Isolated tracheo-oesophageal fistula
Article Outline
- Abstract
- INTRODUCTION
- INCIDENCE AND EMBRYOLOGY
- CLINICAL FEATURES
- INVESTIGATION
- MANAGEMENT
- COMPLICATIONS
- PRACTICE POINTS
- References
- Copyright
Abstract
Congenital isolated tracheo-oesophageal fistula is a rare malformation that presents with a characteristic triad of symptoms: choking and cyanosis on feeding, recurrent lower respiratory tract infection and abdominal distension. Children are invariably symptomatic from birth, although the symptoms may be intermittent and may vary in severity. A high index of suspicion is required because the symptoms are not specific. Establishing the diagnosis can be difficult and neither radiology nor bronchoscopy is infallible. Surgical division of the fistula is curative.
Keywords: tracheo-oesophageal fistula, bronchoscopy
INTRODUCTION
Congenital isolated tracheo-oesophageal fistula (H-TOF) is a rare but important malformation of the airway. The presenting symptoms are characteristic but non-specific and may be intermittent. Establishing the diagnosis is frequently difficult and consequently often delayed. The literature contains several reports of adults with congenital H-TOF to attest to these difficulties.1
INCIDENCE AND EMBRYOLOGY
Congenital tracheo-oesophageal fistula without oesophageal atresia has become known, by common usage, as the “H” fistula, although the fistula usually takes an oblique path down from the trachea to the oesophagus. This anomaly accounts for about 4% of tracheo-oesophageal malformations and has an incidence of around 1:50
000–80000 births. The embryology of tracheo-oesophageal malformations is poorly understood but recent work involving an animal model is providing new information. Rat pups exposed to Adriamycin during gastrulation develop a variety of congenital malformations apparently identical to the VATER association seen in humans. Current interest is focused on the role of the notochord and, at a molecular level, sonic hedgehog (Shh). The notochord is responsible for specification of the fate of the cells in the surrounding foregut mesenchyme, this being mediated by the protein encoded by the Shh gene. The notochord is malformed in Adriamycin-exposed embryos, and this appears to be associated with reduced cellular proliferation in the endoderm destined to become the future oesophagus.2
CLINICAL FEATURES
The classical symptoms of an H-TOF are recurrent chest infections, cyanosis and choking on feeding, and abdominal distension. The frequency with which these symptoms occur according to the larger published clinical series is shown in Table 1. The most significant symptom in a neonate is cyanosis and choking during feeding, which is usually reproducible during a “test feed.” These symptoms are not present when the baby is fed by nasogastric tube, and this may add to the delay. In retrospect, it is usually apparent that symptoms are usually present from birth, although they vary in severity and may seem intermittent. This is probably explicable by the oblique passage of the TOF downwards from trachea to oesophagus and the close apposition of these two structures, which, for much of the time, keeps the fistula occluded. Pressure changes in the trachea or oesophagus and the upward movement of the oesophagus during swallowing, result in transient opening of the fistula and the passage of air from the trachea to the oesophagus, or oesophageal contents to the trachea (Fig. 1).
Table 1. Symptoms of congenital isolated tracheo-oesophageal fistula.
| Beasley and Myers3 | Crabbe et al.4 | Andrassy et al.5 | |
| Choking and cyanosis on feeding | 29/30 | 15/16 | 19/21 |
| Recurrent pneumonia | 22/30 | 16/16 | 17/21 |
| Abdominal distension | 10/30 | 7/16 | 8/21 |
| “Mucousy” | 14/30 | 13/21 | |
| All symptoms | 4/30 |
Figure 1.
(a) Reflux of oesophageal contents into the airway. (b) Passage of air from trachea to oesophagus through the fistula.
The “classical” symptoms of an H-TOF are not specific, making diagnosis difficult. The differential diagnosis is listed in Table 2. Concomitant gastro-oesophageal reflux is present in most babies with an H-TOF. Fortunately, the investigation of infants with a suspected H-TOF and infants with the other conditions listed in Table 2 follows a similar course.
Table 2. Differential diagnosis of congenital isolated tracheo-oesophageal fistula.
| • Gastro-oesophageal reflux |
| • Velopharyngeal incompetence |
| • Laryngeal cleft |
| • Oesophageal stricture |
| • Vascular ring |
| • Other feeding problem |
Other malformations will be present in about 50% of babies, including the same abnormalities found in the more common variants of oesophageal atresia/TOF. Table 3 summarises the associated abnormalities reported in the major series.
Table 3. Associated anomalies.
| Beasley and Myers3 | Crabbe et al.4 | Andrassy et al.5 | |||
| Ventricular septal defect | 5 | CHARGE association | 1 | Malrotation | 1 |
| Vascular ring | 2 | Goldenhaar’s syndrome | 1 | Imperforate anus | 1 |
| Fallot’s tetralogy | 1 | Laryngeal cleft | 1 | Ventricular septal defect | 1 |
| Chromosomal anomaly | 2 | Fallot’s tetralogy | 1 | ||
| Duodenal atresia | 1 | Vascular ring | 1 | ||
| Renal anomaly | 1 | Trisomy 21 | 1 | ||
| Vertebral anomaly | 1 | ||||
Clinical examination of the infant with an H-TOF is likely to be normal unless abdominal distension is marked. The presence of other congenital abnormalities should be noted. It is occasionally possible to diagnose an H-TOF clinically by passing a wide-bore nasogastric tube. The catheter is passed into the stomach and then slowly withdrawn, either placing the end into a jar of water or attaching it to a stethoscope. As the tube passes the fistula, air will be seen, or heard, to escape. Unfortunately, high false-positive and false-negative rates mean that this method cannot be relied on for diagnosis.
INVESTIGATION
The investigation of an infant with a suspected H-TOF needs to bear in mind two objectives – first, confirmation of the diagnosis, and second, location of the site of the fistula – because this will influence the approach to subsequent surgical repair.
Radiology is generally the most appropriate method of investigation. In many cases, these children will have had repeated chest X-rays taken for persistent respiratory infections. Aside from pulmonary changes, gaseous distension of the bowel may be a consistent feature, albeit non-specific (Fig. 2). Rarely, the oesophagus will be distended with air.
Figure 2.
Gaseous distension of the bowel on plain films.
It is highly unlikely that an H-TOF will be identified on a routine contrast swallow: this investigation has little predictive value unless contrast is identified in the airway. The mainstay of diagnosis is a tube injection oesophagogram, a technique that is most easily applicable to infants.6 A nasogastric tube is passed into the stomach, and the infant is strapped prone into a cradle in front of an image intensifier. The tube is slowly withdrawn into the mid-oesophagus and non-ionic water soluble contrast injected. The aim is to achieve local distension of the oesophagus with contrast in the region of the fistula, which will encourage the passage of contrast through the TOF into the airway. The majority of H-TOFs will be located at the level of the thoracic inlet (Fig. 3). Video recording of the examination is important because the fistula may only open transiently.
Figure 3.
Tube injection oesophagogram demonstrating the tracheo-oesophageal fistula at the level of the thoracic inlet.
The main problem with contrast studies occurs when contrast is seen in the airway but the route is obscure (Fig. 4). This is usually the result of the reflux of contrast up the oesophagus and aspiration through the larynx. This should be regarded as suspicious, but not diagnostic, of an H-TOF, and a tube injection study should be repeated, with careful attention to detail. An H-TOF is not infrequently missed on the first examination: in the Melbourne series, 8/30 children required a second study, and 2/30 required a third, before the diagnosis was confirmed.3
Figure 4.
Contrast in the airway from aspiration.
Tube injection oesophagography with the child in a prone position becomes increasingly difficult with older children because of lack of co-operation. In these children, bronchoscopy can be a reliable alternative, although it is easy to overlook a TOF at bronchoscopy. The opening of an isolated TOF is usually small and located on the posterior wall of the trachea at the level of the thoracic inlet within a fold of mucosa (Figure 5, Figure 6). If bronchoscopy is performed with a fibre-optic endoscope under general anaesthesia, the examination must be performed through a laryngeal mask airway rather than an endo-tracheal tube because the tip of the latter will probably lie beyond the fistula. Rigid bronchoscopy is more reliable because of the better optics and also because the posterior wall of the trachea can be gently probed with the tip of a suction catheter, which may drop into the ostium of an obscure fistula.
Figure 5.
Bronchoscopic view of an obscure tracheo-oesophageal fistula (H-TOF). The TOF is just visible on the posterior wall of the trachea as a slit-like opening.
Figure 6.
A large tracheo-oesophageal fistula is totally visible on the posterior wall of the trachea.
Upper gastro-intestinal endoscopy or rigid oesophagoscopy has no role in the diagnosis of an H-TOF as the oesophageal ostium of a TOF is simply too small and well hidden in the folds of mucosa of the oesophagus to be seen. There has been some recent interest in the use of three-dimensional computed tomography and virtual bronchoscopy for the diagnosis of H-TOFs but whether this technique will be of clinical value remains to be seen.7
MANAGEMENT
The aim of treatment is closure of the fistula, which is best achieved by surgical division. Once the diagnosis has been established, surgery should be delayed only if there is prospect of rapid improvement in the chest. In infants, the cessation of oral feeds in favour of nasogastric tube feeding is prudent.
The best approach for the repair of an H-TOF is through a cervical incision rather than a thoracotomy, unless the fistula is located very distally in the trachea. Bronchoscopic cannulation of the fistula with a radiological guide wire and passage of a nasogastric tube aid identification of the H-TOF and repair. The fistula follows a short oblique course from the trachea to the oesophagus. After division of the TOF and removal of the guide wire, the trachea and oesophagus are repaired. It is usually possible to extubate the child at the end of the operation, although small infants may need mechanical ventilation for a short period. Providing the integrity of the repair is certain, oral feeding can resume immediately. If there is any doubt, it is wise to delay oral feeding for a few days until a contrast study has shown the repair to be intact.
There are reports of endoscopic techniques to close TOFs, including destruction of the mucosa of the fistula using transbronchoscopic laser treatment and occlusion of the fistula by injecting glue.8 Neither technique has gained widespread acceptance and certainly neither seems superior to open surgery.
COMPLICATIONS
Historically, the principal risks of surgery are injury to the recurrent laryngeal nerves and recurrence of the fistula, although these complications are now considered uncommon. The recurrent laryngeal nerves run between the trachea and the oesophagus on each side and are at risk of injury during the dissection. Recurrent laryngeal nerve injury will result in a vocal cord palsy, which will become manifest as a hoarse cry and often choking during feeding as a result of laryngeal incompetence and aspiration. If recurrent laryngeal nerve injury is suspected, the diagnosis should be confirmed by direct laryngoscopy. Gastro-oesophageal reflux is a frequent cause of co-morbidity in all children with congenital abnormalities of the oesophagus and airway. Despite aggressive medical treatment, about 10% will require antireflux surgery.4
Oesophageal motility is abnormal in children who have undergone an apparently successful repair of oesophageal atresia; this is thought to be related to an intrinsic, possibly neuromuscular, abnormality in the oesophagus. The functional significance of this is residual dysphagia in the absence of a stricture or other anatomical problem. Similar functional problems have been reported in children following repair of a H-TOF.4
PRACTICE POINTS
References
- . Congenital tracheo-esophageal fistula in adults. Ann. Otol. Rhinol. Laryngol. 1976;85:613–617
- . Mechanisms for the development of esophageal atresia. J. Pediatr. Surg. 2001;36:985–994
- . The diagnosis of congenital tracheo-oesophageal fistula. J. Pediatr. Surg. 1988;23:415–417
- . Management of the isolated tracheo-oesophageal fistula. Eur. J. Pediatr. Surg. 1996;6:67–69
- . Congenital tracheo-esophageal fistula without esophageal atresia. A 22-year experience. Am. J. Surg. 1980;140:731–733
- . Difficulties in diagnosis of congenital H-type tracheo-oesophageal fistulae. Clin. Radiol. 1989;40:150–153
- . H-type tracheo-oesophageal fistula: appearance on three-dimensional computed tomography and virtual bronchoscopy. Pediatr. Surg. Int. 2001;17:642–643
- . Endoscopic treatment of tracheoesophageal fistula using electrocautery and the Nd:YAG laser. J. Pediatr. Surg. 1999;34:464–467
PII: S1526-0542(02)00274-9
doi:10.1016/S1526-0542(02)00274-9
© 2003 Elsevier Science Ltd. All rights reserved.
