pathology, and clinical diagnosis of these anomalies includ-

ing a description of associated anomalies such as spina

bifida, horseshoe kidney, and imperforate anus. E. D. Plass,
an instructor in obstetrics at Johns Hopkins University, sur-
veyed the literature in 1919 and reported 136 verifiable cases
of EA including 92 with an associated TEF.6 By 1931,
Rosenthal had collected data on 255 patients7 and indicated,
as Ladd emphasized in 1944,8 “that atresia of the esophagus
is a much more frequent anomaly than it has usually been
considered to be.”
The history of the surgical treatment of EA-TEF is remark-
able and covers 270 years linking the first description and the
first survivor. In 1869 Timothy Holmes of London, the author
of Surgical Management of Children’s Diseases, was the first to
suggest the possibility of a surgical esophageal anastomosis
in infants who had EA without TEF; however, he added that
“the attempt ought not, I think, be made.”9 The first operative
attempt to correct EA without TEF was undertaken in 1888
by Charles Steele of London, and his report titled “Case of
Deficient Oesophagus” was published in the journal Lancet.10
Using chloroform anesthesia, Steele performed a gastrotomy
and attempted to perforate through what he suspected was
an esophageal membrane by pushing a long, slender steel
CHAPTER 69 probe through the gastrotomy site and up the lower esopha-
geal pouch into the upper pouch, which was simultaneously
pushed down with a bougie.10 The attempt was unsuccessful,
and at autopsy, the esophagus was found “to terminate above
Congenital and below in blind rounded ends an inch and a half apart and
there was no cord or connection between the parts.”10 In 1899
Hoffman performed the first permanent gastrostomy in an
Anomalies of the infant with EA.11 In 1913 H. M. Richter of Chicago described
two infants with EA-TEF, on whom he operated without

Esophagus success. Despite his results, Richter was optimistic about

the eventual success of primary esophageal repair and stated,
“I do not wish to dismiss the idea of immediate union of the
Carroll M. Harmon and Arnold G. Coran two ends of the esophagus.”12 In the same year, however, J.
Brennemann, a prominent pediatrician working in Chicago
with Richter, referred to the fatal outcome of the two infants
who underwent attempted surgical repair by stating that
“the physician who, after making his diagnosis of congenital
atresia of the esophagus, decides to let his little patient die
Historical Background
------------------------------------------------------------------------------------------------------------------------------------------------
undisturbed can amply justify his course.”13 Despite Richter’s
optimism, primary esophageal anastomosis and a myriad of
The documented history of esophageal atresia (EA) began in other operations continued to fail; the mortality rate for EA
the year 1670 with William Durston’s published description remained 100%. In a 1925 textbook of thoracic surgery,
of “A Narrative of a Monstrous Birth in Plymouth . . .” in which Lilienthal proposed that operative treatment might include
he described a blind-ending upper esophageal pouch in the “division of the tracheoesophageal fistula and ‘anastomosis’
right infant of a set of female thoracopagus-conjoined twins.1 of the atretic esophagus by tying each pouch over a rubber
The first description of EA with the typical form of tracheoe- tube stent.”14 In 1936 Mims Gage and Alton Oschner of
sophageal fistula (EA-TEF) appeared in 1697 in the fifth edi- New Orleans, Louisiana, proposed early transabdominal liga-
tion of Thomas Gibson’s The Anatomy of Humane Bodies tion of the cardiac end of the esophagus and gastrostomy
Epitomized.2 Isolated case reports of EA-TEF surfaced in the with early secondary cervical esophagostomy.15 Like Richter,
nineteenth century including one by Thomas Hill in 1840. they noted that “the ideal operation would be the separation
Published in the Boston Medical and Surgical Journal, this report of the esophagus from the trachea and an end-to-end anasto-
described an infant with EA-TEF and associated rectal agene- mosis of the upper to the lower segment.”15 They added that
sis with a rectourinary fistula.3 In 1861 Harald Hirschsprung “such an extensive intrathoracic procedure is not justified in a
of Copenhagen reported 4 cases of his own and collected newborn infant, however, and the operation would always be
10 cases of EA and distal TEF from the literature.4 In 1880 finished as a postmortem procedure.”
Morell Mackenzie,5 another prominent medical figure of the The first published report of a case of EA-TEF treated
nineteenth century, reported 57 cases of congenital esophageal by fistula ligation and primary esophageal anastomosis was
malformation with 37 examples of tracheal or bronchial presented by Robert Shaw of Dallas, Texas. Shaw performed
esophageal fistula. He also discussed at length the embryology, the operation on September 25, 1938, and reported it in
893
894 PART VI THORAX

December 1939.16 In his report, Shaw referred to a personal

communication from Paul C. Samson, who had also per-
formed “an operation very similar” to the one Shaw described.
Samson’s patient died 12 hours after the operation. Shaw’s
patient died on the twelfth day after the operation, apparently
as the result of a transfusion reaction. Shaw performed the
operation without knowing that Thomas Lanman in Boston
had performed four primary esophageal repairs for EA-TEF
between January 2, 1936, and July 27, 1937, with the use
of an extrapleural approach.17 In November 1940, Lanman
reported these 4 cases, 1 other case of primary esophageal
repair, and 27 other cases of EA seen at Boston Children’s
Hospital between September 1929 and February 1940. All
32 patients in the series died. However, Lanman’s fourth
patient treated by fistula ligation and primary esophageal
repair lived 9 days and at autopsy was found to have died
of overhydration, not of mediastinitis or pneumonia. Remark-
ably, Lanman summarized: “In spite of the fatal outcome in all
the 30 operative cases, it is felt that considerable progress
along rational lines is being made. The successful operative
treatment of a patient with this anomaly is only a question
of time.”
According to Humphreys and Ferrer,18 the first survivor of
congenital atresia of the esophagus without TEF was a boy
born in New York on February 16, 1935. The patient was
initially treated by gastrostomy alone, with his “first thoracic
operation performed in 1946.”18 The first survivors of
EA-TEF were a boy born in Minnesota on November 26,
1939, and a girl born in Massachusetts on the next day. After FIGURE 69-1 Cameron Haight, MD, performed the first successful
many failures in caring for infants with EA, N. Logan Leven of primary repair of esophageal atresia with a tracheoesophageal fistula on
the University of Minnesota performed a gastrostomy on a March 15, 1941.
2500-g male infant with EA-TEF on November 29, 1939.
After a failed attempt to close the TEF bronchoscopically
with a coagulation electrode, extrapleural ligation of the fistula dilation. Drs. Haight and Towsley presented this case in
was accomplished on January 5, 1940. When the infant’s February 1942 at the Central Surgical Association meeting
weight increased to 4630 g, a cervical esophagostomy was in Chicago and published the report in 1943.19 In 1943
performed on March 27, 1940. The child thrived, and Leven Haight revised his procedure to a right extrapleural approach
proposed that a subsequent antethoracic esophagoplasty be because he believed that better exposure of the distal segment
performed to reestablish continuity of the gastrointestinal was obtained from this side. He also moved to a modified
tract. In Boston, William Ladd performed a gastrostomy on two-layer, “telescoping” anastomosis in the hope of decreas-
a female infant on November 28, 1939; on March 15, 1940, ing the risk for leak. Between 1939 and 1969, Dr. Haight
he closed the TEF and created a cervical esophagostomy. cared for more than 284 infants with EA and reported a
Although both these patients survived, esophagogastric conti- 52% overall survival rate.20 After Haight’s first success in
nuity was achieved only after multiple operations to create an 1941, reports of survival after direct esophageal anastomosis
anterior skin-lined thoracic neoesophagus. were sporadic; however, many centers soon began reporting
In 1935 Cameron Haight began caring for infants with EA series of successes. One interesting report by Longmire in
at the University of Michigan (Fig. 69-1). His first patient was 1947 describes four consecutive successful primary esopha-
managed unsuccessfully by a gastrostomy alone. In 1939 geal anastomoses including success in a baby who weighed
Haight attempted his first primary repair of EA, which was only 1.4 kg.21
unsuccessful. Four subsequent attempts at primary repair also The history of attempts to classify EA reflects differences
failed; however, on March 15, 1941, the first successful in the terminology, but not in the types of anomalies encoun-
primary repair of EA was performed. The infant girl was an tered. In 1929 E. C. Vogt, a radiologist, recognized and classi-
“unusually robust” 12-day-old child who weighed 8 lb, 4 oz fied the types of anomalies.22 With the success of operative
and had been transferred from Marquette, Mich., to the care repair, many other anatomic classifications were proposed.
of the Chief of Pediatrics, Harry A. Towsley. Dr. Towsley, in In 1944 Ladd introduced a numeric form of classification that
turn, consulted Dr. Haight. consisted of five types with Roman numerals.8 Gross altered
The first successful primary repair of EA-TEF was accom- the numeric system in 1953 to an alphabetic system that is still
plished by using a left extrapleural approach with fistula frequently used (Fig. 69-2).23 In 1962 Swenson returned to a
ligation and a single-layered esophageal anastomosis. An numeric classification and used Arabic numbers instead of Ro-
anastomotic leak developed on postoperative day 6 but man numerals.24 In a 1976 report, Kluth published a com-
was managed without surgical intervention. A stricture later plete listing of all described variations of EA including 10
developed at the anastomosis and responded to a single separate classes and additional subclasses.25
 CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 895

A B C D E F
FIGURE 69-2 Gross classification of anatomic patterns of esophageal atresia. A, Esophageal atresia without a tracheoesophageal fistula. This malforma-
tion is almost invariably associated with a “long gap.” B, Atresia with a proximal tracheoesophageal fistula. It is an uncommon anomaly; the abdomen is
airless, and the diagnosis may be missed unless contrast studies or endoscopy is performed. C, Esophageal atresia with a distal tracheoesophageal fistula,
the most frequently encountered form of esophageal anomaly. D, Atresia with a double (proximal and distal) fistula. Though rare, this form is found more
often than originally thought. E, Tracheoesophageal fistula without atresia (H-type fistula). This anomaly may be missed in the newborn period because
swallowing is possible. It is associated with recurrent cough, pneumonia, and abdominal distention. F, Esophageal stenosis. (From Gross RE: Surgery of
Infancy and Childhood. Philadelphia, WB Saunders, 1953.)

TABLE 69-1 of the foregut was the result of fusion of invaginating lateral
Waterston Risk Groups and Current Survival Figures longitudinal ridges that created a septum dividing the foregut
into a dorsal digestive tract and a ventral respiratory system.
Group Survival (%) Waterston Classification
Formation of the so-called tracheoesophageal septum is
A 100 Birth weight > 2500 g and otherwise healthy believed to begin caudally and end cranially.7 Many of the
B 85 Birth weight 2000-2500 g and well or higher theories concerning the pathogenesis of EA malformations
weight with moderate associated anomalies have been based on this description.7,28–31 However, several
(noncardiac anomalies plus patent ductus
arteriosus, ventricular septal defect, and investigators have more recently found little evidence to
atrial septal defect) support this theory of normal foregut separation and develop-
C 65 Birth weight < 2000 g or higher with severe ment. Reinvestigation of human specimens from the Col-
associated cardiac anomalies lection of Embryos at the Carnegie Institute suggests that
ingrowth of lateral foregut wall ridges and cranial develop-
ment of a tracheoesophageal septum do not occur in the
In addition to anatomic classifications, in 1962 D. J. Waterston, human embryo.32,33 In addition, Kluth and colleagues34 stud-
R. E. Bonham Carter, and Eoin Aberdeen of the Hospital for ied staged chicken embryos via scanning electron microscopy
Sick Children in London developed a classification system rela- in an attempt to clarify normal foregut development. In this
ted to risk factors in infants with EA.26 This risk stratification model they also found no evidence of lateral ridges or an
allowed comparison of case outcomes over time and between epithelial tracheoesophageal septum.
hospitals. The use of this classification scheme—which was A summary of the major theories to explain the embryo-
based on birth weight, the presence of pneumonia, and associ- pathology of the foregut has recently been presented by Kluth
ated congenital anomalies—allowed the identification of factors and Fiegel.35 Theories of foregut occlusion and failure of
that predicted survival and guided operative treatment. This recanalization of the intestinal lumen have been histori-
approach to the classification of EA anomalies has been an cally prominent in discussions regarding the pathogenesis
important contribution to the care of these infants (Table 69-1). of intestinal atresias including EA.36,37
Overwhelming proliferation of esophageal epithelium may
play a role in the development of rare membranous EA, but it
Embryology falls short as a comprehensive theory of typical EA malforma-
------------------------------------------------------------------------------------------------------------------------------------------------
tions.25 Another theory has been put forth by Cozzi and col-
The pathogenesis of congenital EA malformations remains leagues38,39 and others40 in which it is suggested that EA is a
unknown. Although many theories have been offered to component of cephalic neurocristopathy. The observation that
explain the etiology of EA, EA-TEF, and TEF, no single unifying there is a clear association of neural crest-implicated cardio-
theory has been proposed that addresses all the variations seen vascular anomalies (aortic arch, conotruncal and membranous
with this group of anomalies. The pathogenesis is heteroge- ventricular septal defects), as well as thymic, thyroid, parathy-
neous and multifactorial and involves multiple genes and roid, and facial malformations with EA, as seen with the
complex gene-environment interactions. In addition, about DiGeorge syndrome, suggests that the pathogenesis of EA
half of the cases of EA are isolated and the other half are may be related to defective pharyngeal arch development.
associated with other malformations. An animal model in which Adriamycin is used as a terato-
A clear understanding of the pathogenesis of the EA gen and results in embryos with a variety of malformations,
malformation is hindered by the fact that normal foregut including EA and TEF, has been described and promises to
embryology is still obscure. Wilhelm His, Sr., the founder provide new insight into EA pathogenesis.41,42 Merei and
of human embryology, was the first to describe normal devel- Hutson have used this model to develop a theory suggesting
opment of the respiratory system.27 He believed that division that the primitive foregut develops into the trachea, with
896 PART VI THORAX

subsequent EA-TEF anomalies.43 Gittes has recently used this in the United States55 and Australia.56 According to data from
model, as well as human TEF tissue samples, to investigate the the California Birth Defects Monitoring Program database, from
molecular mechanisms that may play a role in the patho- 1983 to 1988 the total prevalence of EA, EA-TEF, and TEF alone
genesis of EA-TEF. He has proposed that the distal fistula tract was reported to be 2.82 per 10,000 live births and stillbirths.57
and esophagus are of respiratory origin and that defects in In Europe, the prevalence rate of EA-TEF during 1980 to 1988
lung morphogenesis account for this aspect of the EA-TEF was 2.86 per 10,000 births, although this rate appeared to be
anomaly.44,45 Insight into the pathogenesis of EA-TEF will also decreasing over time.58 In this European study, 62% of infants
come from a clearer understanding of the genetics of this and with EA-TEF were male, whereas the California database found
associated conditions. More than 50% of cases of EA-TEF are that male-to-female ratios varied considerably between types
associated with other anomalies, and up to 10% of cases are of EA-TEF defects, with ratios of 2.29 for TEF alone, 1.44
found in specific chromosomal or single-gene disorders. for EA-TEF, and 1.33 for EA alone.57 Mothers of white ethnicity
Chromosomal abnormalities associated with EA-TEF include have a higher (>60%) prevalence of EA-TEF and TEF than
DiGeorge syndrome; full trisomies (21, 13, and 18); Opitz nonwhite populations do.59,60 An increased risk for EA-TEF
syndrome; and 13q, 17q, and 16q24 deletions.46 Single-gene with a first pregnancy,59 as well as increasing maternal age,57
mutation examples include MYCN (Feingold syndrome), have been reported (twofold increased risk for women 35
CHD7 (CHARGE syndrome), SOX2 (AEG syndrome), GL13 to 40 years old and threefold increase for those older than
(Pallister-Hall syndrome), MID1 (Opitz G syndrome), and age 40).60
FANCA (Fanconia anemia).46 The rate of multiple births is high for each type of EA-TEF
Recent insights into the genetics involved in VACTERL asso- anomaly (TEF, 3.7%; EA-TEF, 4.9%; and EA alone, 8.8% in
ciation will also provide new information into the etiology of a control population).57 In addition, EA, along with other con-
EA-TEF. For example, the VACTERL-hydrocephalus association genital anomalies, is significantly increased in the offspring of
is related to mutations in the following genes: FANCC, in vitro fertilization patients (OR 3.65:CI 2/53-5/26).61
FANCD1 and 2, FACCG, FANCB, and PTEN. In addition to
VACTERL associations, EA-TEF is also increased in several
other conditions such as Goldenhar syndrome and Martinez- Associated Anomalies
Frias syndrome.47 The development of rat and mouse models ------------------------------------------------------------------------------------------------------------------------------------------------

of EA has recently allowed the parallel study of both normal The early disturbance in organogenesis that results in EA
and abnormal embryogenesis. As discussed earlier, contro- deformities, whatever the exact cause, also affects other organ
versies persist, but the fundamental morphogenetic process systems. Other congenital anomalies are frequently associated
appears to be a rearrangement of the proximal foregut into sep- with EA, and the associated anomaly often significantly alters
arate respiratory (ventral) and gastrointestinal (dorsal) tubes. treatment and affects survival. Numerous reports have sug-
This process depends on the precise temporal and spatial gested that between 50% and 70% of infants with EA have
pattern of expression of a number of foregut patterning at least one associated congenital malformation. The anoma-
genes. Disturbance of this pattern disrupts foregut separa- lies are most common in cases of EA without TEF and are
tion and underlies the development of tracheoesophageal least common in cases of H-type TEF.62 About half of all
malformations.48 Specific murine models of EA-TEF have sug- EA-TEF-associated malformations can be classified as part
gested a number of important genes related to pathogenesis of a recognizable malformation syndrome such as chromo-
that are mostly in developmental pathways including vitamin somal, VACTERL, CHARGE, Fanconi anemia, Opitz G, and
A effectors, retinoic acid receptors a and b (Rara and Rarb), Goldenhar. The other 50% of patients with associated malfor-
sonic hedgehog pathway effectors (Shh, Gli2, Gli3, Foxf1), mations are considered as “nonsyndromic” with multiple
and other homeobox containing transcription factors (Hoxc4, congenital anomalies. In a recent study from France, these
Ttf-1, Pcsk5).46 As suggested by the Adriamycin rat model malformations included cardiovascular system (24%), uro-
for EA-TEF, environmental teratogens have also been implicated genital (21%), digestive (21%), musculoskeletal (14%), and
in the pathogenesis of EA ! TEF. EA has occurred in infants central nervous system (7%).63
born to mothers with prolonged exposure to contraceptive Considering reports over recent decades, cardiovascular
pills49 and exposure to progesterone and estrogen during a anomalies occur most frequently (11% to 49%),39,64–66 followed
pregnancy.50 In addition, EA has been reported in some infants by genitourinary (24%),67 gastrointestinal (24%),68 and muscu-
of hyperthyroid51 and diabetic mothers and after intrauterine loskeletal (13% to 22%)69,70 anomalies (Table 69-2). Associated
exposure to thalidomide and diethylstibestrol.46,52 neurologic anomalies include neural tube defects (2.3%), hydro-
Many studies have described transverse and vertical familial cephalus (5.2%), holoprosencephaly (2.3%), and anophthalmia
cases of all varieties of EA. Empirical risk figures based on the or microphthalmos (3.7%).59 Other reported associated anom-
literature to date give a 0.5% to 2% recurrent risk in siblings of alies include choanal atresia (5.2%), facial cleft (7.2%), abdom-
one affected child; the risk increases to 20% if more than one inal wall defects (4.3%), and diaphragmatic hernia (2.9%).59
sibling is affected.53 The empiric risk for a child born to an One report suggests that the incidence of certain associated
affected parent is 3% to 4%.53 anomalies may vary by world region in that genitourinary anom-
alies were found in 26% of a European cohort of patients versus
only 4% of an Asian cohort.70
Epidemiology
------------------------------------------------------------------------------------------------------------------------------------------------
Complex cardiac deformities may account for most of
the deaths associated with EA malformations. The relative risk
The reported incidence of EA, with or without TEF (EA ! for death in infants with EA malformations associated with a
TEF), is 1:3500 live-born infants; however, it varies geograph- major cardiac anomaly is reported to be 3.47 (95% confidence
ically from 1 in 2440 births in Finland54 to 1 in 4500 births interval [CI], 1.51 to 7.96).65,71 In a series of 153 patients

TABLE 69-2
Incidence of Associated Anomalies
Cardiovascular
Genitourinary
Gastrointestinal
Musculoskeletal
Central nervous system
VACTERL association
Overall incidence
VACTERL, vertebral, anorectal, cardiac, tracheoesophageal, renal, and limb
abnormalities.
with EA, Leonard and colleagues reported that the 1-year
survival rate of infants with congenital heart disease (CHD)
was 67% versus 95% in those without CHD.66 Interestingly,
these authors concluded that although CHD is associated with
higher mortality in EA patients, it is not the direct cause of it.66
The most common single defects identified are ventricular
septal defects (19%), which have reportedly been associated
with a mortality rate of 16%72 and atrial septal defects
(20%).64 Other common cardiac anomalies include the tetral-
ogy of Fallot (5%) and patent ductus arteriosus (13%).64
Coarctation of the aorta is found in 1% to 4% of infants with
EA malformations.64,73 A descending aorta on the right side
has been reported in 4% of infants with EA.64
Some gastrointestinal anomalies associated with EA ! TEF
are anorectal malformations (14%),64 duodenal atresia (2%),64
intestinal malrotation (4%),74 ileal atresia, annular pancreas,
and pyloric stenosis. Genitourinary defects are varied and
include renal agenesis or hypoplasia (1%), hypospadias,
undescended testes, cystic renal disease, hydronephrosis, vesi-
coureteral reflux, ureteric duplication, ureteropelvic and
vesicoureteral obstruction, urachal anomalies, ambiguous
genitalia, and cloacal or bladder exstrophy.59,75,76 Typical
musculoskeletal malformations include limb (15%) and
vertebral (17%) anomalies.64
In 1973 Quan and Smith,77 suggested a broad spectrum
of associated malformations that are associated with EA
malformations. They arranged this association into the acro-
nym VATER: vertebral defects, anal atresia, tracheoesophageal
fistula, esophageal atresia, and renal defects. Opitz defined
the term “association” as the idiopathic occurrence of multiple
congenital anomalies during blastogenesis.47,77 Affected
patients have no family history of malformations, no recogniz-
able teratogen is involved, and no chromosomal abnormality
is observed. As the phenotype expanded, the acronym was
changed to the VACTERL association, which includes verte-
bral, anorectal, cardiac, tracheoesophageal, renal, and limb
abnormalities.78 The incidence of the VACTERL association
in the EA population is approximately 20%.65 Infants specif-
ically with EA-TEF frequently have VACTERL anomalies
(vertebral, 17%; anal, 12%; cardiac, 20%; renal, 16%; limb,
10%), as well as other midline defects (cleft lip and palate,
2%; sacral dysgenesis, 2%; urogenital anomalies, 5%).79
The “renal” category should include urinary tract anomalies
such as megalourethra, urethral duplication, urethral valves,
stricture, and hypospadias. It generally requires a minimum
of three of the components to qualify as a “VACTERL patient.”
Recently, the VACTERL-H association has been described and
includes congenital hydrocephalus.77,80 Infants with EA mal-
formations in association with the VACTERL association have
CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 897
a high mortality rate. Iuchtman and colleagues81 analyzed the
results of 46 infants born with EA and features of the VAC-
"24%
TERL association among 313 patients with EA, TEF, or both.
"21%
In this series, the mortality rate was 24%. Cardiovascular
"21%
anomalies were present in 78% of the patients and were the
"14%
principal cause of death. More recently, Driver and colleagues
"7%
reported that the relative risk for mortality in patients with
"20%
VACTERL associations was 2.54 (95% CI, 1.14 to 4.86).65
50%-70%
EA malformations are increasingly being described in asso-
ciation with other polytopic defects, sequences, syndromes,
and associations. Some of these include Down syndrome82;
the DiGeorge sequence83; the polysplenia sequence; Holt-Oram
syndrome; the Pierre Robin sequence53; Feingold syndrome84;
Fanconi syndrome; Townes-Brock syndrome; Bartsocas-Papas
syndrome; McKusick-Kaufman syndrome53,85–88; the CHARGE
association (coloboma, heart defects, atresia choanae, develop-
mental retardation, genital hypoplasia, and ear deformities)89;
and the schisis association (omphalocele, neural tube defects,
cleft lip and palate, and genital hypoplasia).90,91 Other serious
or lethal associations are trisomy 18, cerebral hypoplasia, and
Potter syndrome (bilateral renal agenesis).92,93 In addition to
the well-described anomalies associated with EA just discussed,
many recent reports describe significant malformations and
associated conditions that may become more frequently recog-
nized. Unilateral pulmonary agenesis associated with EA-TEF
is a rare condition; only 37 cases have been reported since
1874.94–100 Usui and colleagues101 reported a 47% incidence
of tracheobronchial abnormalities such as ectopic or absent
right upper lobe bronchus, congenital bronchial stenosis, and
a decreased ratio of circumferential cartilaginous trachea to
membranous trachea; these abnormalities could play an impor-
tant role in such respiratory conditions as tracheomalacia
and atelectasis associated with EA.101 Several reports also sug-
gest that EA is associated with abnormalities in vagal innerva-
tion that are not related to surgical intervention to repair the
esophageal defect.102
Classification
------------------------------------------------------------------------------------------------------------------------------------------------
As noted previously, numerous classification schemes have
been proposed and used to describe EA. The most useful
and practical classifications are perhaps simple anatomic
descriptions, even though many institutions and surgeons still
use the Gross classification. It is important to note, however,
that there are many variations to the fairly straightforward
anatomic descriptions given in Table 69-3. In 1976 Kluth35
summarized and diagrammed many of the anatomic variants
not described by the many classifications mentioned here.
Waterston’s 1962 classification was based on risk factors
(see Table 69-1). The classification affected the surgical care
of these patients. Infants in the “good”-risk category (A) were
TABLE 69-3
Incidence of Various Esophageal Atresia—Tracheoesophageal
Fistula Anomalies
1. Esophageal atresia with a distal tracheoesophageal fistula 84%
2. Esophageal atresia without tracheoesophageal fistula 6%
3. Tracheoesophageal fistula without esophageal atresia 4%
4. Esophageal atresia with a fistula to both pouches 1%
5. Esophageal atresia with a proximal tracheoesophageal fistula 5%
898 PART VI THORAX

typically treated with immediate operative repair, “moderate”- adjunct in the prenatal diagnosis of EA anomalies suspected
risk infants (B) were managed by delayed repair, and “high”- on ultrasound.111
risk infants (C) were treated by staged repair. Although the Most infants with EA are symptomatic in the first few hours
Waterston classification continues to be used to compare results of life. The earliest clinical sign of EA is usually excessive
between centers, many investigators have questioned its current salivation that results from pooling of secretions in the phar-
validity in caring for these infants.68,92,103 With modern neona- ynx. Typically, the first feeding is followed by regurgitation,
tal critical care, more low-birth-weight infants are surviving and choking, and coughing. Other features are cyanosis with or
more treatment options are available for infants with multiple without feeding, respiratory distress, inability to swallow,
congenital anomalies. As a result, a search for modern criteria and inability to pass a feeding or suction catheter through
for survival has produced several new classification schemes. the mouth or nose into the stomach. If a distal fistula is
In 1989 Randolph and colleagues103 suggested a refine- present, the abdomen distends as inspired air passes through
ment in Waterston’s classification on the basis of the overall the fistula into the stomach. Pulmonary compromise can
physiologic status of an infant with EA. According to a report become significant as gastric fluid refluxes through the TEF,
analyzing the results of 39 infants managed with the physio- spills into the trachea and lungs, and subsequently leads
logic status criteria, more and earlier primary repairs were to chemical pneumonitis. As the abdomen distends with
performed, and excellent survival rates were maintained in air, the diaphragms (the principal muscles of respiration in
stable infants in comparison with 87 infants previously the newborn) elevate and pulmonary status worsens. Aspira-
managed by using Waterston’s criteria.103 tion of saliva from the upper pouch into the trachea further
Other prognostic classifications have recently been pro- exacerbates the pulmonary compromise.
posed. Poenaru and colleagues92 suggested that only severe The diagnosis of EA can be confirmed by passing a firm
pulmonary dysfunction with a preoperative mechanical venti- catheter through the mouth into the esophagus to the point
lation requirement and severe associated anomalies were inde- at which resistance is met. A few milliliters of air can be
pendent predictors of survival. Brown and Tam104 suggested injected through the tube and used as a contrast agent to dis-
that the measured length of the esophageal gap provides a tend the upper esophageal pouch as frontal and lateral films
method of classification to predict morbidity and long-term are obtained (Fig. 69-3). A flexible catheter such as a feeding
outcome. In a review of 357 cases of EA treated at the Hospital tube will often curl up in the upper pouch, also suggesting the
for Sick Children on Great Ormond Street in London between diagnosis of EA (Fig. 69-4). If necessary, 0.5 to 1.0 mL of
1980 and 1992, Spitz and colleagues93 found that birth diluted barium can be used as a contrast agent and injected
weight and major cardiac disease were important predictors into the upper pouch to confirm the diagnosis (Fig. 69-5).
of survival. The Spitz classification is currently the system most Under carefully controlled fluorography, barium may also
commonly used.65,105 Okamoto and colleagues196 recently detect a proximal TEF; in the usual circumstance of a portable
suggested a revision to the Spitz classification (Table 69-4). film performed in the neonatal unit, however, barium identi-
Using receiver-operating characteristic (ROC) curves to com- fied in the tracheobronchial tree is more likely to represent
pare the two classifications, these authors suggest that as overall contrast aspirated through the larynx rather than through a
care of low- and very-low-birth-weight infants improves, the proximal TEF. A very small or short upper blind pouch may
cardiac anomalies play an increasing role in assessing prognosis also suggest the presence of a proximal TEF. Air in the stomach
for infants with EA. and bowel confirms the presence of a distal TEF. Absence of air
in the abdomen typically represents isolated EA without distal
TEF (Fig. 69-6), but the presence or absence of a proximal
Diagnosis and Clinical Findings TEF must be confirmed by bronchoscopy. The incidence of
------------------------------------------------------------------------------------------------------------------------------------------------
proximal TEF is much higher than thought in the past. How-
Unlike many congenital pediatric surgical problems, prenatal ever, it is possible to have absence of intestinal air in the
EA is not commonly diagnosed. Prenatal detection of EA by circumstance of a narrow or occluded distal fistula, which is
ultrasonography relies on the finding of a small or absent identified at the time of operation.112,113 The diagnosis of
stomach bubble and associated maternal polyhydramnios.106 TEF without EA is more difficult and requires a high index
However, with these criteria, the predictive value of prenatal of suspicion on the basis of clinical symptoms. The diagnosis
ultrasonography in the diagnosis of EA is only 20% to is often delayed, and investigations are triggered by repeated
40%.107,108 The ultrasonographic finding of an esophageal coughing or choking during feedings, often with pulmonary
pouch in the middle of the fetal neck in association with poly- infiltrates on chest radiograph suggesting aspiration. The diag-
hydramnios and a small stomach may increase the accuracy of nosis can be made by barium esophagography in the prone
prenatal diagnosis of EA.109,110 A recent report suggests that position (Fig. 69-7); however, bronchoscopy with or without
real-time fetal magnetic resonance imaging may be a useful esophagoscopy is often required to confirm the diagnosis.

TABLE 69-4
Okamoto modification of the Spitz Classification: Predictors of Survival in Cases of Esophageal Atresia
Class Description Risk Survival
Class I No major cardiac anomaly, BW # 2000 g Low 100%
Class II No major cardiac anomaly, BW < 2000 g Moderate 81%
Class III Major cardiac anomaly, BW # 2000 g Relatively high 72%
Class IV Major cardiac anomaly, BW < 2000 g High 27%

FIGURE 69-3 Upper esophageal atresia pouch with air. Also note verte-
bral anomalies.
FIGURE 69-4 Upper esophageal atresia pouch with coiled feeding tube.
As previously discussed, the incidence of recognizable con-
genital defects associated with EA is between 50% and 70%.
The clinical evidence of these other anomalies should, there-
fore, be considered in the diagnostic evaluation for EA. In
addition to a physical examination focused on seeking known
associated defects (such as those of the VACTERL association),
CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 899
additional testing usually includes echocardiography, renal
ultrasonography, and chromosomal analysis. In fact, it is not
unusual for the finding of an anorectal malformation, for
example, to precede the clinical signs and symptoms of a
concomitant EA.
Preoperative Treatment
------------------------------------------------------------------------------------------------------------------------------------------------
Pneumonitis from aspiration of upper pouch secretions and
reflux of gastric acid through the TEF causing worsening
respiratory distress is the most critical preoperative problem
for an infant with EA-TEF. Immediate management includes
measures to prevent further aspiration and treatment of pneu-
monitis. A sump catheter should be positioned in the upper
esophageal pouch to continuously aspirate saliva under
low-pressure suction. The double-lumen Replogle type of
catheter is best for this purpose because the perforations along
the side of the catheter are located only near the tip, which
minimizes the possibility of suctioning oxygenated air away
from the larynx. The infant should be positioned to minimize
reflux of gastric fluid up through the TEF. An upright sitting
position has traditionally been advocated, but some authors
argue that the head-up, prone position is most effective at
minimizing reflux. Broad-spectrum antibiotic coverage and
pulmonary physiotherapy are also initiated. Intravenous fluid
therapy with 10% dextrose and hypotonic saline should
be started to maintain fluid, electrolyte, and glucose bal-
ance. Vitamin K analogue should also be administered before
surgery. Routine endotracheal intubation should be avoided
because of the risk for gastric perforation and worsening res-
piratory distress as the abdomen becomes distended from
ventilation through the TEF.
Operative Repair
------------------------------------------------------------------------------------------------------------------------------------------------
The operative approach to an infant with EA depends greatly
on the specific type of anomaly present and the occurrence of
associated anomalies.
ESOPHAGEAL ATRESIA WITH DISTAL
TRACHEOESOPHAGEAL FISTULA
Emergent operation for EA with distal TEF is seldom necessary,
and a period of 24 to 48 hours between diagnosis and operation
permits full assessment of the infant and treatment of pul-
monary insufficiency including atelectasis and pneumonitis.
In most infants, open thoracotomy or thoracoscopic division
of the fistula with primary anastomosis of the esophagus is
possible and is the operative procedure of choice.
For the open approach, the infant is usually positioned for
a standard right posterolateral thoracotomy, with the right arm
extended above the head and the head slightly flexed. If a
right-sided aortic arch (2.5%) is identified on a preoperative
echocardiogram, a left-sided thoracotomy is preferred; how-
ever, a double aortic arch is not unusual and makes a left
thoracic approach difficult as well.114 A curved skin incision
is made around the lower border of the scapula and extends
from the anterior axillary line posteriorly to the paravertebral
region (Fig. 69-8). Subcutaneous tissue and chest wall mus-
cles are divided using electrocautery; this allows elevation of
the scapula and identification of the fourth intercostal space.
900 PART VI THORAX

A B
FIGURE 69-5 The upper esophageal pouch is outlined by 0.5 to 1 mL of thin barium in the frontal (A) and lateral (B) projections.

FIGURE 69-7 An H-type tracheoesophageal fistula is shown as contrast

FIGURE 69-6 Absence of intestinal air suggests isolated esophageal
atresia without tracheoesophageal fistula. Also note vertebral anomalies.
Some surgeons advocate a muscle-sparing thoracotomy using
an approach between the latissimus dorsi and the serratus
anterior muscles in an attempt to minimize postoperative
shoulder function morbidity.115,116 The thorax is entered
through the fourth intercostal space by dividing the intercostal
muscles, with care taken to avoid incising the pleura. Most
is injected through a nasoesophageal tube. Contrast is noted passing from
the esophagus, through the fistula (arrow), and filling the upper part of the
trachea and larynx.
pediatric surgeons continue to advocate the extrapleural
approach for EA-TEF repair because a substantial anastomotic
leak does not result in empyema but merely causes an esopha-
gocutaneous fistula, which typically closes in 1 to 2 weeks.
Surgeons who use the transpleural route argue that the oper-
ative time is shorter and, with current antibiotic options, the
risk for empyema after a leak is minimal. With the extrapleural
approach, the pleura are gently pushed away from the chest
 CHAPTER 69
FIGURE 69-8 Positioning of an infant for repair of esophageal atresia
with or without a tracheoesophageal fistula. The thorax is entered.
wall; this in turn allows insertion of a rib-spreading retractor.
Moist pledgets, tissue applicators, or gauze can be used to
dissect the pleura anteromedially as the rib spreader is sequen-
tially opened (Fig. 69-9). With further retropleural dissection,
the azygos vein is exposed and may be divided between ties
as the lung and pleura are gently retracted medially. When
the posterior mediastinum is exposed, the upper pouch, distal
TEF, trachea, and vagus nerve are identified. The lower part of
the esophagus is dissected circumferentially at the level of the
fistula, and every effort is made to preserve the vagal fibers that
supply the distal portion of the esophagus. Traction on a heavy
silk suture, tape, or vessel loop passed around the distal end of
the esophagus controls gas flow through the fistula into the
stomach and allows meticulous dissection and exposure for
fistula closure (Fig. 69-10). The latter can be performed with
5-0 or 6-0 permanent or absorbable suture placed in inter-
rupted fashion (Fig. 69-11). Because occasional intratracheal
granuloma has occurred with permanent sutures, we recently
chose to close the fistula with absorbable suture such as poly-
dioxanone (Fig. 69-12). It is important to leave 1 to 2 mm of
esophagus on the tracheal end of the fistula to minimize the
risk for postoperative tracheal stricture, while not leaving an
excessive amount that would act as a tracheal diverticulum.
The air-tightness of the tracheal closure can be tested by filling
the thoracic cavity with warm saline and watching for bubbles
with positive-pressure ventilation.
A small tube should be passed through the distal end of
the esophagus to ensure that the lumen is adequate and
open and to aspirate air distending the stomach. Occasionally,
FIGURE 69-9 Extrapleural dissection is facilitated by using moistened tissue applicators or pledgets to push the parietal pleura away from the chest wall
and subsequently away from the posterior mediastinum.
CONGENITAL ANOMALIES OF THE ESOPHAGUS 901
FIGURE 69-10 Distal esophageal pouch mobilization up to the tracheoe-
sophageal fistula at the tracheal bifurcation. Note the clear view of vagal
branches.
a congenital distal esophageal stricture will be encountered.
Two fine stay sutures on the distal esophageal opening allow
gentle traction for dissection. The extent of distal mobilization
should be minimized to avoid damaging vagal branches and
the segmental blood supply. However, distal mobilization
should be undertaken if it is necessary to ensure a primary
low-tension anastomosis. Identification of the proximal
esophageal pouch can be facilitated if the anesthesiologist
gently pushes on a suction catheter, which is passed through
the mouth into the upper pouch. A traction suture can be
placed through the tip of the pouch, and even through the suc-
tion catheter, to assist in proximal dissection and avoid the
trauma caused by repeatedly applying forceps to the proximal
pouch. Mobilization of the upper pouch should be sufficient
to bring the upper pouch down to the distal esophageal seg-
ment. This usually necessitates dissection up to the thoracic
inlet (Fig. 69-13). Extensive circumferential dissection of
the proximal pouch also allows for identification of an undiag-
nosed proximal TEF. Because the cervical blood supply to the
proximal part of the esophagus is excellent, this type of exten-
sive dissection does not run a significant risk for ischemic
injury. Great care must be taken during the dissection between
902 PART VI THORAX
C ensure distal esophageal patency and the potential for early
FIGURE 69-11 A, The opening of the tracheoesophageal fistula is
closed with 5-0 or 6-0 nonabsorbable suture. B, The feasibility of primary
anastomosis between the two esophageal segments is being asses-
sed. C, A proximal esophagomyotomy being used to gain additional
length. (From Rowe M, O’Neill JA, Grosfeld JL, et al: Essentials of Pediatric
Surgery. St Louis, Mosby-Year Book, 1995.)
the esophagus and the membranous trachea, however, to avoid
inadvertently opening the trachea. Unfortunately, the more
extensive the proximal dissection, the greater the likelihood
of disturbance to vagal branches.
The tip of the upper esophageal pouch is excised or simply
opened and stretched at its lowest point to a diameter corre-
sponding to the distal esophageal lumen. Traction sutures are
placed opposite each other, and an end-to-end anastomosis is
begun by placing interrupted 5-0 or 6-0 sutures in the back
wall with the knots tied on the inside of the lumen; great care
should be taken to include full-thickness esophageal wall
including the mucosa and muscularis (Fig. 69-14). The upper
pouch mucosa often retracts and can easily be missed if the
surgeon is not vigilant. Unless the two ends of the esophagus
are very close, it is best to insert the entire back row of sutures
FIGURE 69-12 Interrupted suture closure of tracheoesophageal fistula.
FIGURE 69-13 Dissection of the upper esophageal pouch.
before tying them. A small feeding tube can then be passed
and advanced across the anastomosis into the stomach to
postoperative enteral feeding if desired. The anterior layer
of the anastomosis is then completed over the tube, with
the knots tied on the outside (Fig. 69-15). The effect of the
type of suture material used in the esophageal anastomosis
on subsequent stricture formation has been a point of debate.
In a review of patients in whom anastomotic strictures devel-
oped, Chittmittrapap and colleagues117 reported a statistically
significant increased rate of stricture when braided silk was
used versus polypropylene or polyglycolic acid suture.
Once the anastomosis is complete, the position of the tube
should be confirmed if it is to remain in place for postopera-
tive gastric decompression and feeding. A chest tube or closed
suction drain is usually placed in the retropleural space and
secured to the lateral chest wall by a loose absorbable suture.
The drain should be placed away from the anastomosis. Some
surgeons leave no chest tube or drain in the chest after a
retropleural dissection.118
In infants with EA and distal TEF, another technique for
handling the fistula and uniting the esophageal ends is the
end-to-side anastomosis developed by Sulamaa and col-
leagues119 and championed by Duhamel and others.109,120,121
With this technique, a single ligature is used to ligate the
 CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 903

fistula; the upper pouch is mobilized down to the lower

esophageal segment, and a wide, oblique, end-to-side anasto-
mosis is performed. Because of concern for recurrent TEF
formation, this method has not been widely accepted. In
a review from a center where both techniques were used,
Poenaru and colleagues identified end-to-side anastomosis
as a significant risk factor for mortality, specifically because
of recurrent TEF, but noted a lower rate of gastroesophageal
reflux (GER).122 A
In another series of 68 infants treated with this technique,
Touloukian and colleagues121 reported a TEF recurrence rate
of 7%, a rate similar to or lower than that seen in large series
of end-to-end reconstructions. In addition, the rate of other
complications such as stricture (5%), anastomotic leak
(8%), and GER (6%) was lower than in most other series using
end-to-end esophageal reconstruction.
Thoracoscopic repair of EA-TEF is being adopted more
widely. This technique typically uses three 2.5- to 5-mm trans-
pleural access trocars, an angled telescope, a video camera,
and small-diameter (2.5 to 5 mm) instruments to identify B
and ligate the TEF (often with a 5-mm clip), mobilize the
proximal esophageal pouch, and perform an end-to-end
esophagoesophagostomy in a fashion that is similar to the
open thoracotomy approach (see Figs. 69-10, 69-12, 69-13,
69-15, and 69-16). Advocates of this approach have suggested
that benefits include superior visualization, improved cosm-
esis, and elimination of the morbidity of neonatal thoracotomy
including scoliosis, winged scapula, chronic pain, shoulder
weakness, chest wall asymmetry, and maldevelopment.123,123a
However, the thoracoscopic EA-TEF repair is technically
demanding and requires advanced endosurgical skills. The
thoracoscopic end-to-end esophageal anastomosis is techni-
cally challenging because of limited working space, lack of
articulation at the end of the instruments (the needle driver C
in particular), and the necessity of tying sutures under
FIGURE 69-14 Single-layer esophageal anastomosis. A, Corner stitches
tension.123 are placed and the knots are tied on the outside. B, The posterior row is
Particular attention must be given to infants with EA and a placed with the knots tied on the inside. C, The anterior row completes
large distal TEF with severe respiratory distress syndrome. the anastomosis over a tube with the knots tied on the outside. (From
Under these circumstances, preoperative endotracheal intuba- Rowe M, O’Neill JA, Grosfeld JL, et al: Essentials of Pediatric Surgery. St
Louis, Mosby-Year Book, 1995.)
tion is often necessary. However, high-pressure ventilation
may worsen lung ventilation and abdominal distention as
inspired air is diverted through the TEF into the stomach, thus
exacerbating respiratory compromise. Affected infants may
require emergency surgical intervention to stabilize the pul-
monary status and avoid gastric perforation, which is typically
fatal in these circumstances.
Several operative approaches have been described to man-
age this problem including gastric division,124 banding of the
gastroesophageal junction,125 and positioning the tip of the
endotracheal tube below the fistulous orifice.126 The latter
technique may be dangerous if not combined with fiberoptic
bronchoscopy to locate the TEF; a fistula at or close to the
carina precludes the safe use of this technique, and inadver-
tent intubation through a large fistula may be fatal.127 Bron-
choscopic placement of a Fogarty balloon catheter through
the fistula has been successfully used to ablate the flow of
ventilated air through the fistula.128 This approach may be
used to stabilize a rapidly deteriorating infant; however, con-
cern has been raised about potential technical difficulties
with this approach and problems with injury to the distal
end of the esophagus in which the balloon is inflated.129 FIGURE 69-15 Completed end-to-end esophageal anastomosis.
Emergency gastrostomy to decompress the air-filled stomach
904 PART VI THORAX

B C

D E
FIGURE 69-16 Thoracoscopic repair of esophageal atresia with a distal tracheoesophageal fistula. A, Position of the trocars, a 6-mm trocar for the
telescope and two 4-mm trocars for instruments. B, The posterior pleura has been incised and the upper pouch is being mobilized. C, The distal fistula
has been dissected free with the azygos vein left intact. D, Fistula being ligated close to the trachea. E, Completed end-to-end esophagoesophagostomy.
(From Bax KMA, Van der Zee DC: Feasibility of thoracoscopic repair of esophageal atresia with distal fistula. J Pediatr Surg 2002;37:192.)

and small bowel has also been used in this situation. This ap-
proach, however, places the infant at risk for pulmonary col-
lapse because each ventilator breath volume bypasses the lung
by preferentially passing through the TEF and escaping via the
low-resistance gastrostomy tube. The use of a gastrostomy
tube with an underwater seal may increase air resistance
and decrease runoff through the fistula. A small balloon cath-
eter passed through the gastrostomy into the lower esophagus
can provide excellent palliation for days to weeks.129a
Emergent thoracotomy and fistula division in an infant
with severe respiratory distress syndrome who requires
high-pressure ventilation is generally the best approach; in a
premature baby, the decision to operate should be made early
because the hyaline membrane disease will worsen in the first
24 to 36 hours and will make the lungs even less compliant. In
many circumstances, the infant remarkably improves after
emergency fistula ligation; this improved condition may allow
primary repair of the esophagus and thus eliminates the need
for a second operation.
LONG-GAP ESOPHAGEAL ATRESIA
Occasionally in infants with EA-TEF, the upper pouch is high
and the distance between the upper and lower esophageal
segments limits the ability to easily complete end-to-end eso-
phagoesophagostomy with acceptable tension. The same is
true for most neonates with isolated EA because the lower
esophagus is not “attached” to the trachea. Although much
has been said and written about the subject, there is no precise
definition of “long-gap” EA. What is amenable to primary
anastomosis by some surgeons may be considered “long
gap,” “very long gap,” or “ultralong gap” by others. In addition,
measurement of gap length can be biased by the method used
to measure the distance between the proximal and distal ends
of the esophagus. Some reports describe preoperative mea-
surement of the gap by inserting a radiopaque tube or bougie
into the upper pouch and placing contrast or a flexible endo-
scope into the distal pouch through a previously placed gas-
trostomy tube.130,131 Other definitions of gap length are based
on intraoperative measurements obtained before or after
mobilization of the proximal pouch and, on occasion, the
proximal and distal pouches. These variations on the method
of determining gap length contribute to the confusion and
debate about how long is long.132
In the case of isolated EA, in which a “long gap” is expected,
most pediatric surgeons would proceed with the operative
placement of a gastrostomy tube, a period of observation,
and an attempted delayed primary repair. It is well documen-
ted that during the first several months of life, the gap between
the two ends of the esophagus tends to lessen because of
spontaneous growth, possibly related in part to reflux of bolus
gastric feedings into the lower esophagus, which makes pri-
mary repair more feasible.133,134 In addition, many preopera-
tive mechanical techniques have been described as a means
to facilitate narrowing of the esophageal gap. The most com-
monly used method is that of upper pouch bougienage, in
which a weighted bougie is passed through the mouth into
the upper pouch, with forward pressure applied daily or twice
daily for 6 to 12 weeks before attempting delayed primary
repair.135 The use of preoperative upper and lower pouch
bougienage to decrease gap length has also been described,136
but some of the early proponents have abandoned this tech-
nique.136a In 1975 Hendren and Hale137,138 reported using
CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 905
an electromagnetic field to pull together metallic “bullets”
placed in the two ends of the esophagus in order to shorten
the gap.
Other innovative operative techniques designed to narrow
the long gap before attempted surgical esophageal anastomo-
ses have been advocated. In 1971 Rehbein and Schweder139
described the use of a nylon thread bridging the gap between
the two ends of the esophagus and attached to silver olives136
residing within each lumen. The olives were pushed toward
each other over time until the two ends of the esophagus
pressed together and created a fistula. Shafer and David140
used a similar technique to create a fistula between the two
ends of the esophagus by simply approximating the two ends
of the mobilized esophagus with tacking sutures, then placing
a bridging silk suture. A modification of this technique has
recently been reported successful in five infants.141 Another
method that facilitates esophageal lengthening is the multi-
staged, extrathoracic elongation technique of Kimura, in
which the upper part of the esophagus is mobilized and
initially brought out as an end cervical esophagostomy. Every
2 to 3 weeks, the esophagus and its cutaneous stoma are
surgically mobilized and translocated down the anterior chest
wall until enough length is achieved to perform an end-to-end
esophageal anastomosis.142,143 Foker has described a novel
technique in which traction sutures on both the proximal
and distal esophageal pouches exit through the chest wall
and are serially pulled in opposite directions until the pouches
approximate (Fig. 69-17). The external traction technique
is reported to induce esophageal growth and expedite approx-
imation of the pouches, thus allowing for earlier primary
repair (10 to 14 days).144,145 Internal traction sutures have
also been used to promote esophageal growth in cases where
the gap is such that an initial anastomosis would perhaps be
under too much tension. Successful thoracoscopic esophageal
elongation for long-gap EA has also been reported.146 There is
also a report of using the Kimura technique for proximal
pouch lengthening and the Foker technique for distal pouch
lengthening.147
In addition to preoperative attempts to elongate the esoph-
agus, many intraoperative techniques have been used to estab-
lish a primary esophageal anastomosis, either at an initial
operation in the newborn period or at the time of a delayed
operation after the elongation methods previously described
have been used. Bagolan advocates using the “Foker traction”
principle intraoperatively, finding that significant length can
be obtained after 20 to 30 minutes of traction on the esopha-
geal ends.147a Livaditis and colleagues148,149 described the use
of a circular myotomy of the upper part of the esophagus to
decrease the tension on an end-to-end esophageal anastomosis
in a piglet model of EA and, in 1973, reported the clinical use
of this technique to lengthen the upper esophageal segment in
order to allow for a primary anastomosis in an infant with
long-gap EA (see Fig. 69-11).150 Since that time, proximal
esophageal circular myotomy to lengthen the upper esopha-
geal pouch and achieve a primary anastomosis in long-gap
EA has been widely used. The use of balloon catheters or
cuffed endotracheal tubes inflated in the upper esophageal
pouch to facilitate mobilization and subsequent circular myot-
omy seems to be a useful modification of this method.151–154
We have occasionally made an additional right cervical inci-
sion to mobilize the upper esophageal pouch up and out of
the neck and perform a more proximal second or third circular
myotomy.132
906 PART VI THORAX
French needle
Roll
(nasal pack) 4–0 silk
passing through
the thoracic
wall using
French needle
Cotton roll
Application
of clips just
above the roll
Traction on
the suture
A B
FIGURE 69-17 Foker technique for lengthening the upper and lower esophageal pouches by external traction sutures. A, External view of crossed
traction sutures and bolsters. B, Internal view. (From Foker JE, Linden BC, Boyle EM Jr, et al: Development of a true primary repair for the full spectrum
of esophageal atresia. Ann Surg 1997;226:533.)
However, significant complications of circular myotomy
have been described and have created a debate among
surgeons regarding its safety. Complications include esophageal
leak, impaction of food particles in the myotomized segment,155
and mild ballooning of the myotomized segment. Severe bal-
looning resulting in an esophageal pseudodiverticulum has
been reported to be a potentially serious complication. Several
reports have noted the formation of a pseudodiverticulum with
a distal anastomotic stricture, as might be expected.152,155,156
If the stricture is not treated, the pseudodiverticulum can
enlarge so much that it causes complications such as dysphagia
and life-threatening respiratory obstruction.156
Several reports have suggested the use of a distal esopha-
geal pouch circular myotomy in addition to proximal myot-
omy to achieve a primary esophageal anastomosis.157,158
Many technical modifications have been suggested to mini-
mize the risk for potential serious complications from circular
myotomy. The use of a spiral upper pouch myotomy with obli-
que suture closure of the muscular layer has been advocated as
a technique that minimizes the chance of diverticulum forma-
tion and, theoretically, improves motility.159,160 Another
method used to elongate the upper pouch involves the crea-
tion of a full-thickness anterior flap of the upper pouch wall
as described by Gough; when folded distally, the flap can be
rolled into a tube and attached to the lower esophageal
segment.161 In 1989 Bar-Maor and colleagues162 emphasized
4–0 silk
One clip
Two clips
the technical details of this method to elongate the upper
pouch. Davenport and Bianchi163 reported good follow-up
results in 25 infants with EA treated by this method over a
5-year period. Variations of this technique have been des-
cribed and consist of the creation of a 5-mm posterior flap164
or anterior esophageal flap,165 followed by an end-to-end
oblique anastomosis with a spatulated distal esophagus.
At the time of attempted initial or delayed primary repair,
the esophageal segments sometimes do not reach together
even though one or more of the methods just presented have
been tried. Several additional techniques have been described
to facilitate an esophageal anastomosis. Despite the long-held
opinion that the blood supply to the distal end of the esoph-
agus is tenuous and may be compromised by mobilization,
many surgeons have found that the distal esophagus can be,
and often is, mobilized to facilitate a primary anastomosis
(Fig. 69-18).166 We have successfully achieved primary
end-to-end esophageal anastomosis by completely mobilizing
the distal esophagus down to and even through the esophageal
hiatus of the diaphragm. With this approach, some of the fun-
dus of the stomach can be brought up into the chest to facil-
itate anastomosis.132
Taking this concept further, Schärli described a combined
abdominal and thoracic procedure in which distal esophageal
elongation was achieved by ligation and division of the left
gastric artery, transverse or diagonal division of the lesser

FIGURES 69-18 Aggressive mobilization of distal esophageal pouch in a
case of isolated esophageal atresia.
curvature of the stomach, and mobilization of the gastric
cardia and upper fundus into the chest to achieve primary
esophageal anastomosis.167,168 A partial fundoplication is
recommended to treat the anticipated GER. Not surprisingly,
postoperative GER has been reported to be a significant prob-
lem after this procedure.169 By use of a similar approach, a
Collis gastroplasty with Nissen fundoplication has also been
described as a means of lengthening the distal end of the
esophagus to avoid esophageal replacement.170,171
In addition to the surgical techniques used to facilitate
saving the native esophagus as the conduit of choice for the
repair of long-gap EA, the use of postoperative head flexion,
paralysis, and mechanical ventilatory support has been
strongly advocated by some to minimize the risk for anasto-
motic disruption when the esophageal anastomosis is under
substantial tension.68,136,172,173
The rationale for this approach is the prevention of disrup-
tive force at the anastomotic site by flexion of the neck and
paralysis of the striated muscles in the proximal part of the
esophagus.136,174
ESOPHAGEAL REPLACEMENT
For an extensive discussion of esophageal replacement, see
Chapter 71. Despite the many methods and innovations
developed to achieve esophageal continuity in infants with
EA, the esophagus may need to be replaced if these methods
fail. Numerous operative procedures have been described for
esophageal replacement in infants with EA, as well as children
with caustic injury.
The choice of esophageal substitute depends on many
factors. Colon replacement, or ileocolon, has been widely
practiced for many years as a method of esophageal replace-
ment. Either the right or left colon can be placed substernally
or behind the hilum of the lung on the right or left side.
Vagotomy and a gastric drainage procedure are typically
performed to avoid stricture or ulceration at the cologastric
anastomosis. Several complications occur after colonic inter-
position including cervical anastomotic leak (30% to 50%
of cases), stricture, and intrathoracic redundant colon with
stasis, gastric reflux, respiratory problems, and diarrhea. Some
surgeons favor a reversed gastric tube as a substitute; in this
procedure, a tubularized portion of the greater curvature is
brought up to the cervical esophagus in the substernal or ret-
rohilar position. Complications are similar to those described
for colonic interposition. A modified gastric tube has also been
CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 907
described in which a portion of the greater curve is fashioned
into a “free” tube graft based on the right gastroepiploic
artery.175 The jejunum has also been used for esophageal
replacement both in a Roux-en-Y fashion and as a free graft
with microvascular anastomosis.176
Gastric transposition has more recently become a well-
established and successful esophageal substitute in infants
with EA.136,177,178 Spitz and colleagues40 have reported good
to excellent outcomes in 90% of 138 infants and children with
EA in whom this method of replacement was used (age at
operation ranged from 7 days to 17 years).179 The survival rate
was 95.4%, the incidence of anastomotic leak was 12%, the
incidence of strictures requiring dilation was 20%, and there
was no deterioration in function over a 10-year follow-up.
A recent review by one of us (AGC) reports excellent outcomes
in more than 110 infants treated with gastric transposition for
long-gap esophageal atresia.180 The technical aspects are dem-
onstrated in Figure 69-19. A laparoscopic approach to gastric
transposition has recently been reported and appears to be
a viable, less invasive option.181
ESOPHAGEAL ATRESIA WITHOUT
TRACHEOESOPHAGEAL FISTULA
Infants born with isolated EA have almost no esophagus in
their thorax. It is important to be aware of this fact in order
to institute an appropriate treatment plan and avoid fruitless
exploratory thoracotomy. The preoperative and operative
treatment of this lesion was discussed extensively earlier in
this chapter. In a review of 69 infants with isolated EA treated
over a period of 50 years, Ein and Shandling182 reported a
52% incidence of prematurity, an 11% incidence of Down
syndrome, and a 10% incidence of duodenal atresia; all these
rates are higher than typically reported in patients with
EA-TEF. A recent 24-year review of the experience from Great
Ormond Street Hospital for Children in London found that
71% of patients were male. Among the associated anomalies
were 19% cardiac, 16% urogenital, 14% vertebral, and 7%
anorectal malformation. A more recent review reports an
85% esophageal salvage rate but also describes the complicated
preoperative and postoperative course for these infants includ-
ing 60% anastomotic leak rate and severe gastroesophageal
reflux disease.183
The clinical manifestations of an infant with isolated EA are
similar to those of an infant with EA-TEF in terms of inability
to swallow; in the former, however, the abdomen is typically
scaphoid, and plain radiographs of the abdomen show no air
in the gastrointestinal tract. A gastrostomy tube should be
placed within the first 24 to 48 hours of life, and delayed pri-
mary repair using the various methods previously discussed
should be attempted. Placement of the gastrostomy tube
allows the early institution of enteral feedings, which leads
to subsequent enlargement of the diminutive stomach and
stretching of the distal esophageal pouch (Fig. 69-20).
Enlargement permits the stomach to be used as an esophageal
substitute if necessary. At the time of placement, a contrast
imaging study can be performed to evaluate the unusual pos-
sibility of an occluded but present distal esophageal fistula to
the trachea, which might suggest the potential for early esoph-
ageal repair.112,113 A rigid bronchoscopy can also be done to
eliminate the possibility of a proximal or occluded distal fis-
tula. A period of observation and enteral nutritional support
908 PART VI THORAX

A B

C D
FIGURE 69-19 A to D, Gastric transposition technique.

is instituted with or without daily upper pouch bougienage.
Over an 8- to 12-week period, the process of esophageal elon-
gation can be monitored by imaging studies that use contrast
material, metallic probes, bougies, or flexible endoscopy
placed in the upper and lower esophageal segments
(Fig. 69-21, A and B).
When no further elongation progress is evident or when
the two ends of the esophagus can be brought to close prox-
imity, an operation is performed. Thoracotomy or a thoraco-
scopic approach with dissection and mobilization of the
proximal and distal esophageal segments is done and primary
anastomosis is attempted. If these efforts do not allow primary
anastomosis, circular myotomy, hiatal mobilization, Kimura
procedure, proximal pouch flap esophagoplasty, or Foker’s
traction-elongation technique can be undertaken. When the
two ends of the esophagus are clearly too distant for primary
esophageal anastomosis despite preoperative attempts at elon-
gation and consideration of these methods (i.e., lower esoph-
agus does not reach above the diaphragm), we have proceeded
with a primary gastric transposition to the cervical esophagus,
typically without a thoracotomy. In these latter circumstances,
many surgeons prefer to proceed to a cervical esophagostomy,
preferably on the left side, with plans for a future esophageal
replacement procedure. Postoperative complications are sim-
ilar to those in infants with repaired EA-TEF (see the later sec-
tion on outcomes); these adverse effects primarily include
anastomotic leak and stricture, dysphagia, and gastroesopha-
geal reflux disease (GERD).
 CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 909

the neonate repeatedly chokes on attempting to feed or has

FIGURE 69-20 Gastrostomy tube with contrast injection in an infant
with isolated EA.
ISOLATED (H-TYPE) TRACHEOESOPHAGEAL
FISTULA
Congenital TEF without EA or the “H” type of TEF (perhaps
more anatomically accurately called the “N” type) occurs in
approximately 4% of esophageal anomalies reported. This
anomaly is usually manifested in the first few days of life when
unexplained cyanotic spells.184 When the infant is coughing
or crying, intermittent abdominal distention can occur as
air passes through the fistula into the stomach. Older infants
and children are more likely to have recurrent bouts of pneu-
monia, typically involving the right upper lobe. A high index
of suspicion is necessary in considering this diagnosis.
The diagnosis of isolated TEF can be suspected if plain
radiographs of the chest show evidence of aspiration pneumo-
nitis with gastric distention. A reliable way to establish the
diagnosis is by tube video esophagography performed while
the infant is prone; a small nasogastric tube is passed into
the distal end of the esophagus, and contrast medium is grad-
ually injected as the tube is slowly withdrawn. The radiologist
performing the study must be familiar with this diagnostic
approach because more than 50% of H-type fistulas may be
missed on routine esophageal contrast swallow studies. Bron-
choscopy with esophagoscopy can typically confirm the diag-
nosis and, if done immediately before surgery to divide the
fistula, allows for the passage of a fine catheter or guidewire
through the fistula to aid in subsequent identification at
surgical exploration.185
Most isolated TEFs can be successfully divided through a
cervical approach.186 The site of a right-sided, low cervical
incision is determined, and the infant is placed with the
head extended and turned to the left. The sternocleido-
mastoid muscle is retracted posteriorly, with division of
the sternal head if necessary; dissection then proceeds medi-
ally to the carotid sheath. Identification of the trachea and
esophagus is facilitated by palpating the endotracheal and

A B
FIGURE 69-21 A chest film taken with a Replogle tube pushing in on the upper esophageal pouch and a contrast injection through the gastrostomy tube
demonstrates a long-gap isolated esophageal atresia (A). After 8 weeks, a repeat study with a bougie coiled in the upper esophageal pouch and a neonatal
flexible endoscope passed through the gastrostomy site and up the distal esophageal pouch allows for assessment of interval narrowing of the gap (B).
910 PART VI THORAX
nasogastric tubes. Division of the inferior thyroid artery and
middle thyroid vein may be necessary to expose the plane
between the trachea and esophagus. The recurrent laryngeal
nerve must be identified and preserved. Identification of the
fistula is facilitated by encircling the esophagus with Silastic
slings, but it is important to be aware that the contralateral
recurrent laryngeal nerve can be damaged during this
maneuver. Once the fistula is identified (it is often higher
than expected), traction stitches should be placed close to
the esophagus through the superior and inferior extent of
the fistula to avoid rotation of the esophagus after division
of the fistula. On the tracheal side, 5-0 polypropylene
sutures are placed at the superior and inferior limits of
the fistula. The fistula is now divided close to the esophagus,
and interrupted 5-0 Vicryl or PDS sutures are placed on the
tracheal side to close the fistula. The esophageal side of the
fistula is closed with absorbable polyglycolic acid sutures
(Fig. 69-22). Some surgeons advocate interposing muscle
tissue between the two opposing suture lines to reduce
the likelihood of recurrence of the fistula. A right thoracot-
omy or thoracoscopic approach can be used on the rare
occasion when the fistula is located well within the thorax
or when a recurrent fistula from a previous EA repair is
being approached.187
Postoperative complications include respiratory distress
secondary to edema of the trachea or injury to the recur-
rent laryngeal nerves. The degree of preexisting lung dis-
ease and concerns about tracheal edema may warrant
leaving the endotracheal tube in place for several days after
the operation. Esophageal leak and recurrence of isolated
TEF are rare.
A B
FIGURE 69-22 H-type tracheoesophageal fistula. A, Anatomic relationship between the fistula and the recurrent laryngeal nerve. B, After division, the
fistula is closed by interrupted suture on both the trachea and esophagus.
ESOPHAGEAL ATRESIA WITH AN UPPER
POUCH FISTULA
There are two versions of this anomaly: (1) proximal pouch
fistula in association with distal TEF (double fistula) and
(2) proximal pouch fistula without distal TEF. Several reports
have cited varying incidences of a fistula between the upper
esophageal pouch and the trachea. The exact incidence of this
type of fistula has probably been underestimated because
some initially unrecognized proximal pouch fistulas have
been reported as recurrent TEFs after repair of EA with distal
TEF,188 and some of these cases have been diagnosed as pure
EA because the upper TEF was missed. A recent report by Bax
and colleagues suggests that proximal pouch fistula without
distal fistula is as high as 5.6% of all patients with EA.189
The diagnosis can be made by a preoperative proximal pouch
contrast study, preferably performed by a skilled pediatric
radiologist under optimum conditions in the radiology suite
using fluoroscopy and video recording or in the neonatal unit
if necessary.190 This study usually shows a small upper pouch
in addition to identifying the fistula. However, a frequent
point of confusion during this study and in the interpretation
of results is the occurrence of spillover of contrast from the
pouch up into the larynx and down into the trachea. Because
of the potential for inaccurate interpretation, some surgeons
rely on preoperative bronchoscopy to make or confirm the
diagnosis of a proximal pouch fistula (Fig. 69-23).190 Unfor-
tunately, a small proximal fistula can also be missed when
endoscopy is used. Another common approach to the diagno-
sis of this anomaly is complete mobilization of the upper
pouch during repair of the EA in order to localize and repair
 CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 911

suggest that infants with EA who appear to be at increased risk

for death and long-term morbidity include those with (1)
lower birth weight (<1500 g)/prematurity,64,136 (2) major
CHD,66,191,193a (3) severe associated anomalies and ventilator
dependency,92 and (4) a long gap length between the two ends
of esophagus.104

Complications
------------------------------------------------------------------------------------------------------------------------------------------------

Despite excellent long-term survival for infants with EA, many

significant complications can occur. In a recent evaluation of
134 EA patients older than 15 years, 49% had complications
in the first year and 54% had complications after 1 year. Pre-
dictors of complications included twin birth, preoperative in-
tubation, birth weight less than 2500 g, long-gap atresia,
anastomotic leak, postoperative intubation longer than 4 days,
and inability to feed at the end of 1 month.197 Complications
can generally be classified as early (including anastomotic
leak, anastomotic stricture, and recurrent tracheoesophageal
fistula) and late (including GERD, tracheomalacia, respiratory
disease, and disordered esophageal peristalsis).
FIGURE 69-23 Bronchoscopic view of an unexpected upper pouch
tracheoesophageal fistula. EARLY COMPLICATIONS
an unsuspected upper pouch fistula. As with the other diag- Anastomotic Leak
nostic methods mentioned, proximal pouch fistulas may also Whether open or thoracoscopic techniques are used, anasto-
be unrecognized during surgical exploration, in part because motic leak at the esophagoesophagostomy occurs in approx-
of the extreme proximal nature of some of these fistulas. imately 13% to 16% of patients with EA.123,198,199 Most
A proximal pouch fistula should be suspected and specifically leaks are clinically insignificant and can be managed with
looked for if the upper pouch is unusually narrow or short adequate drainage and nutritional support (Fig. 69-24). When
at the time of preoperative imaging studies or at surgical a retropleural approach is undertaken and a patent mediasti-
exploration because such findings suggest decompression nal drain is in place, up to 95% of anastomotic leaks close
through a fistula. spontaneously.200 Even when transthoracic repair is followed
Once identified, a proximal fistula should be surgically by disruption and the pleural space is contaminated, adequate
ligated and divided as with a distal TEF. If an upper pouch
fistula is diagnosed after the initial repair of EA with or without
distal TEF, a cervical approach can often be used as previously
described for an H-type TEF.

Outcomes
------------------------------------------------------------------------------------------------------------------------------------------------

Survival rates for infants who have EA with or without TEF

have improved dramatically in the past 50 years. Recent
reviews reported 85% to 95% overall survival as compared
with rates less than 40% before the 1950s.64,65,105,191–194
However, subsets of infants with EA still have a worse progno-
sis. Waterston’s risk classification scheme historically helped
identify infants with EA who were at particular risk for a
poor outcome and as such helped direct treatment options.195
In recent years, other risk classification schemes have been
proposed. A recent modified classification scheme suggests
that revised class I (low-risk group) consisted of patients
without major cardiac anomalies and birth weight greater than
2000 g; class II (moderate-risk group) consisted of patients
without major cardiac abnormalities and birth weight less
than 2000 g; class III (relatively high-risk group) consisted
of patients with major cardiac anomalies and birth weight
greater than 2000 g; and class IV (high-risk group) consisted FIGURE 69-24 Anastomotic leak at the esophagoesophagostomy. The
of patients with major cardiac anomalies and birth weight leak is contained and adequately drained. Right upper lobe atelectasis
less than 2000 g (Table 69-4).196 Other recent evaluations is common.
912 PART VI THORAX
drainage can usually be achieved. This in turn allows sponta-
neous closure of the leak. Breakdown of the anastomosis is
frequently followed by the formation of a stricture at the site
of the leak and is sometimes associated with a recurrent TEF.
Major disruptions of the esophageal anastomosis account
for only 3% to 5% of postoperative leaks and are typically
recognized early (24 to 48 hours) after the initial repair.
The infant frequently deteriorates from tension pneumothorax
or mediastinitis uncontrolled by drainage and antibiotics.
Factors that probably contribute to anastomotic leak include
poor surgical technique, ischemia of the esophageal ends, use
of myotomy, and excessive tension at the anastomotic
site.201,202 In this critical setting, reoperation for control of
sepsis with adequate drainage and attempted repair of the
anastomotic leak are warranted. A pleural or pericardial patch,
with or without an intercostal muscle flap buttress, may help
secure anastomotic closure.203,204 If the esophageal anasto-
mosis is not repairable, cervical esophagostomy and delayed
esophageal replacement may be required.
Esophageal Stricture
Esophageal stricture is a common complication of anastomo-
sis of the esophagus in EA, but the reported incidence varies
widely depending on the criteria used to define a stricture.
In a recent large series, stricture requiring dilatation is
reported to occur in up to 80% of patients.68,105,116,147a Spitz
and Hitchcock91 proposed that stricture be defined as the
presence of symptoms (dysphagia and recurrent respiratory
problems from aspiration or foreign body obstruction) and
narrowing noted on endoscopy or contrast esophagography
(Fig. 69-25). Factors that have been implicated in the patho-
genesis of esophageal stricture include poor anastomotic tech-
nique (excessive tension, two-layered anastomosis, and silk
suture material), long gap, ischemia at the ends of the esoph-
agus, GERD, and anastomotic leak. Reports of thoracoscopic
repair of EA note stricture rates similar to many open
FIGURE 69-25 Barium esophagogram showing an anastomotic stricture.
operations (11% to 13%).123,206 Up to 13% of infants and
children will require removal of an esophageal foreign body
after repair of EA with or without TEF, even in the absence
of a stricture on contrast studies.207 Parents have to be coun-
seled about slow introduction of solids and keeping the food
minced until the child has enough molar teeth to grind the
food adequately because the scar in the area of anastomosis
always remains.
Clinically significant narrowing at the site of the esophageal
anastomosis is traditionally treated by dilation performed via
antegrade or retrograde bougienage. We have found that pass-
ing Savary-Gilliard type dilators prograde over a guidewire has
many advantages: This method allows fluoroscopic assess-
ment during sequential dilations, enables the use of contrast
injection during the dilation session, and typically eliminates
the need for rigid esophagoscopy. More recently endoscopic
balloon dilators have been frequently used with good
results.208,209 This technique has the theoretic advantage of
producing a uniform and radial force at the site of the stricture
rather than the shearing axial force applied with traditional
bougienage.210–213 Many strictures respond to one to three
dilations (53%) in the first months after esophageal repair.
However, a recalcitrant stricture resistant to repeated dilations
will require resection and reanastomosis or even esophageal
replacement. Injection of triamcinolone at the stricture site
during flexible esophagoscopy is used in many centers, mostly
by pediatric gastroenterologists, but repeated injections may
lead to adrenal suppression. Recently, the application of
mitomycin C to the stricture under endoscopic control has
been reported to reduce stricture formation after dilation. It
is crucial to determine whether the esophageal stricture is
associated with GERD by investigation with contrast esopha-
gography, endoscopy, pH monitoring, or any combination of
these studies. Because all EA patients demonstrate GER to
some degree, children are routinely prescribed histamine-2
blockers at the time of discharge. In the presence of an anas-
tomotic stricture or persistence of acid reflux, proton-pump
inhibitors may be preferable. Many strictures do not respond
to dilation attempts if severe GERD continues to bathe the
stricture with acid.212 When medical management of GER
fails and the stricture does not respond to esophageal dila-
tions, an antireflux procedure may be indicated. The stricture
is often successfully treated with dilations after the antireflux
operation; however, care must be given to not dilating the fun-
doplication because that significantly increases the possibility
of recurrent GERD. The most recalcitrant strictures have been
treated by resection and reanastomosis, or more recently by
stenting. However, experience is still limited with the latter,
and a fatal arterio-esophageal fistula from an unrecognized
aberrant right subclavian artery has been reported.*
Recurrent Tracheoesophageal Fistula
Recurrent TEF occurs in 3% to 14% of patients after initial
operative division or ligation.105,121,214–217,200 Recurrent
TEF has been attributed to anastomotic leak with local inflam-
mation and erosion through the previous site of TEF repair.
*From a poster presentation, “Arterio-esophageal Fistula from an Aberrant
Right Subclavian Artery after Esophageal Stenting for Stenosis after Atresia
Repair,” at the Annual Meeting of the Canadian Association of Paediatric
Surgeons, Saskatoon, September 2010, by Lo and colleagues.

Techniques that have minimized the likelihood of recurrent
TEF include the use of a pleural flap,215 vascularized pericar-
dial flap,218–220 and azygos vein flap221 interposed between
the esophageal and tracheal suture lines. Although recurrent
TEF typically occurs in the early postoperative period, it
may not be recognized for months to years. Symptoms can
be typical of those seen with a congenital H-type TEF includ-
ing coughing and choking or cyanosis with feedings; however,
less obvious symptoms such as recurrent pulmonary infec-
tions are more common. The diagnosis may be suggested by
an air-filled esophagus on plain radiographs of the chest. Rou-
tine contrast swallow studies will miss up to 50% of recurrent
fistulas. As with a congenital H-type fistula, contrast esopha-
gography performed in the prone position under videofluoro-
scopy is a reliable method of establishing the diagnosis.
Bronchoscopy with cannulation of the fistula with a 2- to
3-French catheter is also a reliable diagnostic approach and
is invaluable in locating the fistula during the operative pro-
cedure. A recurrent TEF rarely closes spontaneously and
typically requires surgical repair. Thoracotomy with fistula
ligation and division is the operation of choice. To minimize
the chance of recurrent fistulization, which has been reported
in 10% to 20% of patients with a first-time TEF recur-
rence, pleura, intercostal muscle, or pericardium should be
interposed between the esophagus and trachea.222 A recent
large series from the University of Michigan of 26 patients
over 27 years reports postoperative complications to include
leak (27%), esophageal stricture (15%), and recurrent
TEF (11%).223
Endoscopic eradication of TEF by means of various chemi-
cals or diathermy has been increasingly reported.224–226
Diathermy or laser deepithelialization followed by fibrin glue
has been reported in case reports and series with success rates
as high as 85%227,228; however, others continue to report fail-
ure with chemical techniques.229 A recent report of using
argon plasma coagulation reports a 66% success rate.230
LATE COMPLICATIONS
Gastroesophageal Reflux Disease
Gastroesophageal reflux disease (GERD) in infants and chil-
dren with repaired EA is common, and the incidence seems
to be increasing as suspicion and diagnostic investigation have
become more common. In addition, the trend to preserve the
native esophagus with many of the techniques described ear-
lier in this chapter also promotes GERD. The magnitude of the
problem is reflected in the finding that GERD occurs in 30%
to 70% of patients after repair of EA.* The cause of GERD in
this group of infants probably relates to shortening of the
intra-abdominal portion of the esophagus because of anasto-
motic tension and/or esophageal motor dysfunction, either ac-
quired as a result of operative manipulation or intrinsic to the
congenital anomaly itself.232–234 The consequences of GERD
are then magnified by the poor or absent esophageal peristaltic
pump, thus exposing the esophagus to prolonged acid con-
tact. Esophagitis is frequently observed in older children
and adolescents on long-term follow-up of patients with
EA.232,235–238 In several recent long-term follow-up studies,
it appears that adults who had EA repaired as infants continue
*References 105, 147a, 200, 215, 220, 231, 232.
CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 913
to be at risk for GERD and its complications,239,240 including
an apparent increased risk for the development of Barrett
esophagus, perhaps as high as 9%.239–241 In a recent case con-
trol study, 101 patients post EA repair with a mean age of 36
years (range 21 to 57) were compared with 287 controls. The
authors found symptomatic GERD in 34% (vs. 8%), dysphagia
in 85% (vs. 2%), and at endoscopy found an increased inci-
dence of hiatal hernia (28%), Barrett esophagus (11%), esoph-
agitis (8%), stricture (8%), and epithelia metaplasia (21%);
manometry showed nonpropagating peristalsis in 80% and
low distal wave amplitudes in all.242 One case of esophageal
adenocarcinoma was reported in a 20-year-old patient in
whom an EA had been repaired during infancy.243
The diagnosis of pathologic GERD is suspected in patients
with symptoms of vomiting, dysphagia, and recurrent anasto-
motic stenosis, which is occasionally associated with the
impaction of a foreign body or food bolus. In addition, respi-
ratory symptoms such as stridor, cyanotic spells, recurrent
pneumonia, and reactive airways disease may indicate GERD
rather than other conditions such as tracheomalacia.
The diagnosis of pathologic GERD in infants and children
after repair of EA is suggested by upper gastrointestinal con-
trast study. Twenty-four-hour pH probe data, though not as
standardized for children as they are for adults, typically
document pathologic reflux.232 Extensive esophageal mano-
metric studies have consistently documented abnormal
esophageal peristalsis and decreased lower esophageal sphinc-
ter pressures after EA repair; as a result, this test is probably
not helpful in diagnosing GERD.244 Multichannel esophageal
impedance combined with pH monitoring may emerge as a
superior test but is not widely available yet.245
In infants and children with pathologic GERD, aggressive
medical management typically consists of thickening of
feedings, positioning of the infant in a prone or upright
posture, and administration of acid reduction agents such
as histamine-2 blockers, proton pump inhibitors, and proki-
netic agents. However, 45% to 75% of these infants ultimately
undergo antireflux operations because of failed medical
management, failure to thrive, chronic pulmonary infection,
refractory anastomotic stricture, or the development of a distal
esophageal stricture.200,215,220 The choice of antireflux oper-
ation is controversial. The Nissen fundoplication has typically
been considered the best option.234,246–248 However, debili-
tating dysphagia and significant complications including
wrap disruption and recurrent GERD in a third of patients
have been frequent after the 360-degree wrap.200,220,249,250
Because of the generally poor results with Nissen fundoplica-
tion in this setting, some surgeons have used the anterior Thal
partial-wrap fundoplication in these patients. However, as
other surgeons have modified the Nissen fundoplication to
create a short, floppy wrap (1 to 1.5 cm over a large dilator),
this operation continues to be commonly used.246 In addition,
because the esophagus is frequently shortened in the setting of
repair of EA, some surgeons have used the Collis-Nissen
fundoplication to gain intra-abdominal esophageal length.171
Tracheomalacia
Substantial respiratory symptoms occurring after repair of
EA-TEF can be due to tracheomalacia. Tracheomalacia has
been noted to affect up to 75% of pathologic specimens from
patients suffering from EA-TEF; however, the condition is
problematic or significant in approximately 10% to 25% of
914 PART VI THORAX

infants after repair of EA-TEF, approximately half of whom

require surgical correction.215,251–253 It is often difficult to
clinically distinguish these symptoms from those of recurrent
TEF, anastomotic leak, or GERD.254 Tracheomalacia is defined
as generalized or localized weakness of the trachea that allows
the anterior and posterior tracheal walls to come together dur-
ing expiration or coughing. In infants with associated TEF,
structural anomalies of the trachea were identified in 75%
of 40 infants at autopsy, thus suggesting that embryologic
events leading to TEF may contribute to the development of
tracheomalacia.254 The cartilage was shorter than normal
and thereby failed to provide the support necessary to main-
tain a patent airway.254 In infants, the trachea may also be eas-
ily compressed between the aorta anteriorly and the often
dilated upper esophagus posteriorly after repair of EA-TEF;
such compression has been considered a significant contrib-
utor to the pathophysiology of tracheomalacia. Kimura and
colleagues255 studied tracheomalacia with cine computed
tomography and suggested that the primary cause of tracheo-
malacia is related to intrinsic tracheal weakness. This has been
confirmed in the Adriamycin-induced EA-TEF rat model and
may also provide more insight into the pathogenesis of the
anomaly. The level of tracheal collapse is usually in the region
of or just above the original site of the TEF in the distal third of FIGURE 69-26 Tracheomalacia after repair of esophageal atresia-
the trachea, generally at the level of the aortic arch. Perhaps tracheoesophageal fistula. A bronchoscopic view of the tracheal lumen
surprisingly, severe tracheomalacia appears less common in during spontaneous respirations shows almost complete collapse of the
trachea during expiration.
infants with pure EA,256 but it has been reported.182
The clinical manifestations of tracheomalacia are broad and
range from a “brassy” or “barking” cough in mild cases to which the aortic arch would not reach the posterior aspect of
recurrent pneumonia or acute, life-threatening apneic spells. the sternum without undue tension. An anterior mediastinal
Infants with tracheomalacia are often reluctant to feed because approach via a low transverse cervical incision for aortopexy
of difficulty breathing during feeding or cyanotic attacks. and trachelopexy has recently been advocated.263 Recently,
These symptoms usually appear after discharge from the hos- a thoracoscopic approach for aortopexy has been reported
pital, when the infant is a few weeks to a few months of age.256 to be a good alternative to thoracotomy.264–266 Many surgeons
Life-threatening apneic spells were noted in 27 of 32 children use intraoperative bronchoscopy to document the adequacy of
with tracheomalacia described by Filler and colleagues.257 aortopexy with regard to opening the tracheal lumen as
These spells occur during or within 5 to 10 minutes of a meal sutures are being tied or even before suture placement. Other
and are characterized by cyanosis progressing to apnea, bra- surgical treatment options for tracheomalacia associated with
dycardia, and ultimately, cardiorespiratory arrest if not inter- EA have been suggested. The role of various stents in the treat-
rupted. The diagnosis is established by bronchoscopy with ment of tracheomalacia after repair of EA, though successful in
spontaneous ventilation, which reveals a slitlike lumen of case reports, has yet to be validated as a standard treatment
the trachea at the involved area (Fig. 69-26). Because the method.267–270 Filler and colleagues257 recommend that an
symptoms overlap those of a stricture or GER, the initial inves- airway stent be considered for children in whom aortopexy
tigation usually consists of a contrast esophagram; close atten- fails to relieve tracheal collapse or for the rare cases of diffuse
tion to the tracheal air column on the lateral views during tracheobronchomalacia, for which aortopexy is not useful.
such a study will often reveal complete tracheal collapse dur- Recently, glossopexy has been suggested as an early alternative
ing forced expiration (i.e., crying) or when contrast fills a to aortopexy to treat infants with suspected retrodisplacement
distended upper esophagus just above the anastomosis. of the tongue (glossoptosis).271 Tracheostomy is a final treat-
Treatment of tracheomalacia remains controversial. Most ment option if aortopexy fails, which is uncommon. It is fairly
infants with mild to moderate symptoms of tracheomalacia common to detect significant GER in patients with severe tra-
do not require operative intervention because the symptoms cheomalacia; in such cases, we would advocate aortopexy
tend to improve with time. In infants with severe symptoms before considering a fundoplication.
including acute life-threatening events, the operative treat-
ment of choice is aortopexy.258–261 This operation is usually Other Long-Term Quality-of-Life Issues
performed through a left anterior mediastinotomy (Chamber- As previously discussed, the esophagus of an infant with EA
lain approach) or anterolateral thoracotomy, and the ascend- has abnormal peristaltic activity that is secondary to the con-
ing aorta and arch are sutured up to the posterior surface of genital defect itself, as well as secondary to operative repair of
the sternum after partial thymectomy. Lifting the aorta up in the lesion. The incidence of the problem approaches 75% to
this fashion raises the anterior wall of the trachea and opens 100% in children and adults who have been studied with
the tracheal lumen.256,261 A proposed modification of esophageal manometry. Dysphagia is worse after intestinal
this operation has been the use of a flap of pericardium based interposition for esophageal replacement. Occasional dyspha-
at the root of the aorta to be sutured to the sternum in cases in gia has been reported to occur in up to 50% to 90% of adults,
 CHAPTER 69 CONGENITAL ANOMALIES OF THE ESOPHAGUS 915

with 30% having additional problems with choking.253,272,273 The second type of CES has been called idiopathic muscular
This disorder is clinically significant in that it is responsible for hypertrophy or fibromuscular stenosis and in some reports has
many long-term symptoms after EA repair including dyspha- been the most common form of CES.277 Its histologic charac-
gia, esophageal foreign body obstruction, GERD, and recur- teristics are submucosal proliferation of smooth muscle fibers
rent respiratory problems. Although these symptoms may and fibrous connective tissue with normal overlying squa-
improve with time, early problems with feeding intolerance mous epithelium.277 A resemblance to hypertrophic pyloric
and food bolus obstruction can lead to failure to thrive. Many stenosis has been suggested,277,283 but no clear embryologic
children must eat slowly and be selective with food choices. or pathogenic factors explain these lesions.
In a series of EA patients followed for 10 years, Lacher and CES caused by tracheobronchial remnants is described as
colleagues274 found 72% swallowed without problems; the most common type of CES in some reports and is certainly
however, 41% had poor weight gain (<25th percentile). the most described and understood of the three types of CES.
In addition, recent reports have examined long-term respi- It is believed that CES caused by tracheobronchial remnants
ratory outcomes after EA repair. One recent paper reports a occurs as part of a spectrum of anomalies, including EA-TEF,
greater than 60% incidence of recurrent respiratory infections, related to separation of the foregut from the respiratory tract
a 25% decrease in FEV1, and decreased maximal exercise per- around the twenty-fifth embryonic day.121 Tracheobronchial
formance at greater than 5 years beyond EA repair.275 Another tissue is believed to become sequestered in the wall of the
long-term follow-up study found respiratory symptoms in all esophagus and comes to reside in the typical distal location
patients including pneumonia 16%, bronchitis and asthma because of the faster growth rate of the esophagus.
52.3%, cyanosis episodes 9%, and cough 90%.194 In our CES secondary to tracheobronchial remnants was first
experience using small incisions and muscle sparing, thoracic reported in 1936 by Frey and Dusche284 in a 19-year-old
deformities are extremely rare. woman who died with the diagnosis of achalasia.121 In
1964 Holder and colleagues62 reported three cases of distal
Congenital Stenosis esophageal stenosis in their review of 1058 infants with EA,
TEF, or both. After several additional reports,285,286 Spitz,
of the Esophagus in 1973, first demonstrated a clear congenital basis for this
------------------------------------------------------------------------------------------------------------------------------------------------
disorder.287 In addition, 71 cases of CES have been reported
True congenital esophageal stenosis (CES) is a rare condition in the Japanese literature as of 1981279; 6 cases were reported
that has historically been confused with esophageal strictures in the Chinese literature (out of 76 total cases of CES in
secondary to inflammation, especially GERD. CES has been 1987).288 An additional three cases of CES have been
reported to occur once in every 25,000 to 50,000 births276,277; reported, one of each type described earlier.289
as of 2004, fewer than 600 cases have been described in the
world literature. For unknown reasons, the incidence of CES DIAGNOSIS
seems to be higher in Japan than in other parts of the
world.278,279 The incidence of other anomalies associated with Symptoms of CES usually begin in infancy with progressive
CES is reported to be 17% to 33%; these anomalies include EA dysphagia and vomiting, generally after the introduction of
with or without TEF; H-type TEF; cardiac anomalies; intestinal semisolid or solid foods around the age of 6 months. Some
atresia; midgut malrotation; anorectal malformations; hypospa- case reports, however, describe severe symptoms of regurgita-
dias; malformations of the head, face, and limbs; and chromo- tion and respiratory distress in the newborn.277 In some
somal anomalies.277 In addition, a report by Vasudevan and patients, a foreign body in the esophagus may be the first
colleagues280 noted that 5% of their patients with EA-TEF symptom noted.290 The correct diagnosis is frequently diffi-
anomalies had CES with tracheobronchial remnants, whereas cult to establish. Contrast esophagography typically reveals
other reports suggest an incidence as high as 12% to 14% of an abrupt distal esophageal narrowing, most often interpreted
patients with EA had associated CES.278,281 as a stricture related to GERD. Stenosis secondary to fibromus-
Classification schemes for CES have been numerous and con- cular hypertrophy can result in a more tapered narrowing
fusing, in part because of the difficulty in differentiating congen- (Fig. 69-27). CES caused by webs or fibromuscular hypertro-
ital from acquired lesions. The definition and classification phy can be manifested as midesophageal or even upper esoph-
proposed by Nihoul-Fekete and colleagues277 are perhaps the ageal stenosis.291 Over time, the esophagus proximal to the
most clear. CES is defined as an intrinsic stenosis of the esoph- stenosis can dilate, and contrast study results can be inter-
agus, present at birth, that is caused by congenital malformation preted as achalasia.292 Additional studies helpful in differen-
of esophageal wall architecture.277 This classification delineates tiating CES from achalasia and strictures caused by GERD
three forms of CES: (1) a membranous web or diaphragm, (2) include esophageal manometry and pH monitoring. Esopha-
fibromuscular thickening, and (3) stricture secondary to tra- goscopy typically shows esophageal narrowing with normal-
cheobronchial remnants in the wall of the esophagus. appearing mucosa at the level of the stenosis in cases of
A congenital membranous web or diaphragm is reported as CES. Recently, high-frequency endoscopic ultrasonography
the rarest of the three forms of CES.282 It has been considered has been reported to be helpful in the diagnosis of CES.293
to represent a missed form of EA25 and may be analogous to
membranes in other parts of the gastrointestinal tract. It is
TREATMENT
usually a partially obstructing lesion located in the middle
or lower portions of the esophagus. The membrane is covered Treatment of CES should relieve the symptoms of stenosis
on both sides with squamous epithelium and often has an and maintain the antireflux mechanism of the gastroesopha-
eccentric opening. Symptoms typically occur at several geal junction. Bougienage has been successful in treating
months of age as the infant begins consuming solid food. CES secondary to esophageal webs and fibromuscular
916 PART VI THORAX
FIGURE 69-27 Barium esophagogram showing narrowing of the mide-
sophagus as a result of congenital esophageal stenosis.
hyperplasia. Antegrade and retrograde tapered dilators have
been the traditional form of bougienage, but more recent use
of hydrostatic balloon dilation has been successful.277,292,294,295
A series of dilations may be required for resolution of the
stenosis. Membranous webs are typically adequately treated
with dilation, and one report described successful endoscopic
excision of a congenital web.280,289,296,297
Traditionally, most cases of CES secondary to fibromuscu-
lar hypertrophy and tracheobronchial remnants have been
treated by surgical excision, either as the primary approach
or after failed attempts at dilation.294
More recent recommendations suggest dilations should be
aggressively attempted before committing to resection.295
When resection is required, it is important to clearly identify
the location of the stenosis before surgery by contrast esopha-
gography in order to plan the operative approach. A right tho-
racotomy is typically used for stenosis in the midesophagus
and a left thoracotomy for stenosis in the lower part of the
esophagus. An abdominal approach should be used for CES
in the abdominal portion of the esophagus. At exploration,
the extent of the stenosis can be difficult to determine; the
use of a balloon catheter passed beyond the stenosis, inflated,
and pulled back against the stenosis has been suggested as a
helpful technique.298 Alternatively, a bougie passed from
above or preoperative flexible upper endoscopy may demon-
strate the stenotic area. In most cases, the stenosis is less than
3 cm in length and segmental resection and end-to-end
esophageal anastomosis can be accomplished. Care should
be taken to preserve the vagus nerves. Case reports have
documented innovative surgical approaches including steno-
sis resection using limited myectomy, leaving the mucosa
intact,299 as well as thoracoscopic300 and laparoscopic301
approaches. In cases of long fibromuscular hypertrophy
unresponsive to dilation, resection and esophageal replace-
ment with the colon, stomach, or jejunum may be necessary.
If the stenosis is near the gastroesophageal junction, most
surgeons advocate segmental resection with esophageal anas-
tomosis and an antireflux procedure to prevent postoperative
reflux. Modified Hill gastropexy and Nissen fundoplication,
with or without pyloroplasty, have been most commonly
used277,298; however, Collis gastroplasty in combination
with Nissen fundoplication has been reported to be effective
for managing esophageal shortening and postoperative
GERD.278,294,302
RESULTS
Good long-term results have been reported for dilation and
operative resection.269 In patients who have undergone distal
esophageal CES resection without an antireflux procedure,
significant GERD has developed and required a subsequent
antireflux operation. Reported complications from treatment
by dilation include esophageal perforation and failure of ther-
apy. Complications of resection and anastomosis include
esophageal leak with mediastinitis, which can be successfully
treated by mediastinal drainage.270,272
Laryngotracheoesophageal Cleft
------------------------------------------------------------------------------------------------------------------------------------------------
A laryngotracheoesophageal cleft (LTEC) is a rare congenital
anomaly consisting of a midline communication between
the larynx, trachea, and esophagus. The malformation was
possibly first described in 1792 by Richter in his doctoral the-
sis, in which he described an infant who choked and vomited
on feeding.303 The infant died, but because no autopsy was
performed, the diagnosis was unconfirmed. An infant with
LTEC was next described in 1949 by Finlay.304 In 1955
Pettersson305 performed the first successful correction of a
laryngotracheal cleft.
As with EA-TEF, the embryogenesis of LTEC is not
completely understood. The long-held theory is that there is
an arrest of the cranial extension of the tracheoesophageal sep-
tum that permits the persistence of an esophagotrachea.306–308
More recent studies have suggested that as with EA-TEF, initial
normal development is followed by a far-reaching fusion of
the trachea and esophagus.32 Although no consistent pattern
of inheritance has been seen, sporadic familial associations
have occurred; LTEC is reported with the “G” syndrome and
the Pallister-Hall syndrome.309 The incidence of LTEC favors
males by a ratio of 5:3.310
Various associated congenital anomalies occur in the setting
of LTEC including gastrointestinal, genitourinary, and cardiac
malformations.310 EA with TEF occurs in 20% to 37% of
patients with LTEC.310,311,312 Other associated gastrointestinal
malformations include anal defects (21%), anomalies of rotation
and fixation (13%), and meconium ileus (8%).310,313 Genitou-
rinary anomalies occur with an incidence of 14% to 44%
and include hypospadias, inguinal hernias, undescended testes,
and renal agenesis.310,313 Cardiovascular anomalies, identi-
fied in 16% to 33% of patients, include ventricular septal
defects, coarctation of the aorta, and transposition of the great
vessels.310
 CHAPTER 69
To more adequately delineate therapy, several classification
schemes have been described. In reporting the first surgical
repair in 1955, Pettersson described three types of clefts:
type I, limited to the larynx and involving part or all of the
cricoid plate; type II, extending beyond the cricoid lamina
to the cervical trachea; and type III, involving the entire
trachea down to the carina. In 1991 Ryan and colleagues314
suggested a type IV in which the cleft extends beyond the
carina to involve one or both mainstem bronchi.
Symptoms of LTEC vary depending on the extent of the
cleft, but most patients immediately after birth exhibit respi-
ratory distress aggravated by feeding. Additional symptoms
can include a characteristic toneless or hoarse cry, cyanosis,
choking, increased secretions, and recurrent aspiration pneu-
monia. It is not unusual for the severity of associated anoma-
lies to obscure the presence of LTEC, especially if it is a
minimal type I or II lesion. On the basis of the common symp-
toms of LTEC, the diagnostic evaluation typically proceeds
along the line of the more commonly suspected diagnosis of
EA-TEF or tracheomalacia. Contrast esophagography demon-
strates rapid confluence of contrast material in the upper part
of the esophagus and trachea; however, it is often difficult to
know whether this is secondary to spillover at the level of the
larynx or passage of contrast through a cleft. Rigid endoscopy
under general anesthesia is the definitive method for diagnosis
of LTEC, but it can still be difficult to identify a cleft unless the
index of suspicion is high and unless the frequent mucosal
infolding at the level of the subglottic region is pushed open
to reveal the cleft (Fig. 69-28).
Management of infants with LTEC begins by maneuvers to
minimize aspiration and stabilize the airway. Ryan and col-
leagues314,315 recommended avoiding endotracheal intuba-
tion before surgery if possible, but it is often necessary to
intubate the trachea and perform tube gastrostomy before
definitive surgical repair of extensive LTECs. Operative proce-
dures vary depending on the severity of the cleft. Asymptom-
atic type I clefts may require no operative intervention, and
symptomatic clefts have been successfully repaired endoscop-
ically.310 Type II LTEC can be approached through a lateral
pharyngotomy, posterior pharyngotomy, or anterior laryngo-
fissure. The lateral exposure has been most often reported;
A B
FIGURE 69-28 Endoscopic view of a type III laryngotracheoesophageal cleft. A, Proximal end of cleft may be difficult to appreciate at first glance. B, Distal
end of cleft ends just proximal to the tracheal bifurcation.
CONGENITAL ANOMALIES OF THE ESOPHAGUS 917
this method permits easy access to the cleft and allows asym-
metric incisions in the mucosa, thereby avoiding contiguous
suture lines. The major disadvantage to this approach is
its risk to the recurrent laryngeal nerves. The anterior laryn-
geal approach exposes the larynx and upper part of the trachea
and has no risk for recurrent laryngeal nerve injury. Con-
cern about postoperative laryngeal stability has been raised,
and some surgeons recommend the use of stents during the
healing process.310,316
Operative management for types III and IV LTEC requires a
combined cervical and thoracic approach. Donahoe and col-
leagues311,314,315 have described the use of a specifically
designed bifurcated endotracheal tube with flanged ends that
can be positioned during bronchoscopy to suspend the tra-
chea anteriorly with a ureteral catheter sling (Fig. 69-29).
The endotracheal tube allows for confident airway control
during the multiple position changes frequently required dur-
ing the operation. A right thoracotomy is performed, and the
tracheoesophageal cleft is exposed retropleurally. The tra-
cheoesophageal groove is incised on the right and opened
from the carina to the thoracic inlet. Separation of the two
tubes is completed by incising along the left side of the com-
mon esophagotracheal wall, and an approximately 1-cm flap
of esophagus running the length of the trachea should be left
to help create the neomembranous portion of the trachea. It is
important to size the esophageal flap correctly in order to
avoid stenosis or a floppy posterior wall, which can result
in tracheal obstruction. As the dissection proceeds, it is also
important to push away and thus protect the left vagus and
recurrent laryngeal nerves. Closure of the trachea and esoph-
agus is accomplished in a caudal-to-cranial fashion with run-
ning polypropylene suture for the esophagus and interrupted
polypropylene for the trachea. A right cervical incision is then
made to expose the upper portion of the esophagus, trachea,
pharynx, and larynx; the repair is continued with inclusion of
a three-layer repair of the larynx and placement of a trache-
ostomy tube, which can be custom-designed to avoid undue
pressure against the posterior tracheal repair. Some surgeons
prefer an anterior approach and divide the larynx and trachea
in the midline, as described for type II defects.317 Cardiopul-
monary bypass has been used both as a supporting technology
918 PART VI THORAX

Sling in tracheostomy
site to catch tube
introduced from above
us Clavicle
ch
bron
L.
Pharynx
Esophagus opened

R
Extent of cleft

.b
ro
nc
hu
Tube secured anteriorly s

Holes cut
for upper B
lobe bronchi
A

C
FIGURE 69-29 Repair of a type III laryngotracheoesophageal cleft. A, Stabilization of a bifurcated endotracheal tube is done at bronchoscopy with a loop
passed through a tracheotomy, which draws the endotracheal tube forward. B, A cervical and thoracic approach allows retropleural exposure of the cleft.
A longitudinal incision is made in the right tracheoesophageal groove below the tracheal rings. The incision is extended inferiorly and across the esophagus
and up the left side, with approximately 1 cm of esophageal wall left attached to the trachea to allow adequate tissue to close the trachea. C, The trachea
has been closed with interrupted sutures, and the esophagus is closed in running fashion up to the thoracic inlet. Closure of the laryngeal portion of the
cleft and the lateral pharyngeal wall is not yet accomplished. (From Donahoe PK, Gee PE: Complete laryngotracheal cleft: Management and repair. J Pediatr
Surg 1984;19:143.)

during the anterior approach318 to repair a type IV cleft and as
support in the form of extracorporeal membrane oxygenation
for one patient treated via a lateral thoracic and cervical
approach.
Few patients with these complex anomalies are treated at
any single institution.310,315,319 Postoperative survival rates
continue to be rather poor and range from 50% to 75%,
depending on the severity of the malformation, associated
anomalies, and prematurity. Anastomotic leakage is reported
to occur in approximately 50% of repairs and typically re-
quires reoperation via a different approach.310 In addition,
an inability to wean patients from the ventilator, pharyngoeso-
phageal dysfunction, and GER are common postoperative
complications. A reduction in morbidity and mortality de-
pends, in part, on earlier recognition so that secondary com-
plications can be prevented. At present there is no clear
consensus about the best surgical approach; however, treat-
ment is clearly complex and should be undertaken at centers
with multidisciplinary expertise.
The complete reference list is available online at www.
expertconsult.com.