ABSTRACT
Background
Adult, benign tracheoesophageal fistula is a life‐threatening condition. The aim of this study is to report our experience with its surgical treatment, including short‐ and long‐term outcomes.
Methods
We conducted a retrospective, single‐center review of patients who underwent surgical repair of tracheoesophageal fistula at a single center from 1997 to 2023. We compared clinical characteristics and outcomes between patients operated on by different surgical techniques. Furthermore, we analyzed long‐term outcomes including survival, decannulation rate, and resumption of oral intake.
Results
Overall, 60 patients underwent fistula repair during the study period. The main etiology was post‐intubation injury (78.3%). The airway defect was repaired by direct suture, airway resection‐anastomosis or flap/patch interposition in 8 (13.3%), 30 (50.0%) and 22 (36.7%) cases, respectively. The esophageal defect was repaired by double‐layered suture in 56 (93.3%) cases and esophageal diversion in 4 (6.7%). Morbidity was observed in 32 patients (53.3%), including fistula relapse in 2 cases. In‐hospital mortality occurred in two patients (3.3%). The overall survival, the cumulative incidence of decannulation, and of resumption of oral feeding at 5 years were 70.5%, 68.4%, and 86.6%, respectively.
Conclusions
Surgical correction of tracheoesophageal fistula is challenging and requires an individualized approach. Healing of the fistula can be achieved in the majority of cases, including patients with extensive defects or under mechanical ventilation. Long‐term outcomes are less satisfactory, and seem to be determined mainly by patients' comorbidities.
Keywords: post‐intubation injury, resorbable patches, tracheoesophageal fistula
We conducted a retrospective, single‐center review of patients who underwent surgical repair of tracheoesophageal fistula at a single center from 1997 to 2023. We compared clinical characteristics and outcomes between patients operated on by different surgical techniques. Furthermore, we analyzed long‐term outcomes including survival, decannulation rate, and resumption of oral intake.
1. Introduction
Adult, benign tracheoesophageal fistula (TEF) is a life‐threatening pathology, which interferes with nutrition and causes infective complications [1, 2]. Its main causes include post‐intubation injury and iatrogenic damage, particularly following surgical procedures such as esophagectomy or laryngectomy, or endoscopic procedure such as self‐expandable metallic stent (SEMS) placement [3, 4].
Because this pathology is rare, only few high volume centers have extensive experience in its surgical treatment. Consequently, there's a relatively limited number of case‐series presenting early and late outcomes after surgery for TEF [1, 2, 3, 5, 6, 7, 8]. The few available surgical case series report rates of fistula recurrence that range between 3.3% and 11% [3, 5, 6, 7]. However, follow‐up data—including survival and long‐term functional outcomes, such as resumption of oral intake or breathing without an airway appliance—are inconsistently reported and expressed using heterogeneous metrics (e.g., crude rates among operative survivors or among long‐term survivors), thereby limiting comparability.
Over time, different surgical techniques have been proposed for repair of TEF, and in particular, for closure of the airway side of the fistula. Direct repair of the membranous trachea or tracheal resection/anastomosis represent the two traditional surgical options [9]; while for more challenging cases, the use of muscle flaps [10], esophageal flaps [8], or synthetic prostheses [11, 12], have been described. Unfortunately, because of the limited number of included subjects, very few studies have attempted to compare the characteristics and outcomes among patients undergoing different types of airway repair [6].
The aims of this study are the following. First, to analyze our institutional experience in the surgical treatment of TEF. Second, to explore the associations between patient characteristics, surgical techniques, and clinical outcomes. Third and to analyze long‐term functional outcomes, including decannulation and resumption of oral feeding.
2. Patients and Methods
2.1. Patients and Variables
The institutional review board of Padua University Hospital approved this study (protocol number 271n/AO/22) on 30 June 2022, and the need for individual patient consent was waived.
The records of all consecutive patients undergoing surgical treatment for TEF at the Thoracic Surgery Unit of Padua University Hospital from January 1997 to December 2023 were retrospectively reviewed. Patients with malignant fistulae, those treated non‐surgically, and those with bronchoesophageal fistulae (e.g., fistulae located more caudal than the carina) were excluded.
2.2. Preoperative Management
Upon clinical suspicion, the diagnosis of TEF was always confirmed by bronchoscopy and esophagoscopy. When feasible, rigid bronchoscopy was performed in order to better define the site of the fistula and the presence of any associated airway injury. In every case, the endoscopic examination retrieved the following information: distance of the fistula from the cricoid cartilage, its longitudinal diameter, and the distance from the carina (i.e., the total tracheal length). Computed tomography (CT) was routinely performed to provide additional information about mediastinal and pulmonary infection and damage. Preoperative management included adequate nutritional support with enteral feeding, targeted antibiotic therapy for infection control, and weaning from mechanical ventilation. When necessary, the endotracheal or tracheostomy tube was repositioned with the cuff distal to the fistula. Whenever feasible, surgery was deferred until infection was controlled and adequate nutritional and respiratory status had been achieved. However, in cases where no further improvement could reasonably be expected without fistula repair, surgery was undertaken as a salvage measure despite persistent infection or ongoing need for mechanical ventilation.
2.3. Surgical Strategy
The surgical approach was selected based on fistula location, patient body habitus, and the planned repair technique. Generally, cervicotomy, median sternotomy, and right thoracotomy were used for fistulae in the upper, middle, and lower trachea, respectively, using the sternal notch/body of T1 and the dome of the aortic arch as landmarks. However, obese patients or those with a short neck and limited extension were more likely to require a sternotomy also for fistulae more cranial than the sternal notch; while, conversely, lean patients with a good cervical extension could be approached by cervicotomy for more caudal defects. In uncertain cases, surgery started with a cervicotomy, with the option to extend to a sternal split or full sternotomy. More recently, in some patients with large fistulae involving both the intra‐ and extrathoracic trachea, (extending across the level of the sternal notch/body of T1), we adopted as a surgical approach a posterior right thoracotomy in the 3rd intercostal space, with or without resection of the 3rd and 2nd ribs [13]. While anterior approaches (e.g., cervicotomy and sternotomy) are well suited for segmental airway resection‐anastomosis, this posterior approach was preferred for non‐circumferential types of airway repair.
After division of the fistulous tract, the airway was repaired by one of the three techniques illustrated in Figure 1. The choice of procedure was based not only on fistula size, but also on its etiology, location, and the patient's general condition. Small fistulae (< 1 cm) without associated tracheal stenosis could be repaired by direct suture with interrupted 4/0 polydioxanone (PDS; Ethicon, Sommerville, NJ). Tracheal resection and anastomosis (Figure 1B) was adopted more commonly, especially for post‐intubation fistulae, regardless of the presence of overt airway stenosis, since cuff injury is expected to always create some circumferential tracheal damage [6]. In patients with large airway defects (> 3 cm), especially when associated increased technical complexity (e.g., low tracheal location), and/or with conditions increasing the risk of anastomotic failure (e.g., mechanical ventilation, septic syndrome, diabetes, malnutrition), bridging the defect by interposition of either a muscle flap or an absorbable patch (Figure 1C), was often preferred to segmental resection anastomosis, as previously described [11]. Pedicled flaps were also used in patients undergoing other types of airway repair, serving both as a buttress for the airway anastomosis and as a barrier between the trachea and the esophagus. Their use was considered particularly important in larger fistulae, when the suture lines of the tracheal and esophageal repairs had substantial overlap, whereas in cases of smaller fistulae they were not always required. The intercostal muscle flap was favored for intrathoracic fistulae while the sternocleidomastoid or the strap muscles were used for cervical fistulae. A tracheostomy was performed in cases of extensive tracheal resection to reduce tension on the airway anastomosis; in patients with compromised respiratory or neurological status who might require positive‐pressure ventilation; and in those with glottic pathology associated with a risk of aspiration. The esophageal defect was directly closed whenever possible in two layers (mucosal and muscular layers) with running 4‐0 PDS sutures. In the presence of extensive esophageal damage, the affected esophageal segment was resected down to the level of the diaphragm (esophagectomy) and a left cervical end‐esophagostomy was performed (esophageal diversion). In the postoperative period, extubation was performed as soon as possible.
FIGURE 1.
Illustration showing surgical options for closure of the airway side of the defect during repair of tracheoesophageal fistula: direct suture (A), tracheal resection and anastomosis (B), interposition of resorbable patches (or muscle flaps) (C).
2.4. Postoperative Management
Flexible bronchoscopies were scheduled at postoperative day 1 and day 7 to check for healing of the airway. Additional bronchoscopies were performed if clinical indicated. The integrity of the esophageal suture was assessed by oral contrast radiography, normally on postoperative day 7, after which, in the absence of contraindications, oral intake could be resumed. Laryngoscopy was performed to assess vocal cords' and swallowing function, and speech or swallow rehabilitation was initiated when clinically indicated. Patients with a postoperative tracheostomy, in the absence of glottic pathology, were managed with a T‐tube as soon as possible, and followed up periodically. Decannulation was considered on an individual basis, when the general conditions and the status of the trachea permitted (e.g., absence of malacia or stenosis). Likewise, for patients who underwent esophageal diversion, reconstruction was considered, during follow‐up visits, if the general conditions permitted.
2.5. Definitions
Septic syndrome, malnutrition, and Charlson Comorbidity Index (CCI) were defined according to established algorithms [14, 15, 16]. The diameter of the fistula was retrieved by the measurements performed during the preoperative endoscopy, while its location along the airway (upper, middle, or lower third of the trachea) was determined on preoperative CT scans. The trachea was divided into thirds using transverse planes at the level of the sternal notch/body of T1 and the dome of the aortic arch as anatomical landmarks.
2.6. Statistical Analyses
Numerical data are presented as median and interquartile range, while categorical data are presented as absolute number and percentage. Difference between patients' subgroups were assessed by Wilcoxon rank‐sum test or chi‐squared test, as appropriate. Overall survival was estimated with the Kaplan‐Meier method. The cumulative incidence function was used to estimate the rate of decannulation and the rate of resumption of oral feeding over time, taking into account the competitive risk of death. All analyses were performed with the statistical software R (R Foundation for Statistical Computing, Vienna, Austria).
3. Results
3.1. Patient's Characteristics
The cohort consisted of 60 patients, the majority of whom were referred from other hospitals (53/60), while a smaller proportion were referred from other units within our center (7/60). Main characteristics are reported in Table 1. The most common etiology was post‐intubation injury (n = 47, 77.8%), followed by complication of surgical or endoscopic procedures (n = 8, 13.3%). Endoscopic procedures consisted in esophageal (n = 2) or tracheal (n = 2) self‐expandable metallic stent placement; while surgical procedures consisted in esophagectomy for esophageal cancer (n = 2), esophageal myotomy (n = 1) and cervical spine fixation (n = 1). On preoperative endoscopic assessment, the median diameter of the fistula, as measured from the airway side, was 19 mm (interquartile range (IQR): 10–26 mm).
TABLE 1.
Baseline and perioperative patients' characteristics.
| Baseline characteristic | N | Perioperative characteristics | N |
|---|---|---|---|
| Age | 60 (50–69) | Surgical approach | |
| Sex | Cervicotomy | 42 (70.0%) | |
| Male | 29 (48.3%) | Thoracotomy | 13 (21.7%) |
| Female | 31 (51.7%) | Cervicotomy + sternotomy | 2 (3.3%) |
| Charlson comorbidity index | 3.5 (2–5) | Cervicotomy + thoracotomy | 1 (1.7%) |
| Diabetes | 15 (26.8%) | Sternotomy | 1 (1.7%) |
| Arterial hypertension | 18 (32.1%) | Cervicotomy + sternal split | 1 (1.7%) |
| Peripheral vascular disease | 14 (24.6%) | Airway repair procedure | |
| Cardiac disease | 14 (24.6%) | Direct closure | 8 (13.3%) |
| Stroke | 9 (16.1%) | Resection/anastomosis | 30 (50.0%) |
| Paraplegia/hemiplegia | 10 (17.9%) | Flap or patch interposition | 22 (36.7%) |
| Neurologic impairment | 12 (21.1%) | Esophageal repair procedure | |
| History of cancer | 8 (13.6%) | Double‐layered suture | 56 (93.3%) |
| Body mass index (kg/m2) | 22.5 (19.4–25.5) | Esophageal diversion | 4 (6.7%) |
| Malnutrition | 9 (15.3%) | Tracheostomy | 34 (56.7%) |
| Obesity (BMI > 25 kg/m2) | 8 (13.6%) | Pedicled muscle flap | 46 (76.7%) |
| Septic syndrome | 25 (43.1%) | Sternocleidomastoid muscle | 17 (28.3%) |
| Tracheostomy on admission | 40 (66.7%) | Strap muscles | 14 (23.3%) |
| Diameter of the fistula (mm) | 19 (10–26) | Intercostal muscle | 12 (20.0%) |
| Associated tracheal stenosis | 12 (20.0%) | Thymus | 2 (3.3%) |
| Etiology of the fistula | Serratus anterior | 1 (1.7%) | |
| Post‐intubation | 47 (78.3%) | Intensive care unit stay (days) | 3 (1–6) |
| Post‐surgical/endoscopic | 8 (13.3%) | Mechanical ventilation (days) | 0 (0–1) |
| Congenital | 3 (5.0%) | Morbidity | 32 (53.3%) |
| Other | 2 (3.4%) | Mortality | 2 (3.3%) |
| Previous surgical repair | 10 (16.7%) | Postoperative hospital stay (days) | 23 (16–39) |
| Location of airway defect | |||
| Upper trachea | 43 (71.7%) | ||
| Middle trachea | 11 (18.3%) | ||
| Lower trachea | 6 (10.0%) | ||
| Tracheal or esophageal stent a | 5 (8.5%) | ||
| Positive pressure ventilation a | 8 (13.3%) | ||
| ECMO a | 1 (1.7%) |
Note: Data are reported as median (interquartile range) or n (%).
At the time of surgery. ECMO: Extracorporeal membrane oxygenation.
At the time of surgery, 40 patients had a tracheostomy in place, 8 patients were assisted by positive pressure ventilation and 5 patients had either an esophageal (n = 3) or a tracheal (n = 2) stent in place.
The majority of patients were operated on by cervicotomy (n = 42, 70%), while a sternal split or a full median sternotomy was used in 4 patients, either alone (n = 1) or in combination with cervicotomy (n = 3). Finally, 13 patients were approached by thoracotomy. In particular, a standard right posterolateral thoracotomy was performed in 8 cases; 4 patients were approached by a high posterior right thoracotomy; while one patient, who had a double incomplete aortic arch with atresia of the distal left arch was approached by left posterolateral thoracotomy [17].
The airway defect was repaired by direct closure in 8 patients, tracheal resection‐anastomosis in 30, and by bridging of the defect with either a resorbable patch (n = 19 patients) or a muscle flap (n = 3) in the remaining cases. The resorbable patch consisted in Gore Bio‐A tissue reinforcement (W.L. Gore & Associates Inc., Newark, DE) in 17 cases and Vicryl knitted mesh (Vicryl; Ethicon) in 2. The regional muscle flaps used for airway closure consisted in strap muscles in 2 cases and intercostal muscle in 1.
The esophageal defect was repaired in 56 patients (90.7%) by a two‐layered suture, while 4 patients, who had an extensive esophageal damage, underwent esophageal diversion.
In 46 patients (76.7%) a vascularized tissue flap was placed as a separation between the trachea and esophagus, including all 19 patients who underwent repair by resorbable patch. For this purpose, regional muscle flaps consisted in sternocleidomastoid muscle (n = 17), strap muscles (n = 14), intercostal muscle (n = 12), thymic fat pad (n = 2) and serratus anterior (n = 1). At the end of the operation, 34 patients (56.7%) were left with a tracheostomy, including 29 of the 40 patients who had a tracheostomy prior to the operation.
3.2. Postoperative Outcomes
The overall morbidity rate was 70% (38 patients). Complications are shown in Table 2. Infectious complications were the most common cause of morbidity, occurring in 20 patients (33.3%), followed by respiratory failure (16.6%). Among surgical complications, 3 patients who had undergone tracheal resection/anastomosis experienced dehiscence of the anastomosis: in two cases the dehiscence was small and was treated conservatively, while one patient required reoperation and redo of the tracheal anastomosis. The fistula relapse was observed in 2 patients (3.7%). One patient, who was on mechanical ventilation at the time of surgery, relapsed 2 months after surgery, after having been transferred to a rehabilitation facility: he died of repeated infectious complications and respiratory failure. The second case was a patient affected by end‐stage liver disease, who developed relapse of the fistula 3 months after surgery, and was treated by two re‐operations. Definitive healing of the fistula was achieved after the second reoperation, 21 months after initial surgery. Details and postoperative course of patients who experienced airway anastomotic complications are reported in Supporting Information S1: Table S1. In‐hospital mortality occurred in 2 patients: in the above described case of fistula recurrence, and in an additional patient due to respiratory failure. This second patient was hospitalized for mediastinitis following polytrauma, and had already undergone multiple operations before diagnosis of the fistula. He was on veno‐venous ECMO and on mechanical ventilation at the time of surgery. The postoperative period was complicated by renal insufficiency, sepsis and respiratory failure.
TABLE 2.
Postoperative complications.
| Surgical complications | N | Medical complications | N |
|---|---|---|---|
| Wound dehiscence a | 4 | Infectious complications c | 20 |
| Tracheal dehiscence | 3 | Respiratory failure d | 10 |
| Surgical site bleeding b | 3 | Cardiac complications e | 4 |
| Granulations | 3 | Sialorrhea | 2 |
| Fistula relapse | 2 | Delirium | 2 |
| Tracheal stenosis | 2 | Acute renal failure | 1 |
| Tracheomalacia | 2 | Mucus plugging | 1 |
| Vocal fold paralysis | 1 | Deep vein thrombosis | 1 |
| Endobronchial bleeding | 1 | Hyperammonemia | 1 |
Dehiscence of cervicotomy in 2 cases and sternotomy in 2 cases.
Hemothorax in 2 cases, cervical hemorrhage in 1 case.
Signs or symptoms of infection associated with isolation of a new pathogenic agent, requiring new antibiotic treatment.
Positive pressure ventilation for more than 96 h after surgery, or any worsening of respiratory status requiring positive pressure ventilation.
Includes atrial fibrillation in 2 patients, sustained hypotension in 1, and cardiac arrest in 1.
3.3. Differences Between Patients' Subgroups
The study period ranged from 1997 to 2023, and, the only significant modification among perioperative characteristics was the introduction of resorbable patches as a method for closure of airway defects (Supporting Information S1: Table S2). Table 3 reports the baseline patient's characteristics and the outcomes according to the different surgical procedures for tracheal repair. The location of the airway defect differed significantly among surgical techniques (p = 0.012). Patients treated with non‐circumferential airway repair techniques (direct closure or flap/patch interposition) more commonly had fistulas involving the middle or lower trachea, whereas resection/anastomosis was mainly performed for upper tracheal defects. Compared with patients undergoing other types of repair, those treated with flap or patch interposition had a higher comorbidity index and larger airway defects. Esophageal diversion was performed exclusively in this group. These patients also exhibited worse perioperative outcomes, including a higher complication rate and longer postoperative stay.
TABLE 3.
Patients' characteristics by type of airway repair procedure.
| Characteristic |
Direct closure, N = 8 |
Resection/anastomosis, N = 30 |
Flap or patch interposition, N = 22 |
p‐value |
|---|---|---|---|---|
| Age (years) | 58.5 (55.25–69.25) | 52.5 (34.5–67.75) | 61.5 (55–68.75) | 0.276 |
| Sex | 0.688 | |||
| Male | 3 (37.5%) | 14 (46.7%) | 12 (54.5%) | |
| Female | 5 (62.5%) | 16 (53.3%) | 10 (45.5%) | |
| Charlson comorbidity index | 3 (2–3.5) | 2 (0.75–4) | 4 (3–5) | 0.030 |
| Diabetes | 2 (28.6%) | 6 (21.4%) | 7 (33.3%) | 0.644 |
| History of cancer | 1 (14.3%) | 0 (0.0%) | 7 (31.8%) | 0.004 |
| Diameter of the fistula (mm) | 8.50 (5.00–16.25) | 17.5 (10.0–23.75) | 25.00 (15.0–37.5) | 0.009 |
| Etiology of the fistula | 0.270 | |||
| Post‐intubation injury | 6 (75.0%) | 26 (86.7%) | 15 (68.2%) | |
| Other | 2 (25.0%) | 4 (13.3%) | 7 (31.8%) | |
| Location of airway defect | 0.012 | |||
| Upper trachea | 4 (50.0%) | 26 (86.7%) | 13 (59.1%) | |
| Middle trachea | 1 (12.5%) | 4 (13.3%) | 6 (27.3%) | |
| Lower trachea | 3 (37.5%) | 0 (0.0%) | 3 (13.6%) | |
| Positive pressure ventilation | 1 (12.5%) | 3 (10.0%) | 4 (18.2%) | 0.690 |
| Esophageal repair procedure | 0.025 | |||
| Double‐layered suture | 8 (100.0%) | 30 (100.0%) | 18 (81.8%) | |
| Esophageal diversion | 0 (0.0%) | 0 (0.0%) | 4 (18.2%) | |
| Morbidity | 1 (12.5%) | 15 (50.0%) | 16 (72.7%) | 0.012 |
| Postop hospital stay (days) | 19.5 (15.5–23) | 19.5 (13.25–34.75) | 32.5 (20–53.50) | 0.028 |
| Mortality | 0 (0.0%) | 0 (0.0%) | 2 (9.1%) | 0.167 |
| Fistula relapse | 0 (0.0%) | 0 (0.0%) | 2 (9.1%) | 0.167 |
Note: Data are reported as median (interquartile range) or n (%). Bold p‐values indicate statistical significance at the 0.05 level.
3.4. Long‐Term Outcomes
After discharge, 3 patients were lost to follow‐up. Median follow‐up for the other 55 patients who survived the operation was 42 months (IQR: 19.5–92 months). At end follow‐up, 16/55 patients (29.1%) died. The causes of death were unknown in 6 patients, while in remaining patients they consisted in myocardial infarction in 2 cases, chronic respiratory failure in 2, neoplasia in 2, cachexia in 2, medullary aplasia and cerebrovascular and stroke in 1 case each. The 1‐, 2‐ and 5‐ survival rates (95% CI) were 90.4% (82.8%–98.8%), 86.0% (49.3%–76.9%) and 70.8% (57.7%–86.9%), respectively (Figure 2).
FIGURE 2.
Overall survival following surgery for tracheoesophageal fistula.
Of the 34 patients who were discharged with a tracheostomy, 14 patients were decannulated after a median (IQR) follow‐up time of 4.5 (2.75–17.25) months. Among the remaining 20 patients, 12 died before the end of follow‐up, while the reason for persisting tracheostomy in the remaining 8 patients consisted in diffuse airway malacia in 2 cases, severe neurological impairment in 2, emphysema with oxygen requirement in 2, tracheal stenosis after attempted decannulation in 1, while one patient who was operated on in 2023 is currently awaiting removal of the T‐Tube. The cumulative incidence of decannulation (95% CI) was 60.4% (46.7%–71.6%) at 1 year, 64.1% (50.3%–74.9%) at 2 years and 68.4% (54.4%, 79.0%) at 5 years (Figure 3A). At univariable logistic regression, variables significantly associated with a persistent tracheostomy were older age, higher CCI, having a tracheostomy and being on mechanical ventilation at the time of surgery (Table 4).
FIGURE 3.
Cumulative incidence functions of tracheostomy‐free breathing (A), and resumption of oral intake (B) after surgery for tracheoesophageal fistula, accounting for the competing risk of death.
TABLE 4.
Univariable logistic regression for the risk of persistence of the tracheostomy and resumption of oral feeding at last follow‐up.
| Characteristic | Persistent tracheostomy | OR (95% CI) | p‐value | |
|---|---|---|---|---|
| Yes (n = 20) | No (n = 40) | |||
| Age | 64 (56.5–69.5) | 56.5 (37.5–68) | 1.05 (1.01–1.11) | 0.010 |
| Charlson comorbidity index | 5 (3–6) | 3 (1–4) | 1.49 (1.14–2.06) | 0.003 |
| Diabetes | 8 (42.1%) | 7 (18.9%) | 3.12 (0.92–11.0) | 0.068 |
| Obesity | 5 (25.0%) | 3 (7.7%) | 4.00 (0.87–21.6) | 0.074 |
| Preoperative tracheostomy | 17 (85.0%) | 23 (57.5%) | 4.19 (1.17–20.0) | 0.026 |
| Diameter of the fistula (mm) | 22.5 (10.0–30.0) | 15.0 (10.0–25.0) | 1.05 (1.00–1.10) | 0.057 |
| Etiology of the fistula | ||||
| Other | 4 (20.0%) | 9 (22.5%) | Ref | |
| Post‐intubation | 16 (80.0%) | 31 (77.5%) | 1.16 (0.32–4.82) | 0.825 |
| Positive pressure ventilation at surgery | 7 (35.0%) | 1 (2.5%) | 21.0 (3.31–412) | < 0.001 |
| Airway repair procedure | ||||
| Direct closure | 2 (10.0%) | 6 (15.0%) | Ref | |
| Resection/anastomosis | 6 (30.0%) | 24 (60.0%) | 0.75 (0.13–5.99) | 0.758 |
| Flap or patch interposition | 12 (60.0%) | 10 (25.0%) | 3.60 (0.66–28.5) | 0.165 |
| Characteristic | Resumption of oral feeding a | OR (95% CI) | p‐value | |
|---|---|---|---|---|
| No (n = 9) | Yes (n = 49) | |||
| Age | 70 (63–71) | 58 (49–67) | 1.06 (1.00–1.16) | 0.082 |
| Charlson comorbidity index | 6 (5–6.5) | 3 (1–4) | 1.96 (1.27–3.51) | 0.007 |
| Diabetes | 3 (37.5%) | 11 (23.4%) | 1.96 (0.36–9.39) | 0.403 |
| Obesity | 3 (33.3%) | 5 (10.2%) | 4.40 (0.75–23.4) | 0.081 |
| Septic syndrome | 5 (55.6%) | 20 (41.7%) | 1.75 (0.41–7.86) | 0.445 |
| Diameter of the fistula (mm) | 30 (25–40) | 15 (10–25) | 1.15 (1.07–1.29) | 0.002 |
| Etiology of the fistula | ||||
| Other | 3 (33.3%) | 10 (20.4%) | Ref | |
| Post‐intubation | 6 (66.7%) | 39 (79.6%) | 0.51 (0.11–2.76) | 0.398 |
| Positive pressure ventilation at surgery | 4 (44.4%) | 4 (8.2%) | 9.0 (1.69–51.2) | 0.010 |
| Esophageal repair procedure | ||||
| Double‐layered suture | 7 (77.8%) | 47 (95.9%) | Ref | |
| Esophageal diversion | 2 (22.2%) | 2 (4.1%) | 6.71 (0.71–64.1) | 0.078 |
Note: Data are reported as median (interquartile range) or n (%). Bold p‐values indicate statistical significance at the 0.05 level.
Information about eating status is missing for 2 patients.
Resumption of oral intake before death or end of the follow‐up was achieved in 51 of the 58 surviving patients (89%). Of the 4 patients who underwent esophageal diversion, 2 underwent esophageal reconstruction after 6 and 14 months, one patient died after 5 months due to cachexia, and one was lost at follow‐up after 3 months from surgery. The cumulative incidence of resumption of oral intake (95% CI) was 76.9% (63.4%–86.0%) at 1 year, 82.8% (69.5%–90.6%) at 2 years, and 86.6% (73.2%–93.6%) at 5 years after surgery (Figure 3B). At univariable analysis, variables associated with failure to resume oral intake were higher CCI, larger fistula size and being on mechanical ventilation at the time of surgery (Table 4).
4. Discussion
The management of patients affected by TEF is challenging. General principles include weaning from mechanical ventilation, control of infection and institution of enteral nutrition prior to performing surgical repair; however, this is not always possible for most compromised patients. In these cases, the choice to perform surgical correction, at the price of an increased risk of failure, is taken in light of the absence of other options to control airway contamination or improve the respiratory conditions of the patient. Mechanical ventilation is known to be associated with dehiscence of airway sutures in related clinical settings, such as anastomotic complications after lung transplantation [18], or bronchopleural fistula after pneumonectomy [19]. Concerning studies focusing on TEF, Foroulis et al. reported on a series of 13 patients, 5 of whom were on mechanical ventilation at the time of surgical correction. Of these, 3 patients died, while no other fatality occurred among non‐ventilated patients [20]. Similarly, in our series, the only two postoperative deaths, and one of the two cases of fistula relapse, occurred among the 8 patients who were on mechanical ventilation. These findings seem to portend a worse prognosis in mechanically ventilated patients. However, the association between mechanical ventilation and poorer outcomes is likely multifactorial and may largely reflect the worse general conditions of ventilated patients, rather than a direct causal effect. Regardless, because in our experience the majority of such patients healed, we do not consider mechanical ventilation as an absolute contraindication for surgery.
Over time, the number of patients affected by TEF who were referred to our Center has doubled (Supporting Information S1: Table S1). Post‐intubation injury remained the most prevalent cause of TEF; however, a slight increase in other causes of TEF was observed. Among other causes, we found those fistulas arising from surgical or endoscopic procedures at the level of the esophagus (e.g., esophagectomy or stent placement) particularly challenging. In fact, the extent of injury at that level often precluded the possibility to repair the esophagus/neo‐esophagus, which was resected and diverted with a cervical esophagostomy. Furthermore, as observed by Balakrishnan et al., esophagectomy results in a loss of vascularization of the trachea, which may increase the risk of airway complications if airway resection–anastomosis is attempted [21]. Therefore, non‐circumferential techniques of airway repair were preferred for those patients.
At our institution, we started using resorbable patches in 2011 [22], and, since then, we have gained increasing experience and confidence with their use. Probably, some patients approached by this technique could not have benefited from other conventional surgical options. Compared to patients operated on by resection/anastomosis or direct suture, those operated on by this technique had worse baseline conditions, including a higher CCI and longer airway defect (Table 3). Moreover, resorbable patches were often chosen as a technique of non‐circumferential tracheal repair in patients with post‐surgical/endoscopic TEF who concurrently necessitated esophageal diversion. Therefore, the results observed in terms of morbidity, mortality, length of postoperative stay and fistula relapse should be interpreted in light of these elements.
Overall, our outcomes compare favorably with those of other more recent case‐series, where the rate of fistula relapse after surgical treatment for TEF ranges from 3.3% to 11% [3, 5, 7, 8], and surgical mortality ranges from 0% to 5.7%.
Concerning long‐term outcomes, from available data, it can be inferred that survival in these patients is hampered mostly by underlying conditions such as respiratory failure, neoplasia or malnutrition.
Other functional outcomes (breathing without tracheal appliance and resumption of oral intake) are inconsistently reported in the literature, often as crude rates among operative survivors or among long‐term survivors. However, in consideration of the concurrent risk of death, we believe that the cumulative incidence function may provide a more reliable estimate. The 5‐year cumulative incidence of decannulation and of resumption of oral intake (68.4%, and 86.6%, respectively) are comparable with the rates reported in the Literature, ranging from 70% to 98% for breathing without an airway appliance and from 81% to 100% for resumption of oral feeding [3, 5, 6, 7]. In the majority of our long term survivors, the cause for a persistent tracheostomy was mainly attributable to underlying comorbidities, including airway malacia, neurologic impairment and oxygen dependency, rather than to technical factors. Moreover, in our exploratory logistic regression analysis (Table 4), patient‐related variables including the Charlson Comorbidity Index (CCI) and the need for mechanical ventilation at the time of surgery, were significantly associated with both tracheostomy persistence and failure to resume oral intake. These findings, taken together, seem to reinforce the concept that long term outcomes of these patients are strongly influenced by their general conditions at the time of surgery. In our view, the suboptimal functional outcomes observed in more compromised patients do not undermine the value of surgery for TEF in this population, which nevertheless achieved fistula closure and ensured postoperative survival in the majority of cases.
5. Conclusion
In conclusions the surgical treatment of TEF is challenging. Preoperative optimization of patient status and accurate planning of the surgical procedure are fundamental to obtain good outcomes. Preoperative mechanical ventilation is a risk factor for fistula relapse and death; however, if weaning from mechanical ventilation is not possible, surgery may still be attempted. Over time we have adapted our surgical management of TEF, by using muscle flaps or resorbable prostheses for more challenging cases of airway repair. This technique allowed us to perform surgery in patients who probably could have not been operated otherwise. Overall, we observed good outcomes in terms of airway healing and postoperative mortality. Functional outcomes, represented by decannulation, resumption of oral feeding, and long‐term survival were less satisfactory, and were strongly associated with preexisting patients' comorbidities.
Author Contributions
Chiara Catelli: conceptualization, investigation, writing – original draft, visualization. Giulia Pagliarini: investigation, writing – original draft, data curation. Marco Schiavon: project administration, supervision, validation, writing – review and editing. Vincenzo Verzeletti: conceptualization, formal analysis, writing – original draft, data curation. Federico Rea: conceptualization, supervision, methodology. Marco Mammana: conceptualization, investigation , writing – review and editing, methodology, formal analysis, project administration. Andrea Dell’Amore: writing – review and editing, validation, project administration, supervision. Giovanni Maria Comacchio: writing – review and editing, data curation, formal analysis.
Funding
The authors have nothing to report.
Ethics Statement
This study was approved by the local institutional review board (protocol number 271n/AO/22).
Consent
The requirement for informed consent was waived due to the retrospective nature of this study, and the anonymized and aggregated nature of the data.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting Information S1
Acknowledgments
All authors agree with the deceased co‐author's (F.R.) inclusion and the corresponding author (M.M) accepts all responsibility for the deceased co‐author's data and contribution. A previous version of this study was presented at the 30th European Conference on General Thoracic Surgery of the ESTS, 19–21 June 2022, The Hague, The Netherlands.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Mathisen D. J., Grillo H. C., Wain J. C., and Hilgenberg A. D., “Management of Acquired Nonmalignant Tracheoesophageal Fistula,” Annals of Thoracic Surgery 52, no. 4 (1991): 759–765, 10.1016/0003-4975(91)91207-c. [DOI] [PubMed] [Google Scholar]
- 2. Couraud L., Ballester M. J., and Delaisement C., “Acquired Tracheoesophageal Fistula and Its Management,” Seminars in Thoracic and Cardiovascular Surgery 8, no. 4 (1996): 392–399. [PubMed] [Google Scholar]
- 3. Muniappan A., Wain J. C., Wright C. D., et al., “Surgical Treatment of Nonmalignant Tracheoesophageal Fistula: A Thirty‐Five Year Experience,” Annals of Thoracic Surgery 95, no. 4 (2013): 1141–1146, 10.1016/j.athoracsur.2012.07.041. [DOI] [PubMed] [Google Scholar]
- 4. Ji F., Nie P., Yi F., and Zhang L., “Management of Esophageal Stenting‐Associated Esophagotracheal Fistula, Tracheal Stenosis and Tracheal Rupture: A Case Report and Review of the Literature,” International Journal of Clinical and Experimental Pathology 8 (2015): 9332–9336. [PMC free article] [PubMed] [Google Scholar]
- 5. Bibas B. J., Guerreiro Cardoso P. F., Minamoto H., et al., “Surgical Management of Benign Acquired Tracheoesophageal Fistulas: A Ten‐Year Experience,” Annals of Thoracic Surgery 102, no. 4 (2016): 1081–1087, 10.1016/j.athoracsur.2016.04.029. [DOI] [PubMed] [Google Scholar]
- 6. Macchiarini P., Verhoye J. P., Chapelier A., Fadel E., and Dartevelle P., “Evaluation and Outcome of Different Surgical Techniques for Postintubation Tracheoesophageal Fistulas,” Journal of Thoracic and Cardiovascular Surgery 119, no. 2 (2000): 268–276, 10.1016/S0022-5223(00)70182-6. [DOI] [PubMed] [Google Scholar]
- 7. Shen K. R., Allen M. S., Cassivi S. D., et al., “Surgical Management of Acquired Nonmalignant Tracheoesophageal and Bronchoesophageal Fistulae,” Annals of Thoracic Surgery 90, no. 3 (2010): 914–918: discussion 919, 10.1016/j.athoracsur.2010.05.061. [DOI] [PubMed] [Google Scholar]
- 8. Yang G., Li W.‐M., Zhao J.‐B., et al., “A Novel Surgical Method for Acquired Non‐Malignant Complicated Tracheoesophageal and Bronchial‐Gastric Stump Fistula: The ‘Double Patch’ Technique,” Journal of Thoracic Disease 8, no. 11 (2016): 3225–3231, 10.21037/jtd.2016.11.80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Grillo H. C., “Repair of Acquired Tracheoesophageal and Bronchoesophageal Fistula,” in Surgery of the Trachea and Bronchi (BC Decker Inc, 2004), 569–580. [Google Scholar]
- 10. Rosskopfova P., Perentes J. Y., Schäfer M., et al., “Repair of Challenging Non‐Malignant Tracheo‐ or Broncho‐Oesophageal Fistulas by Extrathoracic Muscle Flaps,” European Journal of Cardio‐Thoracic Surgery 51 (2017): 844–851, 10.1093/ejcts/ezw435. [DOI] [PubMed] [Google Scholar]
- 11. Mammana M., Comacchio G. M., Schiavon M., et al., “Repair of Adult Benign Tracheoesophageal Fistulae With Absorbable Patches: Single‐Center Experience,” Annals of Thoracic Surgery 109, no. 4 (2020): 1086–1094, 10.1016/j.athoracsur.2019.09.081. [DOI] [PubMed] [Google Scholar]
- 12. Udelsman B. V., Eaton J., Muniappan A., Morse C. R., Wright C. D., and Mathisen D. J., “Repair of Large Airway Defects With Bioprosthetic Materials,” Journal of Thoracic and Cardiovascular Surgery 152, no. 5 (2016): 1388–1397, 10.1016/j.jtcvs.2016.07.074. [DOI] [PubMed] [Google Scholar]
- 13. Mammana M., Verzeletti V., Dell’Amore A., and Rea F., “An Alternative Surgical Access for Posterior Tracheal Defects or Fistulae Situated Between the Cervical and Thoracic Region,” Updates in Surgery 76, no. 2 (2024): 653–656, 10.1007/s13304-023-01682-6. [DOI] [PubMed] [Google Scholar]
- 14. Bone R. C., Balk R. A., Cerra F. B., et al., “Definitions for Sepsis and Organ Failure and Guidelines for the Use of Innovative Therapies in Sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine,” Chest 101, no. 6 (1992): 1644–1655, 10.1378/chest.101.6.1644. [DOI] [PubMed] [Google Scholar]
- 15. Cederholm T., Bosaeus I., Barazzoni R., et al., “Diagnostic Criteria for Malnutrition—An ESPEN Consensus Statement,” Clinical Nutrition 34, no. 3 (2015): 335–340, 10.1016/j.clnu.2015.03.001. [DOI] [PubMed] [Google Scholar]
- 16. Charlson M. E., Pompei P., Ales K. L., and MacKenzie C. R., “A New Method of Classifying Prognostic Comorbidity in Longitudinal Studies: Development and Validation,” Journal of Chronic Diseases 40, no. 5 (1987): 373–383, 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 17. Zampieri D., Pangoni A., Marulli G., and Rea F., “Acquired Tracheoesophageal Fistula Repair, Due to Prolonged Mechanical Ventilation, in Patient With Double Incomplete Aortic Arch,” Monaldi Archives for Chest Disease 88, no. 3 (2018): 974, 10.4081/monaldi.2018.974. [DOI] [PubMed] [Google Scholar]
- 18. Muñoz‐Fos A., Moreno P., González F. J., et al., “Airway Complications After Lung Transplantation—A Contemporary Series of 400 Bronchial Anastomoses From a Single Center,” Journal of Clinical Medicine 12, no. 9 (2023): 3061, 10.3390/jcm12093061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Mammana M., Marulli G., Zuin A., et al., “Postpneumonectomy Bronchopleural Fistula: Analysis of Risk Factors and the Role of Bronchial Stump Coverage,” Surgery Today 50, no. 2 (2020): 114–122, 10.1007/s00595-019-01871-0. [DOI] [PubMed] [Google Scholar]
- 20. Foroulis C. N., Nana C., Kleontas A., et al., “Repair of Post‐Intubation Tracheoesophageal Fistulae Through the Left Pre‐Sternocleidomastoid Approach: A Recent Case Series of 13 Patients,” supplement, Journal of Thoracic Disease 7, no. S1 (2015): S20–S26, 10.3978/j.issn.2072-1439.2015.02.01. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Balakrishnan A., Tapias L., Wright C. D., et al., “Surgical Management of Post‐Esophagectomy Tracheo‐Bronchial‐Esophageal Fistula,” Annals of Thoracic Surgery 106 (2018): 1640–1646, 10.1016/j.athoracsur.2018.06.076. [DOI] [PubMed] [Google Scholar]
- 22. Battistella L., Marulli G., Comacchio G. M., Mammana M., Di Gregorio G., and Rea F., “Successful Treatment of a Recurrent Wide Tracheoesophageal Fistula With a Bioabsorbable Patch,” Annals of Thoracic Surgery 101, no. 5 (2016): e173–e175, 10.1016/j.athoracsur.2015.11.038. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information S1
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.