77:328-9. [Crossref] [PubMed]