Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Esophageal atresia (EA) and tracheoesophageal fistula (TEF) anomalies present the pediatric surgeon with a unique and complex congenital disease, which tests both the diagnostic and technical skill of the surgeon. Most pediatric surgeons consider the surgical correction of these malformations to be the height of neonatal surgical care. In 1959, Dr. Willis Potts wrote, “To anastomose the ends of an infant’s esophagus, the surgeon must be as delicate and precise as a skilled watchmaker. No other operation offers a greater opportunity for pure technical artistry.” Although this statement still remains true, improvements in anesthetic and neonatal intensive care have made repair of these anomalies and their postoperative management much more routine so that a good outcome can be achieved in most cases. Technical advances, including the application of the minimally invasive approach, have also decreased the morbidity from these operations.
The first report of EA was by Durston in 1670, who found a blind upper pouch in one of a pair of thoracopagus conjoined twins, but the initial classic description was by Thomas Gibson in 1697. However, it was not until 1939 when a baby with EA/TEF survived following successful staged repairs described separately by Leven and Ladd. In 1940, Haight described the first survival following primary anastomosis. By the mid-1980s, most neonatal centers were performing primary repair and reporting successful outcomes in up to 90%.
The embryology of the foregut is still subject to controversy. What is known, however, is that during the fourth week of gestation the foregut starts to differentiate into a ventral respiratory part and a dorsal esophageal part. The laryngotracheal diverticulum then invaginates ventrally into the mesenchyme. The traditional theory postulates that the ventral respiratory system separates from the esophagus by the formation of lateral tracheoesophageal folds that fuse in the midline and create the tracheoesophageal septum. At 6–7 weeks of gestation, the separation between trachea and esophagus is complete. Incomplete fusion of the folds results in a defective tracheoesophageal septum and abnormal connection between the trachea and esophagus.
This theory of longitudinal tracheoesophageal folds merging to form a septum has been challenged. In chick embryo studies, these folds could not be demonstrated. Instead, cranial and caudal folds were found in the region of tracheoesophageal separation. According to this theory, EA/TEF would then be due to an imbalance in the growth of these folds. Furthermore, rat studies suggest that EA/TEF results from disturbances in either epithelial proliferation or apoptosis.
More recent studies show that ectopic expression of sonic hedgehog occurs in the tissues between the notochord and the gut. Knockout mice models have helped elucidate the functions of different genes in the development of foregut aberrations such as EA/TEF. Also, the relationship between BMP4 (bone morphogenic protein) and Nog , the gene encoding noggin (which is a BMP antagonist), may also have an impact on the development of EA/TEF.
The birth incidence of EA/TEF is 1 in 2500–3000 live births. There is a slight male preponderance of 1.26:1. There is no evidence for a link between EA/TEF and maternal age when chromosomal cases are excluded. The risk for a second child with EA/TEF among parents of one affected child is 0.5–2%, increasing to 20% when more than one child is affected. The empirical risk of an affected child born to an affected person is 3–4%. The relative risk for EA/TEF in twins is 2.56 when compared with singletons. The concordance rate in twins is low, but the risk among twins of the same gender is high.
Environmental factors that have been implicated include the use of methimazole in early pregnancy, prolonged use of contraceptive pills, progesterone and estrogen exposure, maternal diabetes, and thalidomide exposure. EA is occasionally seen in fetal alcohol syndrome and in maternal phenylketonuria.
Chromosomal anomalies are found in 6–10% of the patients. The total number of trisomy 18 cases exceeds the total number of trisomy 21 cases. As the incidence of trisomy 18 is higher, it would seem to indicate that trisomy 18 is a greater risk for EA development. Three separate genes have been associated with EA/TEF: MYCN haploinsufficiency in Feingold syndrome, CHD7 in CHARGE syndrome, and SOX2 in the anophthalmia–esophageal–genital (AEG) syndrome.
EA may occasionally be part of the Opitz G/BB syndrome, oculo-auriculo-vertebral syndrome, Bartsocas–Papas syndrome, Fryns syndrome, and in Fanconi anemia.
The factor or factors responsible for the early disturbance in organogenesis causing EA may affect other organs or systems that are developing at the same time. EA can be divided clinically into isolated EA and syndromic EA, occurring at roughly the same rate.
The most frequent associated malformations encountered in syndromic EA are:
Cardiac (13–34%)
Vertebral (6–21%)
Limb (5–19%)
Anorectal (10–16%)
Renal (5–14%)
Vertebral anomalies are confined mainly to the thoracic region. An earlier claim that the presence of 13 pairs of ribs is a good indicator of long-gap EA has not been substantiated.
Nonrandom associations have been documented as well. Two of these are the VACTERL association ( V ertebral, A norectal, C ardiac, T racheo- E sophageal, R enal, and L imb abnormalities), and the CHARGE association ( C oloboma, H eart defects, A tresia of the choanae, developmental R etardation, G enital hypoplasia, and E ar deformities). In 1973, VACTERL was originally described as VATER, an acronym made up of V ertebral Defects, A nal Atresia, T racheo- E sophageal fistula with EA, and R adial dysplasia. It was later extended with the C for cardiac anomalies and the L for limb anomalies. In a cohort of 463 patients with EA, 107 (23%) had at least two additional VACTERL defects. Seventeen of these patients had a chromosomal defect or a syndrome without a known genetic defect. Interestingly, as many as 70% of the remaining 90 patients in this study had additional defects other than VACTERL anomalies.
EA and TEF present in many forms, and various classification systems have been used to describe them. It is clear that EA should be thought of as a spectrum of anomalies ( Fig. 27.1 ). The original classification system was devised by Vogt in 1929. Ladd put forth his own classification in 1945 and Gross revised this schema in 1953. These classifications tend to be confusing, as the same subclasses are named differently. For clarity, it seems much better to give descriptive names to the major subtypes.
This is the most common subtype, accounting for about 85% of EA anomalies. In this anomaly, the very dilated proximal esophagus has a thickened wall and descends into the superior mediastinum, usually to a point between the second to third or fourth thoracic vertebrae. The distal esophagus is slender and has a thin wall. It enters the trachea posteriorly either at the level of the carina or 1–2 cm higher. The distance between the esophageal ends varies from very small to quite wide. Very rarely, the distal fistula may be occluded, leading to the misdiagnosis of EA without distal fistula.
Pure EA has an incidence of about 7%. The proximal and distal esophagus end blindly in the posterior mediastinum. The proximal end is dilated and has a thickened wall as in the more common EA/TEF. If there is no concomitant proximal fistula, the upper esophagus ends at the level of the azygos vein. The distal esophagus is short and often suspended by a fibrotic band. The distance between the two segments is considerable, usually precluding immediate anastomosis.
H-type TEF without atresia is usually discussed together with EA because it may be part of the VACTERL association. It occurs with an incidence of about 4%. The fistula starts from the membranous trachea and runs caudad to enter the esophagus. Normally it is short, although the diameter may be variable. The fistula is usually situated at the thoracic aperture or higher in the neck.
The association of a proximal fistula in a patient with pure EA is thought to be about 2% but may be higher than is generally appreciated. In a series of 13 children with EA, but without a distal fistula, a proximal fistula was found in seven. An upper esophageal fistula is usually not found at the end of the pouch. This fistula is similar to the H-type, which starts proximally on the trachea and ends distally in the dilated proximal esophagus. Usually, there is only one proximal fistula, but two or three have been described. The fistula is usually located at the thoracic aperture or higher in the neck. Although limited in length, its diameter may vary from tiny to large. If not diagnosed preoperatively, it may be suspected during operative repair when bubbles are seen on opening the proximal esophagus.
The incidence of EA with proximal and distal fistulas is thought to be <1%. EA with one distal fistula and two proximal fistulas has also been described. Also reported is a near-complete membranous obstruction of the esophagus in conjunction with a single TEF at the level of the membrane, communicating with both parts of the esophagus.
The prenatal diagnosis of EA/TEF relies, in principle, on two nonspecific signs: polyhydramnios and an absent or small stomach bubble. Polyhydramnios is associated with a wide range of fetal abnormalities and is nonspecific. Similarly, the ultrasonographic (US) absence of a stomach bubble may point to a variety of fetal anomalies. The combination of a small stomach together with a dilated cervical esophagus (the pouch sign) has been confirmed to be diagnostic for pure EA in a number of patients. Nevertheless, it is encountered in only a few patients. Current US technology does not allow for the certain diagnosis of EA/TEF. Therefore, definitive counseling of the parents should be guarded. Three-dimensional power Doppler imaging has been used both antenatally and postnatally. For example, aortic arch anomalies have been diagnosed using this modality.
Magnetic resonance imaging (MRI) has been used to identify other fetal thoracic lesions, and may be beneficial in patients deemed to be at risk on prenatal US. Sensitivity in various studies is between 60% and 100%, and the diagnosis is made by the lack of visualization of the thoracic esophagus.
If the pregnancy was complicated by polyhydramnios, passage of a tube or catheter into the stomach should be performed to assess esophageal patency. The same holds true when the child presents with anomalies that fit the VACTERL association (e.g., radial aplasia).
As EA prevents the passage of saliva down the esophagus, saliva accumulates in the proximal esophagus and mouth, and feeding should be withheld until esophageal continuity is confirmed. This is best done with a stiff 10 French catheter inserted either through the nose or mouth. A chest film is then obtained with downward pressure on the tube. With EA, the tip of the tube is found to be slightly curled in the blind upper pouch around T2–T4 ( Fig. 27.2 ). This technique not only identifies the atresia but gives some clue about the length of the upper pouch. Often, the dilated upper esophageal pouch is visualized by air within it. Air in the stomach signifies the presence of a distal TEF. If the tip of the catheter passes beyond the level of the carina, then the diagnosis of EA should be questioned. Esophageal stenosis, tracheal rings, and iatrogenic perforation of the esophagus can be confused with EA. If there is any question about the diagnosis, a small amount of contrast can be dripped into the upper pouch, but this needs to be done with fluoroscopy and under direction of the surgeon to ensure that the contrast is not aspirated.
Radiographs may reveal associated anomalies such as vertebral and rib anomalies, or other problems such as duodenal atresia ( Fig. 27.3 ). The absence of air in the stomach points to EA without distal fistula (see Fig. 27.3A ). Mediastinal US has been described as a helpful adjunct in the diagnosis of pure EA.
The length of the esophageal gap is usually not known preoperatively. Absence of air in the stomach has been linked with a long gap, but has also been described in association with a distal fistula occluded with mucus. Even in long-gap EA (atresia without distal fistula) as recently defined by the International Network of Esophageal Atresia (INoEA), the gap length can vary. In newborns with isolated EA, the first procedure is generally a gastrostomy, which allows for enteral feeding and also allows for assessment of the length and location of the lower pouch. It can be identified radiologically, either with metal bougies, a small gastroscope, or with a contrast injected through the gastrostomy. This can be done at the time of the initial gastrostomy or more routinely 7–10 days later. If bougies or a telescope are introduced into the distal esophagus, the amount of pressure on these instruments will affect the measurement of the gap between the two esophageal segments and may under- or overestimate the gap length. In one report by an experienced surgeon, operative management was linked to the measured gap length: fewer than two vertebrae, then primary anastomosis; two to six vertebrae, then delayed primary anastomosis; more than six vertebrae, then esophageal replacement. In the era of thoracoscopy, an initial thoracic exploration can be considered if the upper pouch appears fairly long. If the lower pouch is identified and seen to be of adequate length, a primary repair can be attempted. If not, then a gastrostomy can be placed and delayed repair planned. More recently, van der Zee et al. have advocated for early thoracoscopic exploration without gastrostomy and the use of an internal traction technique to achieve an anastomosis in the first week or two of life. Although this approach appears promising, more study is needed to validate this aggressive early approach.
In EA/TEF, a longer gap between the two esophageal ends should be expected when the distal fistula is found at the carina. Combined with a short upper pouch, this can mean a long gap exists between the two esophageal segments and may not be amenable to an initial primary repair. Unfortunately, this may not fully be appreciated until the time of exploration. However, because it is usually necessary to ligate the fistula in the early postnatal period, the gap length can be assessed at that time. A thoracoscopic approach in this scenario allows for minimal morbidity if the decision is to ligate the fistula only without esophageal reconstruction. Bronchoscopy can also be performed prior to exploration, not only to assess the site of the distal fistula, but to also look for an upper pouch fistula. As previously mentioned, a proximal fistula has been found with a much higher incidence in cases of pure EA than previously reported, in one series up to 50%.
Echocardiography should be performed prior to operation, as it may reveal cardiac and/or aortic arch anomalies. A right descending aorta, which occurs in about 3% of the cases, may make a left-sided thoracic approach preferable. Renal US and spine radiographs should be obtained as well. Because EA may be part of a syndrome, consultation by a geneticist is recommended at some point.
There is little doubt that better preoperative imaging of the neck and chest allows for better preoperative planning. At present, the two best imaging modalities are computed tomography (CT) and MRI. Although MRI would certainly be preferable due to its absence of radiation exposure, it requires general anesthesia. However, MRI is better for diagnosing cardiac and aortic arch anomalies. CT has also been performed in children with EA/TEF, but in truth both these advanced studies add little to the routine management of this congenital anomaly.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here