The main congenital structural abnormality of the trachea seen in infants and children is tracheo-oesophageal fistula and oesophageal atresia. Other rare congenital abnormalities include absence of the oesophagus, oesophageal duplication cyst, diverticulum, web and stenosis, and laryngotracheal cleft. These will not be further discussed in this chapter, which will concentrate on tracheo-oesophageal fistula, oesophageal atresia, and the very important and controversial disorder of gastro-oesophageal reflux and hiatus hernia. This chapter is largely based on the experience gained in the management of 192 consecutive patients seen at Our Lady's Hospital for Sick Children, Dublin between 1974 and 1988.

Oesophageal atresia, with or without fistula, is the most common congenital anomaly of the oesophagus, occurring in one per 3000 to 4500 live births. The cause of the anomaly is not known, though there is a general consensus that genetic factors are not involved. The condition is more likely to be the result of a major embryological insult: approximately 50 per cent of affected infants have other congenital anomalies, and these are a major cause of death. Several acronyms have been described to include particular associations or groupings of congenital defects: VATER denotes vertebral, anal, tracheo-oesophageal, and renal abnormalities, while VACTERL includes cardiac and limb defects. CHARGE is a new acronym associated with tracheo-oesophageal fistula (TOF), which includes eye defects and mental retardation as well as renal and cardiovascular abnormalities.

In 1967 Waterston identified the birth weight of the infant, the early presence of sepsis, and the presence or otherwise of associated congenital abnormalities as three important prognostic indicators for infants with tracheo-oesophageal fistula. This classification is of historical interest, since medical advances now allow 90 per cent of affected infants to survive.

The classification of tracheo-oesophageal fistulae and their incidence are illustrated in Fig. 1 2055. In the most common anomaly, air passes along the fistula into the upper gastrointestinal tract, where it may cause progressive abdominal distension, resulting in respiratory compromise. Efforts to feed the infant cause coughing, choking, and even cyanosis. The diagnosis should be entertained in any newborn infant presenting at birth with respiratory problems, particularly if the mother had polyhydramnios and/or if other obvious congenital abnormalities are present in the infant. Early diagnosis before feeding is desirable, and will help to prevent pneumonitis developing as a consequence of reflux of gastric contents through the fistula, or following the aspiration of saliva and offered feed. Diagnosis can be made by passing a semi-rigid (8–10 Fr) nasogastric tube through the mouth or nose. If tracheo-oesophageal atresia and fistula is present the tube will not enter the stomach but will curl up in the proximal pouch fundus, perhaps reappearing at the mouth. With the tube in position, erect and lateral radiographs should be taken to include the neck, chest, and abdomen (Fig. 2) 2056. The proximal pouch can usually be seen with the tube lying in it. The radiographs can be used to judge the level to which the fundus of the proximal pouch has reached—usually the second or third thoracic vertebra. The presence of gastric and intestinal air on the abdominal film confirms the presence of a distal fistula, and an associated duodenal obstruction will cause a double bubble appearance. The absence of air in the abdomen in the presence of a proximal blind oesophageal pouch suggests the presence of pure oesophageal atresia without a fistula. The radiographs also show evidence of associated cardiac or vertebral abnormalities and respiratory infection. Administration of a dilute water-soluble contrast medium through a tube into the upper pouch to give a more accurate image of its location is not favoured, because of the risk of aspiration. Before operation the presence of associated abnormalities should be investigated by echocardiograms and ultrasound studies of the abdomen, paying particular attention to the kidneys. Blood is taken for a cross-match and for estimation of coagulation profile, full blood count, and urea and electrolyte levels. Vitamin K is given (1 mg IM) if not already administered.

The condition of the baby on arrival at the tertiary referral hospital is of great importance in the successful management of tracheo-oesophageal fistula. Proper neonatal transfer reduces the risk of pneumonitis from aspiration and acid regurgitation through the fistula into the airway. The infant should be nursed either prone or in the semi-erect position at 45°, with a sump suction catheter in the proximal pouch on low pressure continuous suction. If this is not available, a regular nasogastric tube aspirated every 15 to 30 min will suffice. Antibiotics (a cephalosporin and an aminoglycoside) should be administered and the baby transferred in an incubator in an atmosphere of humidified oxygen.

On arrival the baby should be evaluated with regard to pulmonary status, weight, maturity, and the presence or absence of congenital anomalies and their severity. Primary repair of the fistula is undertaken if the baby is healthy, has no respiratory compromise, and has no other life-threatening anomaly. If there is appreciable pneumonitis and no abdominal distension, surgery can be delayed to allow improvement in respiratory function. Abdominal distension causing respiratory compromise may be an indication for early intervention.

A staged approach to repair, while popular in the past, is rarely used today, and low birthweight in itself is not regarded as a contraindication to primary repair. Primary repair in the sick, premature, low birthweight infant may be delayed until parenteral nutrition has produced a positive nitrogen balance, and until antibiotic therapy and intensive pulmonary toilet have been undertaken. Ligation of the fistula only, without attempted repair, is also occasionally indicated in the sick, premature infant who develops idiopathic respiratory distress syndrome and who requires ventilation. In such infants, the high pressures required for ventilation forces air through the fistula, increasing abdominal distension. Primary repair is performed when the pulmonary status has improved.

Bronchoscopy and oesophagoscopy prior to surgery provide information on the site of the fistula, the length of the pouch, and the degree of tracheomalacia; they also exclude the presence of an upper pouch fistula. The operation is performed through a right fourth interspace posterolateral thoracotomy. A gastrostomy is rarely performed unless there is a problem with the anastomosis or pure oesophageal atresia is present.

Following the division of the intercostal muscles, the parietal pleura is gently stripped from the rib using gauze or finger tip dissection. When a large area is free of pleura, a rib spreader is inserted, permitting an extrapleural approach to the fistula. The spreader is opened slowly and adherent pleura is further swept from the curve of the ribs in a posteromedial direction to reveal the azygos vein, which is divided between ligatures. This exposes the area of the fistula and lower oesophageal segment. Identification of this may be straightforward, but the lower segment may be short or obscured by a right-sided aorta. The fistula and lower segment will frequently be seen to distend with inspiration. A tape is placed around the lower oesophagus, avoiding excessive dissection, and with traction the fistula is traced to its junction with the trachea. A right-angled clamp is placed on the fistula, close to the trachea, and the fistula is divided, leaving the lower pouch to fall open into the mediastinum. The fistula stump is then ligated with one 3/0 silk ligature.

Identification of the upper pouch is facilitated by intermittent pressure by the anaesthetist on a previously positioned large upper pouch tube. A traction suture is placed in the tip of the pouch and the pouch dissected as high as possible, taking care not to enter the closely adherent trachea. An axial, single layer anastomosis is now performed using a 5/0 suture (Prolene, Maxon, PDS). There is always considerable disparity in size and thickness between upper and lower pouches; there may also be a considerable gap. The upper pouch is opened at its apex and the anastomosis begun by placing two sutures at the lateral extremes, and one in the centre, through all layers of the upper and lower pouches. Additional sutures may then be inserted between these primary sutures. A fine Silastic nasogastric feeding tube is passed across the posterior wall into the stomach and the anterior wall is completed. It is important not to insert too many sutures: four or five in the anterior wall will generally suffice. An intercostal, extrapleural drain is then positioned close to, but not sitting on, the anastomosis. The thoracotomy wound is closed using absorbable sutures.

Diagnosis of a H-type fistula (Fig. 1) 2055 is often delayed for months or years, since its presentation is atypical. The most precise way to establish the diagnosis is by bronchoscopy, at which time the opportunity may be taken to pass a fine ureteric or Fogarty type catheter across the fistula. This will help the surgeon locate the fistula, which is approached through a supraclavicular right transverse incision. This exposes the tracheo-oesophageal area and allows the endoscopically placed catheter to be palpated. The fistula is then identified and divided. An interposition of muscle (omohyoid) or adventitial tissue may reduce the risk of recurrence. Other aspects of the technical approach to these repairs are discussed in Section 36.3 290.

The most difficult problem to overcome is a gap too long to permit an end-to-end anastomosis. This is most common when pure atresia is not associated with a fistula. A number of elaborate methods have been described to overcome this. The best method, in our experience, is the ‘wait and grow’ regimen. This programme requires establishment of a feeding gastrostomy and nursing of the infant in the head down position, with suction to the upper pouch to aspirate secretion and saliva. The length of the gap is measured at intervals radiologically, using metal boughs introduced simultaneously through the gastrostomy and orally as markers. When the gap is less than 3 cm axial anastomosis is attempted. Circular myotomy may be necessary to achieve this.

Every effort must be made to avoid the use of stomach, colon, or large bowel as a substitute for the oesphagus. This goal can usually be achieved, even if a salivary fistula develops due to tension at the anastomosis. If the infant's nutrition is maintained by total parenteral nutrition and/or jejunal feeding and sepsis is controlled, the fistula usually heals and any resultant stricture can be dealt with subsequently.

Frequent gentle oropharyngeal suctioning is performed using a catheter marked to ensure that the tip remains just proximal to the anastomosis. Intravenous fluids and antibiotics are continued and transanastomotic tube feeds begun on the first postoperative day. A contrast swallow is obtained on day 7 to 10. If no leak is demonstrated, the extrapleural drain is removed and oral feeds instituted. The transanastomotic tube is removed if oral feeds are tolerated. Gentle bouginage is performed on day 14 if the anastomosis shows significant narrowing.

The need for assisted ventilation postoperatively is governed by the respiratory status of the infant, its birth weight, and the technical difficulty of the anastomosis. A tense anastomosis is probably protected by full respiratory support for some days, avoiding energetic respiratory efforts and coughing by the infant.

Overall mortality is closely related to the severity of associated anomalies. Table 1 561 depicts the results of 192 consecutive patients with tracheo-oesophageal fistula, seen in Dublin between 1974 and 1988. Accumulated mortality at the end of 1979, 1983, and 1988 was 25.5 per cent, 11.1 per cent, and 5.6 per cent, respectively, with an overall 15-year mortality of 19.8 per cent. Nowadays an overall survival rate of 90 to 95 per cent is expected.

Respiratory problems in the form of stridor, cough, and frequent respiratory infections are common in children who have undergone repair of a tracheo-oesophageal fistula. The many causes include tracheomalacia, gastro-oesophageal reflux, oesophageal dysmotility, persistent or recurrent fistula, and congenital heart disease. Investigations should consider each of these possible causes and should include a full cardiorespiratory evaluation.

Recurrent fistula should be suspected if the infant suffers from recurrent pneumonia and chokes at feeds. Aggressive investigation and therapy are essential. Diagnosis is by bronchoscopy or contrast oesophagram. A transpleural repair of the fistula is required. Balloon dilatation is probably the preferred method of dealing with an established stricture. Mild narrowing responds readily to bouginage. Surgical resection of the stricture is occasionally necessary.

Gastro-oesophageal reflux is relatively common after repair of a tracheo-oesophageal fistula. The place of antireflux surgery in its management is controversial, but in the authors' view is rarely required. Only 12 (8 per cent) of patients in our series required an antireflux procedure.

Detailed oesophageal function studies have shown that all patients with a repaired tracheo-oesophageal fistula have some form of oesophageal dysmotility. This may be the most important factor in the development of recurrent respiratory problems. Treatment is supportive, with thickening of feeds and propping up of the infant after feeding. Symptoms improve with time.

Tracheomalacia probably results from inherent weakness of the trachea caused by deficient cartilage. Compression of the soft trachea between the aortic arch in front and the hypertrophied pouch behind contributes to approximation of the tracheal walls during expiration. The condition can be life-threatening, causing cyanosis, respiratory arrest, bradycardia, and even cardiac arrest. These episodes occur at or shortly after feeding, but may not commence until some weeks after repair. Diagnosis is by a combination of radiological and bronchoscopic examination. Prolonged intubation or tracheostomy or aortopexy (anterior suspension of the aorta to the posterior sternal surface) is indicated in severely affected infants.

Oesophageal stricture secondary to the swallowing of a corrosive liquid is a devastating injury now rare in the Western world. The victim is usually a young child who drinks from a bottle of corrosive liquid that has been carelessly stored.

Treatment is aimed at preventing the development of a stricture. The extent of mucosal injury should be evaluated by endoscopy within 24 h after the incident. Experimental evidence not really supported by firm clinical results suggests that the early systemic administration of steroids and antibiotics reduces the inflammatory response and the morbidity. Whether or not steroids are administered, all patients with endoscopic evidence of mucosal injury must be very carefully monitored. Replacement of the damaged oesophagus with a gastric, small intestine or colonic graft may be required.

In 1947 Neuhauser and Berenberg drew attention to a condition which they called chalasia. This was due to laxity, or in more modern terms immaturity, of the cardio-oesophageal junction, resulting in infantile vomiting. Since then the diagnosis, clinical significance, investigation, and treatment of gastro-oesophageal reflux has been widely investigated. This has contributed a great deal to the current management of the condition, and has also identified points of controversy.

It is universally agreed that gastro-oesophageal reflex is the most common cause of recurrent persistent vomiting in infancy. The severity and frequency of vomiting diminish usually as the baby develops, even without treatment. Most infants cease to have a problem by 3 months of age, and a majority of the remainder will be asymptomatic by 1 year of age. In a small percentage of patients, however, troublesome symptoms persist indefinitely. Even in those who recover rapidly, the vomiting and respiratory complications of the condition can cause a great deal of harm during the symptomatic period. There is clear understanding and ready acceptance of the ill-effects of vomiting; but much less agreement on the significance of the respiratory symptoms.

Vomiting results in loss of calories and, if untreated, the infant fails to thrive, shows poor weight gain, and does not attain its expected physical and even intellectual milestones. Apart from poor weight gain, serious harm may occur because of the corrosive effect of regurgitated acid gastric contents on the stratified squamous epithelium of the oesophageal lining. Regurgitation produces oesophagitis and sometimes ulceration of the oesophagus, which may result in the vomitus containing altered (‘coffee grounds’) or fresh blood. Occult bleeding may also cause anaemia. Inflammation may extend into the oesophageal submucosal and muscle layers: oedema and eventual circumferential and longitudinal fibrosis are the natural reparative response to such inflammation, and a shortened, strictured oesophagus may result. This complication occurs in 5 per cent of untreated patients. Since gastro-oesophageal reflux is now treated actively, progression to stricture formation is rare. Occasionally vomiting may not be a dominant symptom, and a child may present with dysphagia due to a stricture, a consequence of chronic occult gastro-oesophageal reflux. The severity of symptoms shows great interpatient variation, and this dictates the intensity of investigation and treatment. A significantly affected infant typically has a history of persistent vomiting, is underweight, and may be anaemic.

The respiratory symptoms and complications which are said to derive from gastro-oesophageal reflux are controversial. The most important respiratory event is near-miss sudden infant death, also referred to as apparent life-threatening events. The suggestion of a relationship between gastro-oesophageal reflux and apparent life-threatening events is of great importance, since the latter may be recurrent, and can result in brain damage and even death in up to 2 per cent of those so affected. It is suggested that, in addition to aspiration of gastric contents into the airways, the presence of acid in the oesophagus initiates reflux obstructive laryngospasm, bronchospasm, reflex centre apnoea, and bradycardia, resulting in profound hypoxia, and sometimes death. The highly sophisticated investigative techniques which can identify silent reflux in infants suggest that there is an increased incidence of gastro-oesophageal reflux in babies with near-miss sudden death. Further research is necessary before the relationship between gastro-oesophageal reflux and apparent life-threatening events can be confirmed. Gastro-oesophageal reflux has also been proposed to play a role in other respiratory conditions, such as pneumonia, recurrent bronchitis, wheeziness, and asthma, particularly in infants. It is suggested that this arises from the aspiration of gastric contents into the lung fields; more recently a possible reflex or neurohumoral mechanism has been proposed. The role of gastro-oesophageal reflux in these respiratory events has not been widely accepted in Europe, although such a causal relationship is more widely accepted in North America. This different attitude has important therapeutic consequences: in North America antireflux surgery is frequently proposed as the preferred method of treatment, while in Europe it is rarely recommended or practiced.

Sandifer's syndrome, in which the patient shows spasmodic contortions of the head, similar to that seen in torticollis, is a very rare clinical manifestation of gastro-oesophageal reflux which is relieved by elimination of reflux.

Appreciable gastro-oesophageal reflux which may occur in children with cerebral palsy or Down's syndrome is often diagnosed late, because the primary concern relates to the presenting condition.

Techniques used in the diagnosis of gastro-oesophageal reflux include barium studies and fluoroscopy, which may be combined with oesophagoscopy. Tests used in the investigation of adult oesophageal problems, including oesophageal manometry, long (24-h) or short (3-h) intraluminal monitoring of pH of the oesophagus, and oesophageal clearance studies are applicable to children. These investigations are frequently combined with continuous recording of the child's respirations. The acid reflux test described by Tuttle is also used, as is technetium scintiscanning, in which radioactive material is placed in the stomach and the oesophageal and lung fields are subsequently scanned for evidence of reflux and aspiration. It is particularly relevant to patients who have respiratory symptoms associated with gastro-oesophageal reflux.

The current variety of investigative techniques has made it easy to diagnose gastro-oesophageal reflux. All infants reflux at some time, the frequency usually diminishing as they grow. It may be difficult to distinguish reflux which is pathological and requires treatment from that which is physiological and may be ignored. Many clinicians consider only reflux, and ignore or consider of little importance the presence of a sliding hiatus hernia (partial thoracic stomach). A reversal of this attitude is required. In one series of 1000 infants with radiological evidence of reflux but no partial thoracic stomach, none developed a stricture or required surgery. However, 30 per cent of those with gastro-oesophageal reflux and hiatus hernia suffered from troublesome symptoms for 4 years or more, and 5 per cent developed a stricture (Fig. 4) 2058.

While it may be an oversimplification to maintain that the distinction between pathological and physiological reflux can be made on the basis of the presence or absence of a hernia (when taken in conjunction with the symptomatology), it seems equally extreme to claim that the presence of such a hernia is irrelevant. It is, however, certainly true that reflux can occur without the presence of a hiatus hernia and alternatively a hiatus hernia may be present without significant reflux. The frequency with which a hiatus hernia is detected in infants with reflux is influenced by the skill and persistence of the radiologist. If the presence of a hernia is not accorded a significant role in the aetiology of the condition, it may not be sought.

Since infantile gastro-oesophageal reflux, with or without hiatus hernia, usually resolves spontaneously as the baby matures, time is the essence of management. Medical treatment should always be tried before an operation is considered, whether the dominant symptom is excessive regurgitation and vomiting, respiratory distress, or a combination of both. The enthusiasm, conviction, and efficiency with which medical measures are applied greatly influence the proportion of patients subjected to surgery, and explain the disparity seen in reports of the role of surgery in gastro-oesophageal reflux. A very small proportion of infants under 12 months of age will require surgery: in one series of 1525 patients under 14 months of age only 4.2 per cent were operated on. Children over 14 months of age represent those whose problems have not resolved with time or who have presented late. Surgery is required in 20 per cent. One of the pillars of conservative management is maintaining the infant in a head-up position of 45 to 60° 24 h a day for at least 3 months. There is disagreement over whether this should be in the supine or prone position. Studies of oesophageal pH suggest that the supine position makes reflux worse. The preferred regimen, therefore, is to nurse the very young infant prone, with the head end of the cot elevated. As the baby develops and becomes mobile, this is not practical during waking hours, and the propped up sitting position should be used. The prone head-elevated position should be adopted for sleeping. The semiseated position allows the infant to relate more easily to its stimulating environment, which is of increasing importance as the baby matures. Thickening of the infant's feeds, and early weaning on to semisolid and spoon feeds also diminish the frequency and amount of vomiting. Vitamin and iron supplementation may also be indicated. The use of other pharmacological agents will be influenced by the severity of the symptoms, their response to other standard treatments, and objective radiological and endoscopic evidence of persisting oesophagitis. Treatment with an H&sub2;-receptor antagonist, and in recent times omeprazole, and an antacid is indicated for persistent oesophagitis, particularly when surgery is contemplated. Drugs such as bethanacol, domperidone, metoclopramide and, most recently, cisapride, stimulate oesophageal motility and increase oesophageal sphincter pressure and gastric emptying; these have been used in some infants with gastro-oesophageal reflux. Results obtained with metoclopramide have been contradictory but generally favourable; cisapride has attracted increasing interest in the past 2 years, with encouraging results. Further clinical experience is necessary, however before the use of these drugs in the management of gastro-oesophageal reflux is established. At the present time, their use should probably be limited to specialists and specialized institutions and to the minority of infants who do not respond to other treatment or who experience apparent life-threatening episodes.

Antireflux surgery, which is usually very successful, has an important role in the management of gastro-oesophageal reflux, but should only be undertaken in the small minority of patients (5–10 per cent) who do not respond to expert medical management. Surgery is indicated more frequently in infants with a hiatus hernia and in older children whose troublesome symptoms persist, even if they are not severe, or who develop complications secondary to reflux. Radiological or endoscopic evidence of narrowing of the oesophagus, signalling incipient stricture, is an indication for intervention. Surgery has a particular place in the management of neurologically damaged children and in those with Down's syndrome, not only to improve weight gain but also as an aid to the general nursing or parental care of these children, who need frequent changes of clothing, because of their persistent vomiting and drooling.

Surgery is also indicated if extensive investigation suggests that respiratory symptoms are due to gastro-oesophageal reflux. Although clinicians believe this to be particularly true of infants with frightening recurrent apnoea spells, the majority of these episodes can be managed successfully by conservative methods. The most popular operation used in children is the Nissen fundoplication, which is usually performed through an upper abdominal midline incision. Modifications of the standard fundoplication operation which may be used in infants and children involve a partial rather than a complete wrap of the gastric fundus around the terminal oesophagus. This is aimed at preventing or minimizing the uncomfortable bloat syndrome, seen in some patients following full fundoplication.

Some surgeons recommend pyloroplasty in selected patients in association with the antireflux procedure, on the basis that gastro-oesophageal reflux is associated with an increased incidence of delayed emptying of the stomach. There is controversy over the existence of such delay in emptying and if it is present it probably resolves following antireflux surgery.

Surgical correction of gastro-oesophageal reflux, with or without hiatus hernia, produces good results. The recurrence of symptoms sufficiently severe to require reoperation is uncommon, with an incidence of zero to 10 per cent. Occasionally patients suffer from postoperative dysphagia, which may be due to too tight a repair: this usually responds to oesophageal dilation. Inability to burp or vomit, and the ‘gas bloat’ or a feeling of abdominal distension, may also occur. The incidence of these symptoms is also related to the tightness of the repair. The effects and incidence of bloat can be minimized if the hiatal repair and fundal wrap are performed around a large intraluminal bougie. The size of the bougie will vary with the size of the child but should not be smaller than 20 FG and may be as large as 35 FG in the older child. This does not produce a high incidence of recurrent reflux: only four of 44 patients (8 per cent) required reoperation in our series.

The quoted incidence of stricture associated with reflux oesophagitis varies. This is understandable since some series report on the incidence as it presents in patients who come to surgery, when a relatively high incidence is to be expected, while other reports relate the incidence in all patients, both medical and surgical, seen and treated for oesophagitis. The 5 per cent incidence reported by Carre would be in this category. The higher reported incidence of stricture in Europe compared to North America is a result of the different clinical and investigative attitudes to hiatus hernia and gastro-oesophageal reflux. Undoubtedly, the best treatment of stricture of the oesophagus is prevention. The formation of a true fibrous stricture is usually associated with shortening of the oesophagus. This combination is a prescription for long-term ill health and, not infrequently, repeated surgery to resect the stricture. Stricture formation can be prevented by careful clinical, radiological and, if necessary, endoscopic supervision of patients with persistent vomiting associated with reflux. Those who do not respond to medical measures in whom vomiting persists should undergo repeated barium oesophagogram and endoscopy with mucosal biopsy. The term ‘stricture’ implies a fibrous or rigid narrowing of the oesophagus seen radiologically or endoscopically; the narrowing is, however, usually due to oedema and muscle spasm which resolves following antireflux surgery and subsequent balloon or bougie dilation. The development of balloon dilatation has been a major advance in the management of oesophageal stricture.

Patients who present with a mature fibrous stricture generally respond to an antireflux procedure, and only very rarely is a direct operation on the stricture, such as resection and anastomosis, necessary.

Amoury RA. Structure and function of the esophagus in infancy and early childhood. In: Ashcraft KW, Holder TM, eds. Pediatric Esophageal Surgery. Orlando: Grune and Stratton 1986: 1–28.

Boix-Ochoa J. Gastro-oesophageal reflux in children. In Jamieson GG, ed. Surgery of the Oesophagus, Edinburgh: Churchill Livingstone, 1988.

Carre IJ. Natural history of partial thoracic stomach (‘hiatus hernia’) in children. Arch Dis Child 1959; 34: 344–53.

Cohen S, Harris LD. Does hiatus hernia affect competence of the gastroesophageal sphincter? N Engl J Med 1971; 284: 1053–6.

Consensus statement. National Institutes of Health consensus development conference on infantile apnoea and home monitoring. Pediatrics 1987; 79: 292–9.

Corbally UT, Muftah M, Guiney EJ. Nissen fundoplication for gastro-oesophageal reflux in repaired tracheo-oesophageal fistula. Eur J Paediatr Surg 1992; 2: 332–5.

Editorial. Gastro-oesophageal reflux and apparent life-threatening events in infancy. Lancet 1988; i: 261–2.

Ein SH, Shandling B, Wesson D, Filler RM. Esophageal atresia with distal tracheoesophageal fistula; associated anomalies and prognosis in the 1980's. J Pediatr Surg 1989; 24: 1055–9.

Hoffer FA, Winter HS, Fellows KE, Folkman J. The treatment of postoperative and peptic strictures after esophageal atresia repair. A program including dilatation with balloon catheters. Paediatr Radiol 1987; 17: 454–8.

Holder TM. Esophageal atresia and tracheoesophageal fistula. In: Ashcraft KW, Holder TM, eds. Pediatric esophageal surgery. Orlando: Grune and Stratton, 1986: 29–52.

Holder TM, et al. A survey of its members by the surgical section of the American Academy of Pediatrics. Pediatrics 1974; 34: 542–9.

Jeffery HE, Rahilly P, Read DJC. Multiple causes of asphyxia in infants at high risk for sudden infant death. Arch Dis Child 1983; 92: 73–5.

Kosloske AM, Jewell PF, Cartwright. Crucial bronchoscopic findings in esophageal atresia and tracheoesophageal fistula. J Pediatr Surg 1988; 23: 466–70.

Machida HM, Forbes DA, Gall DG, Scott RB. Metoclopramide in tracheo-oesophageal reflux of infancy. J Pediatr 1988; 112: 483–7.

Meyers WF, Herbst JJ. Effectiveness of positioning therapy for gastroesophageal reflux. Pediatrics 1982; 69: 768–72.

Neuhauser EBD, Berenberg W. Cardioesophageal relaxation (chalasia) as a cause of vomiting in infants. Pediatrics 1950; 5: 414–20.

Orenstein SR, Orenstein DM. Gastroesophageal reflux and respiratory disease in children. J Pediatr 1988; 112: 847–58.

Orenstein SR, Whitington PF, Orenstein DM. The infant seat as treatment for gastroesophageal reflux. N Engl J Med 1983; 309: 760–3.

Puri P, Blake NS, O'Donnell B, Guiney EJ. Delayed primary anastomosis following spontaneous growth of esophageal segments in esophageal atresia. J Pediatr Surg 1981; 16: 180–3.

Puri P, Ninan GK, Blake NS, Fitzgerald RJ, Guiney EJ, O'Donnell B. Delayed primary anastomosis for esophageal atresia: 18 months' to 11 years' follow-up. J Pediatr Surg 1992; 27: 1127–30.

Randolph JG, Newman KD, Anderson KD. Current results in repair of esophageal atresia with tracheoesophageal fistula using physiological status as a guide to therapy. Ann Surg 1989; 209: 526–30.

Reyes HM, Meller JL, Loeff D. Management of esophageal atresia and tracheoesophageal fistula. Clin Perinatol 1989; 16: 79–84.

Shaul DB, Schwartz MZ, Marr CC, Tyson KR. Primary repair without routine gastrostomy is the treatment of choice for neonates with esophageal atresia and tracheoesophageal fistula. Arch Surg 1989; 124: 1188–90.

Silen U, Hagberg S, Rubenson A, Werkmaster K. Management of esophageal atresia: review of 16 years experience. J Pediatr Surg 1988; 23: 805–9.

Spitz L, Kiely E, Brereton RJ. Esophageal atresia: five year experience with 148 cases. J Pediatr Surg 1987; 22: 103–8.

Tolia V, Calhoun J, Kuhns L, Kauffman RE. Randomized, prospective double-blind trial of metoclopramide and placebo for gastroesophageal reflux in infants. J Pediatr 1989; 115: 141–5.