The spectrum of pathology in infants with esophageal atresia, with or without TEF, is illustrated in Fig. 51-1. These include esophageal atresia with distal TEF, without TEF, with proximal TEF, and with double (proximal and distal) TEF. Isolated TEF without esophageal atresia, so-called H-fistula that usually occurs in the neck rather than midmediastinum, is also encountered. In most infants, esophageal atresia occurs with a single TEF. This form is observed in approximately 85% of affected infants. The upper esophageal pouch ends blindly, and there is a gap between it and the lower esophageal segment. The fistula from the lower esophageal segment usually enters the trachea just above its bifurcation. The blind upper pouch actually may overlap the nearby lower esophageal segment, or the two ends may be separated by a centimeter or more. Esophageal atresia without TEF is the next most common form and is seen in approximately 8% of infants.
Types and frequencies of esophageal atresia with and without TEF.
A newborn in respiratory distress should be evaluated immediately for esophageal atresia with or without TEF. If there is no air in the abdomen, isolated “long gap” esophageal atresia should be suspected because a neonate normally has gas in the abdomen almost immediately after birth. Conversely, if there is an excess of gas in the stomach and intestine, TEF should be suspected. The presence of esophageal atresia can be ruled out by passing a small soft plastic catheter through the nose into the upper esophagus. Failure of the tube to pass into the stomach is diagnostic of esophageal atresia. One also can make this diagnosis by gently blowing air into the upper pouch to distend it under fluoroscopy. A very small amount of water-soluble contrast material is passed into the tube until the distal end of the upper pouch is visualized. Contrast material never should be injected without fluoroscopic guidance because it can overfill the pouch, causing aspiration. These babies often already have aspiration pneumonia before referral to the surgeon. In the rare case (see Fig. 51-1C or D), an upper pouch fistula may be revealed when contrast material is put into the upper pouch. Note that the upper pouch fistula arises from the anterior wall of the blind upper esophagus, not its end, as occurs with the distal fistula. The clinical importance of this is that the unsuspecting surgeon in such a case may mobilize the upper pouch to an extent sufficient to perform the anastomosis while missing the presence of the second fistula. We have seen several such cases. A clinical clue to the presence of a second upper pouch fistula is the intermittent filling of the pouch with air as the anesthesiologist applies positive-pressure anesthesia through the endotracheal tube.
Other anomalies may be present in a neonate with esophageal atresia, most notably imperforate anus, which is a broad spectrum of anorectal anomalies and congenital heart disease. Appropriate consultation with a pediatric cardiologist may be indicated, particularly today, when the ideal time to repair many congenital cardiac defects, which are soon fatal, may be in the neonatal period.
The surgical approach to esophageal atresia is made through the right chest in most patients. Rarely, if there is a right descending arch, it may be easier to approach the repair through the left chest. Neonatal thoracotomy, which transects the chest wall muscles, can result in impaired growth of the thorax on that side as the child matures. An incision that minimizes muscle cutting is that used by the late David Waterston at the Hospital for Sick Children, London, and is our preference. This is a vertical skin incision beginning just below the axilla and moving downward. The chest wall muscles are retracted upward to enter the fourth interspace. Spending a few minutes to bluntly separate the thoracic wall pleura off of the endothorax, keeping it intact, permits a posterolateral extrapleural exposure of the anatomy to be repaired. It is an important safety feature to locate the esophageal anastomosis extrapleurally to avoid widespread mediastinitis if a leak does occur.
There must be constant, close communication between the anesthesiologist and the surgeon because retraction of the mediastinal structures forward may embarrass cardiopulmonary function. Thus the anesthesiologist should warn the surgeon to release the hand-held retractors if there is any sign of cardiac slowing or if blood gas measurements show deteriorating levels. It is best to anticipate the possibility of these events, which proceed much more rapidly in a frail infant than in older patients. Anticipating a possible cardiac arrest in a baby is better than treating the actual event, which is not always salvageable. The surgeon who undertakes major chest surgery in a small baby must be gentle and must work with an experienced pediatric anesthesiologist.
The azygos vein usually is divided to permit the approach to the esophagus, although we have done this surgery by mobilizing the azygos vein and maintaining it intact. If the vein is large and the baby is small, it is advisable to keep the azygos vein intact.
Early in the operation, the lower esophageal pouch should be identified. The vagus nerves will be seen coming from above and running along the sides of the distal pouch. They should be spared. As soon as the entry of the lower pouch into the back wall of the trachea is identified, it should be encircled with a ligature to stop the air blowing down from the trachea into the distal esophagus and from there into the stomach. If the stomach is already greatly overinflated with air from the fistula, we generally decompress it by placing a mushroom catheter into the anterior wall of the stomach through a minilaparotomy using a vertical incision in the midline approximately 2 to 3 cm in length. If a gastrostomy is used to decompress the stomach, we use the Stamm technique with two purse-string sutures of fine nonabsorbable material.
The tube should be brought out through a separate stab wound lateral to the laparotomy incision, with the added precaution of suturing the stomach to the abdominal wall. We have seen death occur when the stomach was not sutured and a leakage occurred at the gastrostomy site causing the stomach contents to spill into the upper abdomen. This is entirely preventable by anticipating this possibility. Regarding use of a stab wound, we have seen several infants die because infection occurred in the minilaparotomy incision when a tube traversed the wound. Infection caused a wide fistula between the stomach and the outside world, an entirely preventable event. Some of our teachers brought tubes through the main laparotomy incision. However, the surgeon who has seen that result in a major gastric fistula will prefer use of a stab incision for the tube.
We prefer simple ligation of the distal fistula, flush with the back wall of the trachea, and usually use a single ligature of 3-0 or 4-0 size. If the fistula is not ligated flush with the trachea, it leaves a tracheal diverticulum, which can cause intermittent episodes of pneumonitis. We have reexplored several infants in whom such a pouch had been left, another complication preventable by thinking of its possibility when doing the fistula interruption. An alternative method for closing the defect in the trachea is to close it with a row of fine interrupted sutures. The suture must have a tiny needle; otherwise, this closure can leak, whereas a ligature should not.
Mobilization of the distal esophageal segment must be done without injury to its thin, easily torn wall. Dr. C. Everett Koop would not allow a pair of forceps on the operative table when his resident mobilized the distal pouch. A pair of blunt, nontoothed forceps can be inserted into the open end of the distal esophagus, gently spreading it to permit mobilization without grasping its wall.
Conversely, the upper esophagus is usually thick walled. It can be mobilized by passing a traction suture through its tip. This can facilitate, as the surgeon holds it first one way and then the other, to circumferentially mobilize the upper pouch. A useful maneuver, also of great help in mobilizing the upper pouch, is to pass into it a blunt-ended esophageal bougie, which is gently pushed down through the mouth by the anesthesiologist. Again, toothed forceps should be avoided to prevent injury to the infant's esophagus. Alternatively, a Bakes' common duct dilator appropriately bent can exert pulsion on the upper pouch, bringing it into view, especially if it is a high pouch located in the neck.
If the upper pouch seems to intermittently inflate as the anesthesiologist inflates the endotracheal tube, a search should be made for an upper pouch fistula. Mobilization of the upper pouch will disclose that it is often adherent to the adjacent back wall of the trachea. The two structures are separated by careful scissor dissection, staying off the back wall of the trachea, which is thin and easily entered. The esophagus is a much better developed structure because it has been obstructed in utero and is not easily transgressed. If a traction suture has been used on the end of the upper pouch, that part should be excised when the anastomosis is performed.
Early in the development of surgery for esophageal atresia, a two-layer anastomosis was used by many surgeons. The two-layer anastomosis consisted of a full thickness of the delicate distal esophageal opening to the mucosa of the upper pouch imbricating with a second layer of muscle from the upper pouch to the esophageal wall of the lower pouch. Today we prefer a single-layer anastomosis using the cut end of both upper and lower pouches. The back row of sutures is placed using 5-0 or 6-0 nonabsorbable sutures with the knots tied in the lumen of the two ends of esophagus. In practice, we place all those sutures first and defer tying them until the entire back row is in place. The assistant then grasps with large blunt forceps the two ends of esophagus, holding them gently together while the surgeon ties all the sutures with no tension. Most of the anterior wall sutures are then placed and tied simultaneously, again without tension on the end courtesy of the assistant. The last two or three sutures then are necessarily placed such that the tie will be on the outside. At the completion of the anastomosis, we generally place two or three sutures in the muscular wall of the upper pouch to exert gentle traction downward to avoid tugging on the two ends of the anastomosis.
A small flap of mediastinal tissue can be used to cover the ligated and divided fistula at its tracheal end. A soft plastic catheter is placed in the mediastinal gutter well lateral to the anastomosis. Hence, if leakage occurs, it will track out along the tube. We anchor the tip of that tube in the upper gutter because many years ago we had a tube migrate into the esophageal lumen, which was a disaster.
If the postoperative course is uneventful, we obtain an oral contrast study at 7 to 10 days. If the anastomosis is intact, the chest tube is removed from the gutter, and feedings are begun by mouth. Some surgeons prefer to avoid using a gastrostomy tube and postoperatively leave a transanastomotic plastic tube through the nose into the stomach. However, infants breathe through the nose, and a transnasal catheter, in our opinion, can cause pulmonary complications. Therefore, we use them reluctantly. A surgeon who has endured a nasal tube for a week or more is not so apt to order one for someone else. This statement is backed up by ample personal experience!
The chest is closed with pericostal interrupted nonabsorbable sutures of appropriate size for the baby. Many of these infants are premature and require smaller sutures than are used for a full-term 7- or 8-lb infant. Our smallest esophageal atresia patient weighed 1 lb, 2 oz. If muscle has been cut, the ends are rejoined with interrupted sutures. The skin is closed with a running subcuticular suture.
Today, more than 90% of these infants should survive. Mortality is now limited mainly to those with other life-threatening anomalies or extreme prematurity, which itself carries considerable risk.
If postoperative contrast study shows narrowing at the anastomotic site, esophagoscopy and dilation are performed postoperatively. The safest way to dilate the esophagus in a small newborn is to pass a filiform catheter doubled back on itself into the stomach from above. A second filiform catheter is put into the stomach through the gastrostomy site. As the upper filiform catheter is passed into the stomach, pulling up on the transgastric filiform catheter usually will grab the tip of that passed from above, bringing it out of the gastrostomy site. Followers of increasing size then can be brought down from above through the anastomosis to dilate it. One must be careful not to overdilate the anastomosis, for we have had the unhappy experience of splitting a new anastomosis longitudinally, requiring reoperation.
There is a high (approximately 25%) incidence of gastroesophageal reflux in babies after surgery for esophageal atresia. In part, this may be related to excessive mobilization of the distal esophagus when the gap between the distal and proximal ends is longer than usual. Principally, however, the reflux is related to esophageal dysmotility that is common in these patients. An antireflux operation may be indicated if gastroesophageal reflux persists. One of our patients developed carcinoma of the distal esophagus secondary to reflux at an age of 20 years.5 She was treated by distal esophagectomy with thoracoabdominal replacement using transverse colon on left colic pedicle (see Fig. 51-2). The patient is now 39 years old and in excellent health and swallowing normally.
A. Adenocarcinoma in a 20-year-old patient after original esophageal atresia repair. B. Resection and restoration of continuity.
Long-gap Esophageal Atresia
Long-gap esophageal atresia (see Fig. 51-1B) is a special problem. It is possible to stretch the two ends of the esophagus by various means. The upper pouch can be elongated by passing blunt esophageal bougies from above several times daily. The principal problem with leaving the upper pouch intact, instead of marsupializing it, is the risk of aspiration if it is not suctioned dry many times each day. Stretching the lower end is a more difficult problem. A bougie can be passed through the gastrostomy site, but that is not as satisfactory as dilating the upper pouch. Alternatively, a Bakes' dilator can be used with ultrasound guidance to visualize its passage into the lower pouch. The late Fritz Rehbein6 of Germany described a method that involved passing sutures through the ends in such a manner as to gradually bring them together. In 1976, we, along with Dr. Richard Hale of the National Magnet Laboratory at Massachusetts Institute of Technology, described an electromagnetic technique to bring the long-gap ends together.7 We put a metal bullet into each end of the esophagus and placed the baby in an electromagnetic field (Fig. 51-3). Intermittently, we turned the magnet on, pulling the bullets together until the blind ends of the esophagus were apposed. Although this was shown to be effective, reluctance of certain members of the nursing service to participate in this activity led us to discontinue this treatment. The objection raised was the need to keep the infant restrained during the treatment, which lasted 4 to 6 weeks. Therefore, for three decades we used colon esophageal interposition by the technique of Mr. David Waterston, with whom we worked in 1962. Prof. Gunther H. Willital of Munster, Germany, has reported an extensive series of patients treated using electromagnetic bougienage.8
Electromagnetic technique using bullets to bring the ends of the esophagus together in long-gap atresia.
Tracheoesophageal fistula without Esophageal Atresia (H-fistula)
Isolated TEF (see Fig. 51-1E) often escapes clinical recognition during infancy. The history is usually one of intermittent respiratory infections. A contrast swallow with films in the lateral or oblique view can make the diagnosis. Sometimes the fistula is not seen, however, especially if the radiologist has scant experience looking for fistulas. Endoscopy can be diagnostic. Searching the back wall of the trachea with a 30-degree endoscope (this can be either a bronchoscope or even a cystoscope without water) will reveal the tracheal side of the fistula on the back wall of the trachea. It is usually located in the lower neck, not at the level of the carina, as in the most common type of esophageal atresia with TEF. A small catheter passed through the endoscope and through the fistula will prove its presence. A useful maneuver when the tracheal end of the fistula is not obvious is to place an endotracheal tube to intermittently inflate the trachea with positive pressure while searching the anterior wall of the esophagus for telltale signs of air entering the anterior wall of the esophagus. A drop or two of methylene blue placed into the tracheal lumen can further help to reveal its point of entry in the anterior wall of the esophagus.
The isolated fistula is exposed using a transverse cervical incision just above the clavicle. We pass a Fogarty catheter through the fistula, which runs obliquely from the trachea down to the esophagus. The Fogarty balloon is inflated and maintained in the inflated position by cross-clamping the upper end of the catheter to trap the air in the balloon. Gentle tugging on the catheter will show where the balloon is located with minimal dissection. It is easy to inadvertently injure the ipsilateral recurrent nerve when dissecting in the tracheoesophageal groove in search of the H-fistula. Awareness of the recurrent nerve should prevent its inadvertent injury. We have encountered two patients in whom the recurrent nerve “did not recur.” Both had an aberrant right subclavian artery that arose as the last great vessel from the aortic arch and passed from the left mediastinum, behind the esophagus, and out the right thoracic apex toward the arm. Awareness of this rare anatomic entity, in which the recurrent nerve does not loop around the subclavian artery and then pass up the tracheoesophageal groove, is essential. If the surgeon thinks of this, that error will not occur. By having the anesthesiologist intermittently pull back gently on the Fogarty balloon, it is possible to feel it with a finger. This allows the surgeon to dissect exactly to the fistula without encircling the trachea or esophagus. The fistula is doubly ligated and divided. Although the ends are at a slightly different level, a wise precaution to prevent recurrence is to interpose tissue between the divided ends. This is a cardinal principle of closing any type of fistula, wherever it occurs. A strap muscle works well for this purpose. We divide the sternohyoid muscle at its upper end, and rotate the upper end of the muscle backward to cover the esophageal ligature. This takes only a few minutes and provides great safety against recurrence.9 The outlook after this surgery should be a normal baby. We have treated one patient who was 10 years old before the diagnosis was made.
There is another rare communication between the gut and the tracheobronchial tree that should be mentioned. The trachea and the primitive gut are adjacent to each other early in development. If there is a persisting communication between the two, as the gut lengthens, that communication can be stretched all the way from the neck down through the thorax, where it can arise from the jejunum or even the biliary tree. We encountered one of these cases, after the initial workup failed to reveal a communication. Repeating the barium study and following the contrast material down to the duodenum disclosed the occult communication. It ran upward and entered the right lower lobe of the lung. Using postural drainage for several years in this 5-year-old boy had been of no benefit in controlling his obvious intermittent aspiration pneumonitis and was exactly the wrong thing to do. By removing that bowel communication, we were able to stop his pneumonia. Severe pulmonary clubbing of the fingernails disappeared. When last seen 45 years ago, he was completely well.10