Tracheo- and broncho-esophageal fistulas in adults

INTRODUCTION — Pathological connections between the esophagus and the trachea or major bronchi are termed tracheoesophageal fistula (TEF) and bronchoesophageal fistula (BEF), respectively. Although uncommon, they can be a major source of significant morbidity and mortality, particularly in patients with esophageal or lung cancer. Since the etiologies and management are similar for TEF and BEF, for the purposes of this topic, both will be referred to as TEF.

This topic will discuss the clinical features, diagnosis, and treatment of patients with TEF and BEF. Bronchopleural and alveolopleural fistulas are discussed separately. (See "Bronchopleural fistula in adults" and "Alveolopleural fistula and prolonged air leak in adults".)

ETIOLOGIES — Most TEFs in adults are acquired and due to esophageal or lung cancer. Congenital TEFs are rare in adults.

●Acquired – Malignancy, typically esophageal or lung cancer, accounts for over 50 percent of TEFs. However, TEFs are rare as a complication of malignancy; about 5 to 15 percent of patients with esophageal malignancy and 1 percent of patients with bronchogenic carcinoma develop TEF [1-4]. Less commonly, TEFs are due to benign conditions (eg, prolonged endotracheal intubation, surgical or endoscopic interventions, mediastinal granuloma) (table 1) [1-4].

●Congenital – Most congenital TEFs present in childhood and are typically associated with esophageal atresia. Rarely, a small congenital H-type TEF (communicates with a normal esophagus) may present in adulthood [5]. (See "Congenital anomalies of the intrathoracic airways and tracheoesophageal fistula".)

CLINICAL FEATURES — TEF should be suspected in patients with a known risk factor who have one or more of the following:

●Frequent coughing following solid and liquid intake

●Recurrent purulent bronchitis or pneumonia

●Recurrent aspiration

●Unexplained malnutrition

Patients who are receiving mechanical ventilation can develop TEF from prolonged endotracheal intubation and may present with acute respiratory distress, worsening oxygenation, loss of tidal volume during ventilation, and gastric distension.

In most patients, the symptoms develop over days to weeks (eg, those due to malignancy) while in others the symptoms may be acute (over hours; eg, those due to intubation). The onset may also depend upon the location and size of the fistula in that large proximal TEFs may present earlier than smaller distal TEFs.

Symptoms of the underlying cause may also be present (eg, cough and hemoptysis from lung cancer or dysphagia and weight loss from esophageal cancer).

Since the condition is uncommon and the symptoms are nonspecific, the diagnosis is often delayed.

DIAGNOSTIC EVALUATION — An esophagram and endoscopy should be performed in patients with suspected TEF. There are no specific laboratory findings, although some patients with lung infection may have a leukocytosis or elevated erythrocyte sedimentation rate. TEFs should be identified and treated promptly since spontaneous closure is rare and if left untreated, patients progress to respiratory failure and death. There are no guidelines on how best to evaluate and diagnose TEFs. Our approach, outlined in this section, is based upon observational series and our experience.

Imaging — The diagnosis of TEF is traditionally made with contrast-enhanced esophagography that demonstrates displacement of the contrast into the lung; barium preparation is preferred over Gastrografin because the latter is extremely hypertonic and can induce pulmonary edema, pneumonia, or death, whereas aspiration of small amounts of barium seems to have little clinical relevance [6]. In patients who cannot swallow (eg, mechanically ventilated patients), contrast studies may not be feasible such that chest computed tomography (CT) is an alternative. Chest CT scan with three-dimensional reconstruction and oral or intravenous contrast medium may localize the fistula, identify details of the fistula etiology, and examine tracheal/esophageal anatomy, all of which are important for evaluating potential therapies.

Endoscopy — In most cases, bronchoscopy and/or endoscopy should be performed to confirm imaging findings and localize the fistula (picture 1 and picture 2). In cases where a malignant etiology is suspected a biopsy should be performed to confirm the diagnosis. However, recognition of smaller fistulas can be challenging especially if the mucosa is red and swollen. Orally administered methylene blue before bronchoscopy with observation of bubbles leaking into the airway has been described as helpful in identifying small fistulas [7].

DIAGNOSIS — The diagnosis of TEF is usually based upon a combination of clinical, radiographic, and endoscopic findings. While contrast imaging, computed tomography, and endoscopy are not always necessary for the diagnosis, all three are typically done since they provide complementary diagnostic and therapeutic information.

MANAGEMENT AND OUTCOMES — Spontaneous closure of TEF is rare. Without treatment, the outcome is poor and can be measured in weeks. The management of TEF requires a multidisciplinary approach involving thoracic surgery, oncology, gastroenterology, and interventional pulmonary experts. Careful assessment of the etiology, size, anatomy (tracheal and esophageal), burden of the underlying disease, and patient comorbidities as well as the risk-benefit ratio of various repair options should be undertaken when deciding about the best treatment approach. In general, surgery with a curative intent is usually performed for benign TEF [8], whereas palliative management is reserved for malignant TEF (algorithm 1 and algorithm 2 and table 1). Since malignancy is the more common etiology, palliative endoscopic/bronchoscopic treatments are more frequently administered.

There is a paucity of data and no consensus or guidelines on how best to manage TEF. Significant variation among clinicians exists but practice is evolving as expertise in interventional pulmonology grows. The approach outlined here is influenced by our expertise in interventional pulmonology and we recognize that this strategy may not always be universally applied, particularly when interventional expertise is not available.

Initial general measures — Initial general measures that should be undertaken include eliminating oral intake, keeping the head of the bed elevated at 45 degrees or greater, administering anti-reflux therapy, frequent oral suctioning, treating pulmonary infection/aspiration pneumonia, and oxygenation with supplemental oxygenation (if indicated). Nasogastric tubes should be removed, if present. A gastrostomy tube is also sometimes placed to suction gastric contents which potentially reduces further leak from gastroesophageal reflux (eg, mechanically ventilated patient or patient with TEF in the lower one-third of the esophagus). In patients who are considered for endoscopic stenting some experts recommend placing a gastrostomy or jejunostomy tube prophylactically for enteral nutrition. If this is not feasible, total parenteral nutrition is appropriate.

If patients are receiving mechanical ventilation, extubation is preferable but is not always feasible. For those who cannot be extubated, bypassing the site of the TEF by advancing the endotracheal tube (ETT) or placing an extra-long tracheostomy tube, should be performed ensuring that the inflated cuff is below the fistula; ETTs or tracheostomy tubes with continuous subglottic aspiration (if available) are preferable. Some experts also lower the tidal volume and positive end-expiratory pressure in an attempt to minimize air leak through the fistula and avoid TEF enlargement, although this maneuver is of unproven benefit. Many experts wait until patients are weaned off mechanical ventilation before attempting surgical repair since it has been shown that positive pressure ventilation is associated with an increased incidence of anastomotic dehiscence and restenosis following surgery [1,9-13].

Treat underlying cause — While therapy is mostly aimed at treating the fistulous communication, efforts targeted at treating the underlying cause should be simultaneously undertaken (table 1). This may involve treating any underlying disorder such as malignancy or infection, or removing the patient from mechanical ventilation, if feasible.

Malignant lesions (palliative therapy) — Patients with malignant TEF are treated palliatively (algorithm 1). While there are rare case reports of malignant TEF being treated surgically (eg, surgically fit patients with minimally invasive localized disease or surgery after oncologic therapy), it is not generally recommended and frequently not feasible [14]. (See 'Patients or lesions suitable for surgery' below.)

The treatment for most malignant TEFs, particularly those >5 mm involves stenting of the esophagus, airway, or both. A smaller proportion of patients have small lesions (eg, ≤5 mm) which are generally treated with local bronchoscopic therapies (eg, clipping or fibrin glue). The choice of palliative intervention depends upon lesion size and location as well as level of expertise and patient comorbidities. In most cases, a complete or partial response can be achieved to allow patients to survive a few more weeks or months with an improved quality of life. The rationale for this approach is based upon our experience and that of others [15].

Fistula in mid-proximal esophagus — The majority of TEFs are located in the middle one-third of the esophagus (since most cancers are located within this region), and most of those are treated with double stenting (ie, sequential placement of tracheal and esophageal stents).

First line: Double (combined) stenting — For patients with malignant TEF involving the mid to proximal esophagus, airway (typically tracheal) stenting in addition to esophageal stenting should be performed. While an airway stent is indicated in those with an associated stenotic airway lesion, it is also appropriate in those without airway stenosis. Unlike distal esophageal stents, stents placed in the mid to proximal portion of the esophagus can result in potential airway obstruction once the esophageal stent is fully expanded; thus, placement of an airway stent prior to the esophageal stent can prevent this complication [16].

One retrospective study evaluated the clinical efficacy of airway stenting in 61 patients with TEF [17]. Almost every patient who underwent double stenting had a complete response (defined as no leakage of contrast medium after radiography and clinical symptom resolution without recurrence for more than two weeks) compared with two-thirds of patients who had airway stenting alone. In another retrospective study of 30 patients with large fistulas and airway stenosis, patients who received a double stent had a greater mean survival than those who received an airway stent alone (110 versus 24 days) [18]. In another study of 112 patients with TEF, double stenting was associated with improved survival when compared with airway stenting alone (252 versus 219 days) [19].

Airway stent — The airway stent should be placed first followed by the esophageal stent, preferably in one setting, although one procedure is not always feasible; this reduces the risk of airway obstruction by the esophageal stent and decreases the risk of esophageal stent migration.

Airway stents should cover at least 20 mm beyond the proximal and distal margin of the fistula (picture 3). In addition, the stent should be 10 to 20 percent larger than the internal airway diameter at the fistula site. Straight stents are often used for proximal airway TEFs while L- or Y-shaped ones are used for lower airway TEFs.

Factors that affect success include apposition between the stent and esophageal and/or tracheal wall around the fistula, adequacy of stent covering of the fistula, and the degree of associated esophageal or tracheal stenosis. A variety of airway stents are available for clinical application in the tracheobronchial tree for treatment of TEF. Options include covered self-expanding metal stents (ie, tubular mesh that is partially or fully covered), silicone stents, or hybrid stents. Although there are no comparative studies available, we prefer to use covered self-expanding metal stents since they have the following advantages over silicone stents:

●Preferable inner diameter to wall thickness ratio.

●Excellent expansion and adherence to the tracheal wall creating a good seal (silicone stents do not self-expand so they do not conform to the tracheal wall as readily).

●Fewer episodes of stent migration (due to the above); although fully covered metallic stents may migrate more than partially covered, both migrate less than silicone stents.

●Less airway obstruction due to improved secretion clearance.

●Ease of insertion using a flexible bronchoscope (silicone stents require a rigid bronchoscope).

●Available sizes for larger airways (airway diameter >18 mm and ≤20 mm; silicone stents not available for large airways)

●Ease of deployment, which decreases the risk of expanding the size of the TEF orifice at the time of stent placement.

The disadvantages of using a metallic over a silicone stent are:

●Greater tumor in growth and esophageal spill (of food particles and liquid material) into the trachea through partially covered metal stent interstices (at the proximal and distal end of the stent, which is not covered).

●Difficulty removing (often due to tumor infiltration and excellent adherence properties of the partially covered stents).

●Less durability due to increased risk of stent fracture (500 to 1000 days).

Thus, a fully covered metallic stent may be preferred for patients with malignant fistulas [20,21], those in whom short-term use is planned, for TEF associated with tortuous airways, those with an airway diameter >18 and ≤20 mm, or TEF located in upper and mid airway (stents placed in the upper airway have a greater chance of migrating than those in the lower airway). On the other hand, a silicone stent may be preferred when long-term use is needed (eg, >500 days or when TEF is associated with benign airway stenosis) since they last longer and are more durable than metallic stents.

Details regarding the insertion, complications, and follow-up of airway stents are provided separately. (See "Airway stents", section on 'Insertion technique' and "Airway stents", section on 'Follow-up' and "Airway stents", section on 'Complications'.)

Esophageal stent — The upper margin of the esophageal stent should be slightly higher than the upper margin of airway stent, which is thought to reduce the risk of esophageal stent migration (picture 4). The stent length and diameter are chosen according to the location of the lesion, size of the TEF and degree of esophageal stenosis. Stents must cover 2 cm beyond the proximal and distal margins of the TEF, should be wide enough to press firmly against the esophageal wall, and are usually placed under direct visualization through endoscopy or under fluoroscopy. Technical details and efficacy of stenting for esophageal lesions are provided separately. (See "Endoscopic stenting for palliation of malignant esophageal obstruction".)

Second line options — Second line options include:

●Airway stent alone – Airway stenting alone (picture 3) (ie, without concomitant esophageal stent) is indicated in rare circumstances when an esophageal stent is not feasible. Examples include TEF located in the high proximal esophagus (where an esophageal stent is often not feasible since patients do not tolerate them due to pain), TEF associated with a completely occluded esophageal lumen that precludes the passage of an esophageal guide wire, and severe esophageal stenosis that is at risk of rupture from esophageal stenting.

●Esophageal stent – If an esophageal stent is feasible but an airway stent is not, due to the size of the airway (≥20 mm), then an esophageal stent may be placed, taking into consideration that airway encroachment and worsening of fistula size are potential complications. (See 'Esophageal stent' above.)

Distal esophagus fistulas — Distal esophageal fistulas are less common than mid or proximal ones and are more likely to communicate with major bronchi than with the trachea (ie, BEF). Management is dependent upon whether or not endobronchial stenosis is present.

No airway stenosis: Esophageal stent — For patients with large TEFs/BEFs in the distal esophagus without concomitant airway stenosis, an esophageal stent (picture 4) alone is indicated. An airway stent is not needed since esophageal stents in this location are not typically associated with airway compromise or worsening TEF [16]. (See 'First line: Double (combined) stenting' above.)

Few studies have selectively compared outcomes using individual stenting strategies. Nonetheless, one prospective study reported survival rates of 263 days in those with esophageal stents compared with 219 days for those who received an airway stent alone and 252 days for those who received a combined airway-esophageal stent [19].

If an esophageal stent migrates (eg, because there is no esophageal stenosis), TEF is persistent or worsened by the stent, or the airway is compromised by the stent, esophageal stent replacement with a concomitant airway stent are options (ie, double stenting). (See 'First line: Double (combined) stenting' above.)

Airway stenosis: Double stent — For patients with large TEFs in the distal esophagus with concomitant airway stenosis (typically bronchial), double stenting is indicated. Bronchial stents are not more difficult or less feasible to place than tracheal stents. Bronchial stents can be placed up to distal left main stem bronchus and distal right bronchus intermedius. A frequent problem that is encountered with right BEF when the communication is at the level of the right upper lobe (RUL) bronchus is that in order to seal the defect, the RUL needs to be removed from ventilation (ie, "jailed") with a stent that is placed from the right main stem (RMS) bronchus and the bronchus intermedius (BI). Another problem with right-sided BEF is that there can be a size mismatch between the RMS and the BI so placing a stent of ideal size and fit can be an issue. In this case, a custom three-dimensional printed silicone stent can be considered, although experience with this novel approach is limited [22]. (See 'First line: Double (combined) stenting' above.)

Benign lesions (potentially curative) — The intent of treating benign TEF is curative with surgery. However, not all patients or lesions are suitable for surgical repair. Importantly, definitive surgical repair cannot be performed unless the underlying disorder is curable and site of potential anastomosis is disease-free (algorithm 2). Palliative procedures are frequently performed as a bridge to surgery while the underlying disorder is being treated.

Patients or lesions suitable for surgery — Surgical repair is technically difficult surgery that uses a cervicotomy, cervicosternotomy, or thoracotomy approach; thus, expertise in both esophageal and tracheal surgery are critical for success. The surgical approach for TEF depends upon the size of the fistula [1,23,24]:

●For small lesions, the fistula is divided and repaired using one or two layers of omental or muscle flaps (between the esophagus and trachea) over a nasogastric tube.

●Large fistulas with tracheal injury may require major esophageal and/or tracheal surgery including any combination of the following: esophageal diversion (esophagostomy) or resection, full thickness skin graft esophageal reconstruction, tracheal or laryngotracheal resection and reconstruction, and muscle flap interposition.

Immediate extubation following surgery is the goal since it is thought that postoperative mechanical ventilation may lead to wound dehiscence and fistula recurrence. (See 'Initial general measures' above.)

Successful fistula closure following surgical intervention for benign TEF has been reported in 75 to 94 percent of patients with median follow-up times between 23 months and 12.5 years [25-29]. The majority are able to resume oral intake (>70 percent) and do not require prolonged mechanical ventilation after surgery. One retrospective study reported that in patients with benign TEF, reintervention was less likely in those who had surgery compared with patients who had non-surgical interventions (eg, stenting) [14].

Surgery is frequently fraught with complications (up to 50 percent) [23-25,27,30]. Perioperative mortality as high as 11 percent has been reported. Major complications include wound dehiscence, recurrent TEF, pneumonia, vocal cord paralysis, and tracheal stenosis.

Patients or lesions not suitable for surgery — Patients with benign lesions who are not good surgical candidates (eg, critically ill patients) or who have lesions not suitable for surgery (eg, large fistulas that prohibit reconstruction or resection or fistulas that involve or are close to the major vessels [eg, following esophagectomy]) should undergo palliation or local therapy depending on size. Such interventions are also indicated in patients who need a bridge to surgery (eg, patients on mechanical ventilation or receiving antibiotics for infection).

Fistulas >5 mm: Palliative interventions — Patients in this subgroup are treated with palliative stenting similar to those with malignant TEF. However, most experts deploy only one stent, airway or esophageal, rather than double stenting. The choice of stent depends upon which site can achieve the best "fit" (ie, better sealing effect) and local expertise. When one stent does not seal the defect, then double stenting is appropriate. This approach is based upon the rationale that, unlike malignant lesions, the ideal goal of stenting for benign lesions is to provide a bridge to curative surgery; this strategy maximizes the sealing effect with a single stent and minimizes the risk of stent-related complications. For example, double stenting may interfere with the healing process long term because of the constant friction and radial pressure between stent walls, and unlike malignant fistulas, benign fistulas have less risk of airway obstruction after deployment of an esophageal stent in the mid to proximal region, thereby decreasing the need for double stenting. The details of palliative interventions are discussed above. (See 'Malignant lesions (palliative therapy)' above.)

In few selected patients, TEF healing might be achieved by a mini-invasive approach using an Amplatzer septal occluder, which is available in different sizes and intended for cardiac septal defect closure. It is composed of a nitinol mesh, has two self-expandable disks connected by a thin diameter waist, and ensures mechanical closure of the two sides of the fistula (image 1), making a potential platform for subsequent tissue ingrowth [31].

Three-dimensional designed metallic and silicone segmented Y airway stents have been used to construct personalized stents that are congruent to the airways [32]. However, such technology is not widely available and requires considerable communication between the physician and the manufacturer and a significant level of experience from all parties involved.

Fistulas ≤5 mm: Local therapies — Local therapies are usually preferred in patients with small benign lesions based upon the rationale that risks of stent insertion and their complications outweigh the benefits; stenting is reserved for those who fail local therapies.

Endoscopic clip placement — Gastrointestinal over-the-scope-clipping (OTSC) is a new generation technique that allows closure of gastrointestinal defects including fistulas. The OTSC device is attached to an applicator integrated onto the tip of the endoscope. It requires soft and expandable tissue in order to launch, and is therefore only applied to the gastrointestinal side of the TEF. The major benefits of such devices are speed (procedural time is on average measured in minutes), ease of deployment, and persistent sealing. Small case reports suggest adequate fistula closure following OTSC in TEF [33-36], but further studies are needed to adequately assess outcome in this population with this technique.

Occlusive therapies — Closure of small TEFs with local injection of tissue adhesive, fibrin glue, vascular plugs, septal occluders, or silicon rings has been used in individual cases with variable success [37-42]. However, such methods are uncommonly used owing to their temporary effect and dissolution of occlusive material two weeks following injection leading to recanalization of the fistula.

Investigational therapies — Laser and argon plasma coagulation thermal ablation have also been used in children for refractory congenital TEF in an attempt to promote re-epithelialization [43,44]. However, success has not been reported in adults with non-congenital TEF. (See "Endoscopic palliation of esophageal cancer".)

FOLLOW-UP — In most circumstances patients should undergo a repeat contrast-enhanced esophagram within 48 hours following therapy to make sure complete seal has been achieved and to prevent further pulmonary soiling. In addition, patients should be monitored for symptom relief and maintenance of the response. A complete response is one where no leak is identified and the patient can resume eating and drinking for a minimum of two weeks. A partial response is one where the leak is reduced and symptoms improve for the same time period; however, patients cannot eat. In the latter, reintervention may be considered depending upon the original procedure; this may involve stent replacement, placement of a concomitant airway stent (if not already in place), and/or additional local therapies. There should be a low threshold to re-image and perform endoscopy or bronchoscopy in those who develop new or progressive symptoms that suggest fistula recurrence (up to 11 percent in benign acquired TEF) or a stent-related complication (like obstruction), which is not unusual [25].

SUMMARY AND RECOMMENDATIONS

●Definition – Tracheoesophageal fistula (TEF) is a pathological connection between the esophagus and trachea. Most TEFs in adults are acquired and due to esophageal or lung cancer (table 1). (See 'Etiologies' above.)

●Diagnostic evaluation – TEF should be suspected in patients with a known risk factor who have coughing following solid and liquid intake, recurrent purulent bronchitis or pneumonia, and malnutrition. The diagnosis is usually based upon a combination of clinical, radiographic (esophagram and/or chest computed tomography), and endoscopic findings that demonstrate a leak or fistulous communication (picture 1 and picture 2). (See 'Diagnostic evaluation' above.)

●Initial management – The management of TEF requires a multidisciplinary approach (thoracic surgery, oncology, gastroenterology, interventional pulmonary). Initial therapeutic measures include eliminating oral intake, keeping the head of the bed elevated ≥45 degrees, administering anti-reflux therapy, oral suctioning, treating pulmonary infection, oxygenation, and hyperalimentation. Efforts targeted at treating the underlying cause should simultaneously be undertaken. (See 'Initial general measures' above and 'Treat underlying cause' above.)

●Mid-proximal esophagus fistula – For patients with malignant TEF located in the mid to proximal esophagus, we suggest combined esophageal and airway stenting (double stenting) rather than an esophageal or airway stent alone (algorithm 1) (Grade 2C); for those in whom an esophageal stent is not feasible, an airway stent alone is appropriate. Since spontaneous closure is rare and TEF leads to progressive deterioration, stenting is a palliative measure targeted at relieving symptoms or delaying death. (See 'Fistula in mid-proximal esophagus' above.)

●Distal esophagus fistula – For patients with malignant TEF located in the distal esophagus who are without concomitant airway stenosis, we suggest an esophageal stent rather than a double stent (algorithm 1) (Grade 2C); double stenting can be reserved for those with associated airway stenosis or those who fail esophageal stenting alone. (See 'Distal esophagus fistulas' above.)

●Benign lesions – For patients with benign TEF, who are good surgical candidates and in whom surgery is technically feasible, we recommend surgical repair rather than nonsurgical treatments (eg, stenting) (algorithm 2) (Grade 1A) based upon the rationale that this therapy is curative. For patients with benign TEF >5 mm in whom surgical repair is not feasible or patients who need a bridge to surgery, palliative measures (typically single stenting with esophageal or airway stent) are indicated. For patients with benign TEF <5 mm who are not surgical candidates, local endoscopic therapies (eg, clipping, fibrin glue) are appropriate based upon the rationale that risks of stent insertion and their complications outweigh the benefit in this population; stenting is reserved for those who fail local therapies. (See 'Benign lesions (potentially curative)' above.)

●Follow-up – Following therapy, patients should undergo a repeat contrast-enhanced esophagram within 48 hours and clinical assessment for symptom response. There should be a low threshold to re-image and perform endoscopy in those who develop new or progressive symptoms that suggest fistula recurrence or a stent-related complication. For patients with benign TEF who undergo successful surgery, the outcome is typically good. For patients with malignant TEF who undergo palliative approaches, symptoms typically improve but recurrence from the underlying cause is common. (See 'Follow-up' above.)