- Email: [email protected]
Barrett's esophagus in children: A consequence of chronic gastroesophageal reflux
CASTROENTEROLOGY 1984;88:318-23
Barrett’s Esophagus in Children: A Consequence of Chronic Gastroesophageal Reflux BEVERLY
BARRETT
DAHMS
and
FRED C. ROTHSTEIN
Departments of Pathology and Pediatrics, Case Western Reserve University and Rainbow Babies’ and Children’s Hospital, Cleveland, Ohio
It is widely acknowledged that Barrett’s esophagus in adults is an acquired condition resulting from prolonged gastroesophageal reflux. Barrett’s esophagus is rare in childhood, even though gastroesophageal rejlux occurs commonly in the pediatric age group. When a columnar-lined esophagus is present in children, it is often regarded as a congenital of chronic anomaly rather than as a consequence gastroesophageal reflux. Over a 5-yr period (1978 1982), we retrospectively studied Barrett’s esophagus in children 19 yr of age or younger who were evaluated for gastroesophageal reflux and whose symptoms warranted esophagoscopy and esophageal biopsy. Esophageal biopsies were performed on 103 patients with gastroesophageal reflux. Thirteen children [age range, 8 mo-19 yr) had Barrett’s esophagus, for a prevalence of 13%. Gastroesophageal reflux was documented in these children by upper gastrointestinal radiographs or pH monitoring. Radiographs demonstrated esophageal stricture in 5 of the 13 children; none had hiatal hernia. Children presented with symptoms suggestive of gastroesophageal reflux and esophagitis: vomiting, abdominal pain, odynophagia, dysphagia, and heartburn. All children had a past history of excessive regurgitation during infancy. Histologically, three types of columnar epithelium were present: gastric fundic type (11 patients), junctional-type columnar epithelium reminiscent of gastric cardia (7 patients), and specialized columnar (metaplastic intestinal) type (2 patients). We believe that Barrett’s esophagus is more common in children than had previously been appreciated. In these children, we
Received March 14. 1983. Accepted August 30, 1983. Address requests for reprints to: Beverly Barrett Dahms, M.D.. Department of Pathology, University Hospitals of Cleveland, 2085 Adelbert Road, Cleveland, Ohio 44106. 0 1984 by the American Gastroenterological Association 0016-5085/84/$3.00
suggest that the distal columnar-lined resulted from chronic gastroesophageal not a congenital anomaly.
esophagus reflux and is
In the past decade, it has become widely acknowledged that Barrett’s esophagus in adults is an acquired condition resulting from prolonged gastroesophageal reflux (GER) (l-6). Reports of Barrett’s esophagus in childhood are rare, even though GER in the pediatric population is common and has been the subject of several recent reviews (7-10). When columnar epithelium has been described in the esophagus of children, there is a tendency to regard it as a congenital heterotopic anomaly, rather than a consequence of chronic GER (11-14). This study was prompted by the recent appearance of a number of children at our hospital with the classic symptomatic, endoscopic, and histologic manifestations of Barrett’s esophagus. Their medical histories were indicative of gastroesophageal reflux from infancy. The purpose of this paper is to document the occurrence of Barrett’s esophagus in children who present with symptomatic gastroesophageal reflux.
Methods We retrospectively reviewed the biopsy specimens of children, 19 yr of age and younger, who were evaluated for GER and whose symptoms warranted esophagoscopy and esophageal biopsy. All children were referred to the gastroenterology service at Rainbow Babies’ and Children’s Hospital over a 5-yr period (1978-1982) for symptoms suggestive of GER. Clinical evaluation included barium swallow and upper gastrointestinal (UGI) series or esophageal pH monitoring (15) to document GER. Not every patient with a diagnosis of GER underwent esophagoscopy. Only those children with GER and symptoms of esophagitis or those who failed traditional medical management for their GER had esophagoscopy and biopsy studies performed. Esophageal manometry was performed
February
BARRETT’S
1984
ESOPHAGCIS IN CHILDREN
319
to assess lower esophageal sphincter function in some children. Endoscopy was performed using an Olympus pediatric endoscope (Olympus Corporation of America, New Hyde Park, N.Y.). All esophageal mucosal biopsy specimens were obtained under direct visualization either by grasp forceps (through the endoscope) or by suction biopsy (Ruben tube) technique. In most cases, specimens were taken at one level per endoscopic session, usually 2-3 cm proximal to the lower esophageal sphincter (LES). This method of endoscopic visualization of the LES is not as reliable as is manometric localization of the sphincter. However, this retrospective series is based on observations made during routine procedures, and esophageal biopsy by manometric guidance is not routine clinical practice. Two or three different pieces of tissue were usually obtained per biopsy. The level at which biopsy specimens were obtained was recorded in each case. In 2 patients (No. 2 and No. 4 in Table l), multiple biopsy specimens were obtained from different levels of the esophagus at one endoscopic session. All biopsy specimens were fixed in Bouin solution, embedded in paraffin, stained with hematoxylin and eosin, and serially sectioned. Biopsy specimens were diagnosed as having Barrett’s esophagus if (a) they were taken at least 2 cm proximal to the LES by direct visualization at endoscopy and [b) they demonstrated the abnormal columnar epithelium previously described (16,17) as characteristic of this condition: i.e., (a) gastric fundic type mucosa, (b) junctional epitheliurn reminiscent of gastric cardia, or (c) specialized columnar epithelium with a villiform surface and goblet cells. In
order to avoid confusion with normal esophageal-cardiac junction in some of the lower biopsy specimens, the presence of junctional epithelium with cardiac-type mucous glands was considered insufficient by itself to include patients in the study unless Barrett’s mucosa was apparent endoscopically. Patients with a history of esophageal atresia and tracheoesophageal fistula at birth were excluded from the series because of the known high incidence of columnar epitheli-
Table
Barrett’s
1. Clinical
und
Pathologic
Findings
Symptoms
in Children
With
(18).
urn in the esophagus in this condition
Results Clinical Features Esophageal biopsies of 103 children and adolescents were performed over the 5-yr period 19781982, during evaluation of gastroesophageal reflux. Thirteen of these children demonstrated the abnormal columnar epithelia that have come to be recognized as Barrett’s esophagus, for a prevalence of 13%. Patients are tabulated in Table 1. The 13 children with Barrett’s esophagus were 8 mo to 19 yr of age and presented with symptoms of GER or peptic esophagitis. All children had a history of excessive regurgitation during infancy. Eight children presented with a history of acute emesis. Two children were failing to thrive because of caloric losses secondary to chronic vomiting. Two children presented with acute hematemesis, and at least 5 had Esophagus
Radiology
Esophagoscopy
Histology
-I: ._
.z
2
2
+
ND
+
ND
E ; ++ ++
Yes Yes
4
-
2
*
f
5
+
+ +
11 3
14
+
+
+
+
ND
ND
4
15
t
._
+
+
+
ND
5
2.92
*
+
+
+
12.0
+
+
_
+
9.5
_
+
ND
+
ND
+
11.0
+
22.0
++ ++
NO Yes
+t
No
if + -+ ++ + + ++ ++
NO
4
*
2
f
4
t
a 3
7
+
4
6
12
7
6
I
a
7
_
9
3 ..5
i-
+
_
10 11
6 19
+ +
_ +
+ _
-
ND
ND
12
4
+
+
_
-
ND
ND
13
0.67
+
+
+
-
+
Abbreviations:
+
7.0
+ = present; - = absent; GER = gastroesophageal reflux; ND = not done; ” Location gitis + = erythema, friability; esophagitis ++ = ulcers, exudate.
No
NO No Yes Yes
4 3 3 3 4
+
+ -+ + _
4
t + + + +
NO
3
-
No
3
+
LES = lower esophageal = number of centimeters
+ +
+ + + + -t
sphincter; esophaproximal
to LES.
320
DAHMS
AND
GASTKOENTEROI,(~(;‘r’
ROTHSTEIN
had hematemesis in the past. Nine children had Hemoccult-positive stools at the time of initial evaluation. It was difficult to document the presence of certain symptoms (i.e., heartburn and odynophagia) in some children because of either their young age (3 pa(1 patient). tients] or their inability to communicate Heartburn was the most common complaint in the older children (9 patients) and this occurred postprandially, while supine and with exercise. Heartburn was associated with an increased frequency of burping in several children. Nine of the older children had complaints of odynophagia-for liquids (2 patients) and solids (4 patients). The younger children experienced chest and epigastric discomfort. Pain was often severe enough to awaken the child from sleep. Six children had dysphagia. In the cases of the younger children, mothers would complain that their child would prefer liquids to solids and often cough and choke when solids were ingested. Several children had pulmonary complaints that may have been secondary to GER (19-2 1). Five had a history of asthma, 4 had night cough of >3 months’ duration, and 3 had radiographically demonstrated pneumonia on at least two occasions. Upper gastrointestinal radiographs were performed on all children and GER was demonstrated in 9. Five patients had radiographic evidence of either distal or midesophageal strictures, and these patients also had esophageal ulcerations demonstrated. No child had radiographic evidence of a hiatal hernia or gastric outlet obstruction.
Vol. 86.
No.
2
On 8 children, including those 4 with negative UGI studies, pH monitoring was performed. All 8 had GER. Esophageal manometry was performed in 7 children. Six had decreased LES pressures. Two had of a very short LES (SO.5 cm). No specific pattern motor abnormalities of the esophagus was apparent. At esophagoscopy, strictures, located 5-10 cm above the LES, were noted in 5 children. Esophagitis was present in all children, varying in severity from erythema and moderate friability (3 children] to purulent exudates and ulcerations (10children). The presence of columnar epithelium, as evidenced by a salmon-pink color of esophageal mucosa, was apparent to the endoscopist in only 5 children.
Pathology As indicated in Table 1, three histologic types of columnar epithelium were demonstrated in the distal one-third of the esophagus, just proximal to the LES. Abnormal gastric fundic epithelium, prewas the most commonly obsent in 11 patients, served columnar epithelium (Figure 1); in some patients, it was the only columnar epithelium obtained from the esophagus. Characteristically, this mucosa appeared atrophic compared with normal gastric fundus. It was similar to fundic mucosa in being covered by uniform tall columnar cells and containing coiled tubular glands with parietal and chief cells. Seven patients demonstrated the second type of columnar epithelium seen in Barrett’s esophagus, the junctional type reminiscent of gastric cardia.
February
1984
BAKRETT’S
Figure
ESOPHA(:IJS
IN CHILDKEN
32 1
2. This junctional-type epithelium is reminiscent of gastric cardia, with a tall columnar epithelial surface and mucous glands. Note inflammation of the lamina propria. This biops) specimen was obtained from a 2.5. yr-old girl and was taken 4 cm proximal to the lower esophageal sphincter. (H&E. * 120.)
There was considerable variation in the appearance of junctional epithelium from patient to patient. In some, the mucosa closely resembled gastric cardia [Figure 2) with uniform tall columnar surface epithelial cells and a wide glandular layer composed of cardiac-type glands. In others, the surface was more villiform and the glands atrophic and distorted (Figure 3). One patient demonstrated a mixture of cardiac and fundic glands. In all cases, the junctional epithelium exhibited sufficient atrophy, inflammation, or distortion that it was unlikely to be confused with normal gastric cardiac mucosa. The majority of the 7 patients with junctional epithelium demonstrated one of the other types of columnar epitheliurn at other levels of the esophagus. Only 2 patients demonstrated the third type of epithelium, the distinctive specialized columnar epithelium pathognomonic of Barrett’s esophagus (Figure 4). This columnar epithelium has also been called the intestinal metaplastic type because of its villiform surface, goblet cells, and mucosal glands. In addition to these three types of columnar epithelium, squamous mucosa was also obtained from 10 patients. Inflammation was noted in all biopsy specimens of squamous mucosa and consisted of a scattering of neutrophils in lamina propria and epithelium. Eosinophils were present in three biopsy specimens. In several instances, an abnormal squamocolumnar junction was demonstrated, such as the junction between squamous and specialized columnar epithelium in Figure 4. Patient 13 demonstrated continuity of squamous mucosa with gastric fundictype epithelium. None of the columnar epithelia in these children demonstrated dysplasia, although some showed mild nuclear changes consistent with inflammation and repair.
Patients No. 2 and No. 4 had multiple biopsy specimens of the esophagus taken at different levels during one endoscopy session. In each case, gastric fundic mucosa lined the most distal portion of the
This variation of junctional-type epithelium has some features of both gastric and intestinal mucosa. The surface is \rilliform, but is composed of tall columnar epithelium devoid ot goblet cells. Mucous glands are reminiscent of cardiac glands, but are markedly atrophic. The lamina propria is severely inflamed Biopsy specimen is from patient No. 6. (H&E, x100.)
322
Figure
DAHMS
AND
ROTHSTEIN
GASTROENTEROLOGY
Vol.
86, No.
2
4. This photomicrograph shows the specialized columnar type of Barrett’s epithelium, reminiscent of intestinal mucosa because of the villiform surface and intestinallike epithelium containing goblet cells. Note the squamocolumnar junction, upper left. Biopsy specimen is from patient No. 4. (H&E. x60.)
esophagus, nearest the LES. Specialized columnar epithelium was the most proximal type of columnar epithelium, biopsied 5 and 8 cm above the LES, respectively.
Discussion Barrett’s esophagus is more common in children than previously appreciated. The clinical presentation, diagnostic findings, and histopathology are similar in these children to those findings described in adult patients with GER and Barrett’s esophagus (4-6,16,22). The 13% prevalence of Barrett’s esophagus in this series of children with GER is comparable to the 11% incidence of Barrett’s esophagus reported in adult patients presenting with GER 1221. These children presented with symptoms suggestive of GER and esophagitis: excessive regurgitation during infancy, vomiting, pain, odynophagia, dysphagia, and heartburn. Gastroesophageal reflux was documented by radiographic studies, as well as esophageal pH monitoring, and supported by the finding of decreased sphincter pressure at esophageal manometry. Five of the 13 patients had radiographic evidence of esophageal stricture, but none had a hiatal hernia. Histologic examination of the columnar epithelium in these children revealed the same three types of columnar epithelium previously described in adults by Paul1 et al. (16).Our patients were not subjected to the detailed esophageal mapping described by those authors, but the relationship of the various types of epithelia to the LES appears similar. Gastric fundic epithelium was most distal nearest the LES, and specialized columnar epitheli-
urn was most proximal. The prevalence of specialized columnar epithelium with distinctive intestinallike villi and goblet cells was much lower in this series of children than in adult series of Barrett’s esophagus. The reason for this is unclear. Perhaps a longer time period is required to develop this type of mucosa rather than the other two types. The 2 children demonstrating specialized columnar epithelium were among the oldest patients in the series. We feel part of the difficulty in recognizing Barrett’s esophagus in children is its similar endoscopic appearance to esophagitis. This differs from adults, in whom the distinction between inflamed squamous epithelium and columnar epithelium is usually more readily apparent through the endoscope. In this series of children, the pathologist’s report of columnar epithelium in the esophagus was often a surprise. Barrett’s esophagus may be overlooked in children if biopsy is not performed at the time of endoscopy. We suggest that in these children, the columnarlined esophagus results from GER, just as it does in adults. Each of these children had symptoms of GER for many months or years before their biopsy. Previous studies documenting the transformation of esophageal squamous epithelium to columnar epithelium indicate that the change may be accomplished within 30 mo in human patients with GER (3)and even faster in experimental animals (23,24). Therefore, the age of the children in this series does not preclude transformation from squamous to columnar esophageal mucosa after the onset of reflux. Esophagitis was not well appreciated as a pediatric entity until the last 10 yr, when the introduction of small flexible endoscopes enabled detailed direct
February
1984
gastrointestinal examination of children. Prior to that, when a columnar-lined esophagus was observed in a child, it was thought to represent a congenital anomaly. Barrett (11) originally thought the condition represented a congenitally short esophagus with partial intrathoracic stomach. This assumption may have arisen partly because the embryonic esophagus is lined by columnar epithelium prior to the 25th week of gestational age (25). However, the embryonic epithelium is superficial and ciliated and does not develop glands. Except for being columnar, it bears no resemblance to any of the three types of epithelia seen in Barrett’s esophagus. Another source of confusion regarding columnar epithelium in the esophagus of children is the widely misquoted 1941 autopsy study of Rector and Connerley (26) in which some form of gastric mucosa was found in the esophagus of 7.6% of 1000 sequentially autopsied infants and children between the newborn period and 14 yr of age. Careful reading of this paper and examination of its figures reveals that the authors included as “gastric mucosa” the finding of a few cardiac-type glands in the lamina propria beneath stratified squamous epithelium, or an isolated strip of columnar epithelium on the surface, adjacent to squamous epithelium. In addition, in nearly all of the patients in that study, the abnormality was found in the superior two-thirds of the esophagus. In no instance was gastric mucosa with parietal cells found in the inferior one-third of the esophagus, nearest the LES. The work of Rector and Connerley does not support the view that Barrett’s esophagus in children is a congenital anomaly. We do not mean to imply that columnar epitheliurn is never a congenital anomaly in childhood. Certainly, heterotopic gastric cardiac and fundic mucosa may be found occasionally in the postcricoid region of the esophagus (26,27). Patients with esophageal atresia and tracheoesophageal fistula may have gastric epithelium present at birth in the Such cases, however, upper esophageal pouch (18). are rare, and would not ordinarily be confused with a child presenting with GER and showing Barrett’s epithelium in the distal one-third of the esophagus. References Goldman MC, Beckman RC. Barrett syndrome: case report with discussion about concepts of pathogenesis. Gastroenterology 1960;39:104-10. Adler RH. The lower esophagus lined by columnar epitheliurn. J Thorac Cardiovasc Surg 1963;45:13-31. Mossberg SM. The columnar-lined esophagus (Barrett syndrome)-an ar.quired condition. Gastroenterology 1966; 50:671--K.
BARRETT’S
ESOPHAGIJS
IN CHILDREN
323
4. Burgess JN, Payne WS, Anderson HA, et al. Barrett esophagus: the columnar-lined lower esophagus. Mayo Clin Proc 1971: 46:728-34. lower esopha5. Naef AP, Savary M. Ozzello L. Columnar-lined gus: an acquired lesion with malignant predisposition. Report on 140 cases of Barrett’s esophagus with 12 adenocarcinomas. J Thorac Cardiovasc Surg 1975;70:826-34. 6. Messian RA, Hermos JA, Robbins AH, et al. Barrett’s esophagus: clinical review of 26 cases. Am J Gastroenterol 1978;69: 458-66. 7. Euler AR, Ament ME. Gastroesophageal reflux in children: clinical manifestations, diagnosis, pathophysiology, and therapy. Pediatr Ann 1976;5:678-89. reflux. J Pediatr 1981;98:859-70. 8. Herbst JJ. Gastroesophageal 9. Herbst JJ, Meyers WF. Gastroesophageal reflux in children. Adv Pediatr 1981:28:159-86. 10. Weissbluth M. Gastroesophageal reflux. A review. Clin PI,diatr 1981;20:7-14. 11. Barrett NR. Chronic peptic ulcer of the oesophagus and oesophagitis. Br J Surg 1950;38:175-82. a review of some controversial 12. Barrett NR. Hiatus hernia: points. Br J Surg 1954;42:231-43. AS. The oesophagus lined with gastric 13. Allison PR, Johnstone mucous membrane. Thorax 1953:8:87-101. 14. Borrie J, Goldwater L. Columnar cell-lined esophagus: assessment of etiology and treatment. J Thorac Cardiovasc Surg 1976;71:825-34. of gastroesophageal reflux in 15. Euler AR, Ament ME. Detection the pediatric age patient by intraluminal pH probe measurrment (Tuttle test). Pediatrics 1977;60:65-8. 16. Paul1 A, Trier JS, Dalton MD, et al. The histologic spectrum of Barrett’s esophagus. N Engl J Med 1976;295:476-80. 17. Trier JS. Morphology of the epithelium of the distal esophagus in patients with midesophageal peptic stricture. Gastroenterotogy 1970;58:444-61. 18. Emery JL. Haddadin AJ. Gastric-type epithelium in the upper esophageal pouch in children with TE fistula. J Pediatr Surg 1971:6:449-53. 19. Christie DL. O’Grady LR, Mack DV. Incompetent lower esophageal sphincter and gastroesophageal reflux in recurrent acute pulmonary disease of infancy and childhood J Pediatr 1978: 93:23-7. 20 Euler AR, Byrne WJ, Ament ME, et al. Recurrent pulmonar> disease in children: a complication of gastroesophageal reflux. Pediatrics 1979;63:47-51. 21 Rothstein FC, Halpin TC. High incidence of pulmonary symptoms in infants evaluated for esophageal disease. Ann Otol Rhino1 Laryngol 1980;89:450-3. 22 Goldman H, Antonioli DA. Mucosal biopsy of the esophagus. stomach, and proximal duodenum. Human Path01 1982; 13:423-48. con23. Bremner CG, Lynch VP, Ellis FH. Barrett’s esophagus: genital or acquired? An experimental study of esophageal mucosal regeneration in the dog. Surgery 1970;68:209-16. 24. Wong J. Finckh ES. Heterotopia and ectopia of gastric epithelium by mucosal wounding in the rat. Gastroenterolog! 1971;60:279-87. changes in the oesophageal epi25. Johns BAE. Developmental thelium in man. J Anat 1952:86:431-4%. ML. Aberrant mucosa in the esophagus 26. Rector LE, Connerley in infauts and children. Arch Pathol Lab Med 1941;31:285-94. heterotopias of the alimentary tract. 27. Taylor XL. The epithelial I Pathol Bacterial 1927:30:415-49.
Barrett’s Esophagus in Children: A Consequence of Chronic Gastroesophageal Reflux BEVERLY
BARRETT
DAHMS
and
FRED C. ROTHSTEIN
Departments of Pathology and Pediatrics, Case Western Reserve University and Rainbow Babies’ and Children’s Hospital, Cleveland, Ohio
It is widely acknowledged that Barrett’s esophagus in adults is an acquired condition resulting from prolonged gastroesophageal reflux. Barrett’s esophagus is rare in childhood, even though gastroesophageal rejlux occurs commonly in the pediatric age group. When a columnar-lined esophagus is present in children, it is often regarded as a congenital of chronic anomaly rather than as a consequence gastroesophageal reflux. Over a 5-yr period (1978 1982), we retrospectively studied Barrett’s esophagus in children 19 yr of age or younger who were evaluated for gastroesophageal reflux and whose symptoms warranted esophagoscopy and esophageal biopsy. Esophageal biopsies were performed on 103 patients with gastroesophageal reflux. Thirteen children [age range, 8 mo-19 yr) had Barrett’s esophagus, for a prevalence of 13%. Gastroesophageal reflux was documented in these children by upper gastrointestinal radiographs or pH monitoring. Radiographs demonstrated esophageal stricture in 5 of the 13 children; none had hiatal hernia. Children presented with symptoms suggestive of gastroesophageal reflux and esophagitis: vomiting, abdominal pain, odynophagia, dysphagia, and heartburn. All children had a past history of excessive regurgitation during infancy. Histologically, three types of columnar epithelium were present: gastric fundic type (11 patients), junctional-type columnar epithelium reminiscent of gastric cardia (7 patients), and specialized columnar (metaplastic intestinal) type (2 patients). We believe that Barrett’s esophagus is more common in children than had previously been appreciated. In these children, we
Received March 14. 1983. Accepted August 30, 1983. Address requests for reprints to: Beverly Barrett Dahms, M.D.. Department of Pathology, University Hospitals of Cleveland, 2085 Adelbert Road, Cleveland, Ohio 44106. 0 1984 by the American Gastroenterological Association 0016-5085/84/$3.00
suggest that the distal columnar-lined resulted from chronic gastroesophageal not a congenital anomaly.
esophagus reflux and is
In the past decade, it has become widely acknowledged that Barrett’s esophagus in adults is an acquired condition resulting from prolonged gastroesophageal reflux (GER) (l-6). Reports of Barrett’s esophagus in childhood are rare, even though GER in the pediatric population is common and has been the subject of several recent reviews (7-10). When columnar epithelium has been described in the esophagus of children, there is a tendency to regard it as a congenital heterotopic anomaly, rather than a consequence of chronic GER (11-14). This study was prompted by the recent appearance of a number of children at our hospital with the classic symptomatic, endoscopic, and histologic manifestations of Barrett’s esophagus. Their medical histories were indicative of gastroesophageal reflux from infancy. The purpose of this paper is to document the occurrence of Barrett’s esophagus in children who present with symptomatic gastroesophageal reflux.
Methods We retrospectively reviewed the biopsy specimens of children, 19 yr of age and younger, who were evaluated for GER and whose symptoms warranted esophagoscopy and esophageal biopsy. All children were referred to the gastroenterology service at Rainbow Babies’ and Children’s Hospital over a 5-yr period (1978-1982) for symptoms suggestive of GER. Clinical evaluation included barium swallow and upper gastrointestinal (UGI) series or esophageal pH monitoring (15) to document GER. Not every patient with a diagnosis of GER underwent esophagoscopy. Only those children with GER and symptoms of esophagitis or those who failed traditional medical management for their GER had esophagoscopy and biopsy studies performed. Esophageal manometry was performed
February
BARRETT’S
1984
ESOPHAGCIS IN CHILDREN
319
to assess lower esophageal sphincter function in some children. Endoscopy was performed using an Olympus pediatric endoscope (Olympus Corporation of America, New Hyde Park, N.Y.). All esophageal mucosal biopsy specimens were obtained under direct visualization either by grasp forceps (through the endoscope) or by suction biopsy (Ruben tube) technique. In most cases, specimens were taken at one level per endoscopic session, usually 2-3 cm proximal to the lower esophageal sphincter (LES). This method of endoscopic visualization of the LES is not as reliable as is manometric localization of the sphincter. However, this retrospective series is based on observations made during routine procedures, and esophageal biopsy by manometric guidance is not routine clinical practice. Two or three different pieces of tissue were usually obtained per biopsy. The level at which biopsy specimens were obtained was recorded in each case. In 2 patients (No. 2 and No. 4 in Table l), multiple biopsy specimens were obtained from different levels of the esophagus at one endoscopic session. All biopsy specimens were fixed in Bouin solution, embedded in paraffin, stained with hematoxylin and eosin, and serially sectioned. Biopsy specimens were diagnosed as having Barrett’s esophagus if (a) they were taken at least 2 cm proximal to the LES by direct visualization at endoscopy and [b) they demonstrated the abnormal columnar epithelium previously described (16,17) as characteristic of this condition: i.e., (a) gastric fundic type mucosa, (b) junctional epitheliurn reminiscent of gastric cardia, or (c) specialized columnar epithelium with a villiform surface and goblet cells. In
order to avoid confusion with normal esophageal-cardiac junction in some of the lower biopsy specimens, the presence of junctional epithelium with cardiac-type mucous glands was considered insufficient by itself to include patients in the study unless Barrett’s mucosa was apparent endoscopically. Patients with a history of esophageal atresia and tracheoesophageal fistula at birth were excluded from the series because of the known high incidence of columnar epitheli-
Table
Barrett’s
1. Clinical
und
Pathologic
Findings
Symptoms
in Children
With
(18).
urn in the esophagus in this condition
Results Clinical Features Esophageal biopsies of 103 children and adolescents were performed over the 5-yr period 19781982, during evaluation of gastroesophageal reflux. Thirteen of these children demonstrated the abnormal columnar epithelia that have come to be recognized as Barrett’s esophagus, for a prevalence of 13%. Patients are tabulated in Table 1. The 13 children with Barrett’s esophagus were 8 mo to 19 yr of age and presented with symptoms of GER or peptic esophagitis. All children had a history of excessive regurgitation during infancy. Eight children presented with a history of acute emesis. Two children were failing to thrive because of caloric losses secondary to chronic vomiting. Two children presented with acute hematemesis, and at least 5 had Esophagus
Radiology
Esophagoscopy
Histology
-I: ._
.z
2
2
+
ND
+
ND
E ; ++ ++
Yes Yes
4
-
2
*
f
5
+
+ +
11 3
14
+
+
+
+
ND
ND
4
15
t
._
+
+
+
ND
5
2.92
*
+
+
+
12.0
+
+
_
+
9.5
_
+
ND
+
ND
+
11.0
+
22.0
++ ++
NO Yes
+t
No
if + -+ ++ + + ++ ++
NO
4
*
2
f
4
t
a 3
7
+
4
6
12
7
6
I
a
7
_
9
3 ..5
i-
+
_
10 11
6 19
+ +
_ +
+ _
-
ND
ND
12
4
+
+
_
-
ND
ND
13
0.67
+
+
+
-
+
Abbreviations:
+
7.0
+ = present; - = absent; GER = gastroesophageal reflux; ND = not done; ” Location gitis + = erythema, friability; esophagitis ++ = ulcers, exudate.
No
NO No Yes Yes
4 3 3 3 4
+
+ -+ + _
4
t + + + +
NO
3
-
No
3
+
LES = lower esophageal = number of centimeters
+ +
+ + + + -t
sphincter; esophaproximal
to LES.
320
DAHMS
AND
GASTKOENTEROI,(~(;‘r’
ROTHSTEIN
had hematemesis in the past. Nine children had Hemoccult-positive stools at the time of initial evaluation. It was difficult to document the presence of certain symptoms (i.e., heartburn and odynophagia) in some children because of either their young age (3 pa(1 patient). tients] or their inability to communicate Heartburn was the most common complaint in the older children (9 patients) and this occurred postprandially, while supine and with exercise. Heartburn was associated with an increased frequency of burping in several children. Nine of the older children had complaints of odynophagia-for liquids (2 patients) and solids (4 patients). The younger children experienced chest and epigastric discomfort. Pain was often severe enough to awaken the child from sleep. Six children had dysphagia. In the cases of the younger children, mothers would complain that their child would prefer liquids to solids and often cough and choke when solids were ingested. Several children had pulmonary complaints that may have been secondary to GER (19-2 1). Five had a history of asthma, 4 had night cough of >3 months’ duration, and 3 had radiographically demonstrated pneumonia on at least two occasions. Upper gastrointestinal radiographs were performed on all children and GER was demonstrated in 9. Five patients had radiographic evidence of either distal or midesophageal strictures, and these patients also had esophageal ulcerations demonstrated. No child had radiographic evidence of a hiatal hernia or gastric outlet obstruction.
Vol. 86.
No.
2
On 8 children, including those 4 with negative UGI studies, pH monitoring was performed. All 8 had GER. Esophageal manometry was performed in 7 children. Six had decreased LES pressures. Two had of a very short LES (SO.5 cm). No specific pattern motor abnormalities of the esophagus was apparent. At esophagoscopy, strictures, located 5-10 cm above the LES, were noted in 5 children. Esophagitis was present in all children, varying in severity from erythema and moderate friability (3 children] to purulent exudates and ulcerations (10children). The presence of columnar epithelium, as evidenced by a salmon-pink color of esophageal mucosa, was apparent to the endoscopist in only 5 children.
Pathology As indicated in Table 1, three histologic types of columnar epithelium were demonstrated in the distal one-third of the esophagus, just proximal to the LES. Abnormal gastric fundic epithelium, prewas the most commonly obsent in 11 patients, served columnar epithelium (Figure 1); in some patients, it was the only columnar epithelium obtained from the esophagus. Characteristically, this mucosa appeared atrophic compared with normal gastric fundus. It was similar to fundic mucosa in being covered by uniform tall columnar cells and containing coiled tubular glands with parietal and chief cells. Seven patients demonstrated the second type of columnar epithelium seen in Barrett’s esophagus, the junctional type reminiscent of gastric cardia.
February
1984
BAKRETT’S
Figure
ESOPHA(:IJS
IN CHILDKEN
32 1
2. This junctional-type epithelium is reminiscent of gastric cardia, with a tall columnar epithelial surface and mucous glands. Note inflammation of the lamina propria. This biops) specimen was obtained from a 2.5. yr-old girl and was taken 4 cm proximal to the lower esophageal sphincter. (H&E. * 120.)
There was considerable variation in the appearance of junctional epithelium from patient to patient. In some, the mucosa closely resembled gastric cardia [Figure 2) with uniform tall columnar surface epithelial cells and a wide glandular layer composed of cardiac-type glands. In others, the surface was more villiform and the glands atrophic and distorted (Figure 3). One patient demonstrated a mixture of cardiac and fundic glands. In all cases, the junctional epithelium exhibited sufficient atrophy, inflammation, or distortion that it was unlikely to be confused with normal gastric cardiac mucosa. The majority of the 7 patients with junctional epithelium demonstrated one of the other types of columnar epitheliurn at other levels of the esophagus. Only 2 patients demonstrated the third type of epithelium, the distinctive specialized columnar epithelium pathognomonic of Barrett’s esophagus (Figure 4). This columnar epithelium has also been called the intestinal metaplastic type because of its villiform surface, goblet cells, and mucosal glands. In addition to these three types of columnar epithelium, squamous mucosa was also obtained from 10 patients. Inflammation was noted in all biopsy specimens of squamous mucosa and consisted of a scattering of neutrophils in lamina propria and epithelium. Eosinophils were present in three biopsy specimens. In several instances, an abnormal squamocolumnar junction was demonstrated, such as the junction between squamous and specialized columnar epithelium in Figure 4. Patient 13 demonstrated continuity of squamous mucosa with gastric fundictype epithelium. None of the columnar epithelia in these children demonstrated dysplasia, although some showed mild nuclear changes consistent with inflammation and repair.
Patients No. 2 and No. 4 had multiple biopsy specimens of the esophagus taken at different levels during one endoscopy session. In each case, gastric fundic mucosa lined the most distal portion of the
This variation of junctional-type epithelium has some features of both gastric and intestinal mucosa. The surface is \rilliform, but is composed of tall columnar epithelium devoid ot goblet cells. Mucous glands are reminiscent of cardiac glands, but are markedly atrophic. The lamina propria is severely inflamed Biopsy specimen is from patient No. 6. (H&E, x100.)
322
Figure
DAHMS
AND
ROTHSTEIN
GASTROENTEROLOGY
Vol.
86, No.
2
4. This photomicrograph shows the specialized columnar type of Barrett’s epithelium, reminiscent of intestinal mucosa because of the villiform surface and intestinallike epithelium containing goblet cells. Note the squamocolumnar junction, upper left. Biopsy specimen is from patient No. 4. (H&E. x60.)
esophagus, nearest the LES. Specialized columnar epithelium was the most proximal type of columnar epithelium, biopsied 5 and 8 cm above the LES, respectively.
Discussion Barrett’s esophagus is more common in children than previously appreciated. The clinical presentation, diagnostic findings, and histopathology are similar in these children to those findings described in adult patients with GER and Barrett’s esophagus (4-6,16,22). The 13% prevalence of Barrett’s esophagus in this series of children with GER is comparable to the 11% incidence of Barrett’s esophagus reported in adult patients presenting with GER 1221. These children presented with symptoms suggestive of GER and esophagitis: excessive regurgitation during infancy, vomiting, pain, odynophagia, dysphagia, and heartburn. Gastroesophageal reflux was documented by radiographic studies, as well as esophageal pH monitoring, and supported by the finding of decreased sphincter pressure at esophageal manometry. Five of the 13 patients had radiographic evidence of esophageal stricture, but none had a hiatal hernia. Histologic examination of the columnar epithelium in these children revealed the same three types of columnar epithelium previously described in adults by Paul1 et al. (16).Our patients were not subjected to the detailed esophageal mapping described by those authors, but the relationship of the various types of epithelia to the LES appears similar. Gastric fundic epithelium was most distal nearest the LES, and specialized columnar epitheli-
urn was most proximal. The prevalence of specialized columnar epithelium with distinctive intestinallike villi and goblet cells was much lower in this series of children than in adult series of Barrett’s esophagus. The reason for this is unclear. Perhaps a longer time period is required to develop this type of mucosa rather than the other two types. The 2 children demonstrating specialized columnar epithelium were among the oldest patients in the series. We feel part of the difficulty in recognizing Barrett’s esophagus in children is its similar endoscopic appearance to esophagitis. This differs from adults, in whom the distinction between inflamed squamous epithelium and columnar epithelium is usually more readily apparent through the endoscope. In this series of children, the pathologist’s report of columnar epithelium in the esophagus was often a surprise. Barrett’s esophagus may be overlooked in children if biopsy is not performed at the time of endoscopy. We suggest that in these children, the columnarlined esophagus results from GER, just as it does in adults. Each of these children had symptoms of GER for many months or years before their biopsy. Previous studies documenting the transformation of esophageal squamous epithelium to columnar epithelium indicate that the change may be accomplished within 30 mo in human patients with GER (3)and even faster in experimental animals (23,24). Therefore, the age of the children in this series does not preclude transformation from squamous to columnar esophageal mucosa after the onset of reflux. Esophagitis was not well appreciated as a pediatric entity until the last 10 yr, when the introduction of small flexible endoscopes enabled detailed direct
February
1984
gastrointestinal examination of children. Prior to that, when a columnar-lined esophagus was observed in a child, it was thought to represent a congenital anomaly. Barrett (11) originally thought the condition represented a congenitally short esophagus with partial intrathoracic stomach. This assumption may have arisen partly because the embryonic esophagus is lined by columnar epithelium prior to the 25th week of gestational age (25). However, the embryonic epithelium is superficial and ciliated and does not develop glands. Except for being columnar, it bears no resemblance to any of the three types of epithelia seen in Barrett’s esophagus. Another source of confusion regarding columnar epithelium in the esophagus of children is the widely misquoted 1941 autopsy study of Rector and Connerley (26) in which some form of gastric mucosa was found in the esophagus of 7.6% of 1000 sequentially autopsied infants and children between the newborn period and 14 yr of age. Careful reading of this paper and examination of its figures reveals that the authors included as “gastric mucosa” the finding of a few cardiac-type glands in the lamina propria beneath stratified squamous epithelium, or an isolated strip of columnar epithelium on the surface, adjacent to squamous epithelium. In addition, in nearly all of the patients in that study, the abnormality was found in the superior two-thirds of the esophagus. In no instance was gastric mucosa with parietal cells found in the inferior one-third of the esophagus, nearest the LES. The work of Rector and Connerley does not support the view that Barrett’s esophagus in children is a congenital anomaly. We do not mean to imply that columnar epitheliurn is never a congenital anomaly in childhood. Certainly, heterotopic gastric cardiac and fundic mucosa may be found occasionally in the postcricoid region of the esophagus (26,27). Patients with esophageal atresia and tracheoesophageal fistula may have gastric epithelium present at birth in the Such cases, however, upper esophageal pouch (18). are rare, and would not ordinarily be confused with a child presenting with GER and showing Barrett’s epithelium in the distal one-third of the esophagus. References Goldman MC, Beckman RC. Barrett syndrome: case report with discussion about concepts of pathogenesis. Gastroenterology 1960;39:104-10. Adler RH. The lower esophagus lined by columnar epitheliurn. J Thorac Cardiovasc Surg 1963;45:13-31. Mossberg SM. The columnar-lined esophagus (Barrett syndrome)-an ar.quired condition. Gastroenterology 1966; 50:671--K.
BARRETT’S
ESOPHAGIJS
IN CHILDREN
323
4. Burgess JN, Payne WS, Anderson HA, et al. Barrett esophagus: the columnar-lined lower esophagus. Mayo Clin Proc 1971: 46:728-34. lower esopha5. Naef AP, Savary M. Ozzello L. Columnar-lined gus: an acquired lesion with malignant predisposition. Report on 140 cases of Barrett’s esophagus with 12 adenocarcinomas. J Thorac Cardiovasc Surg 1975;70:826-34. 6. Messian RA, Hermos JA, Robbins AH, et al. Barrett’s esophagus: clinical review of 26 cases. Am J Gastroenterol 1978;69: 458-66. 7. Euler AR, Ament ME. Gastroesophageal reflux in children: clinical manifestations, diagnosis, pathophysiology, and therapy. Pediatr Ann 1976;5:678-89. reflux. J Pediatr 1981;98:859-70. 8. Herbst JJ. Gastroesophageal 9. Herbst JJ, Meyers WF. Gastroesophageal reflux in children. Adv Pediatr 1981:28:159-86. 10. Weissbluth M. Gastroesophageal reflux. A review. Clin PI,diatr 1981;20:7-14. 11. Barrett NR. Chronic peptic ulcer of the oesophagus and oesophagitis. Br J Surg 1950;38:175-82. a review of some controversial 12. Barrett NR. Hiatus hernia: points. Br J Surg 1954;42:231-43. AS. The oesophagus lined with gastric 13. Allison PR, Johnstone mucous membrane. Thorax 1953:8:87-101. 14. Borrie J, Goldwater L. Columnar cell-lined esophagus: assessment of etiology and treatment. J Thorac Cardiovasc Surg 1976;71:825-34. of gastroesophageal reflux in 15. Euler AR, Ament ME. Detection the pediatric age patient by intraluminal pH probe measurrment (Tuttle test). Pediatrics 1977;60:65-8. 16. Paul1 A, Trier JS, Dalton MD, et al. The histologic spectrum of Barrett’s esophagus. N Engl J Med 1976;295:476-80. 17. Trier JS. Morphology of the epithelium of the distal esophagus in patients with midesophageal peptic stricture. Gastroenterotogy 1970;58:444-61. 18. Emery JL. Haddadin AJ. Gastric-type epithelium in the upper esophageal pouch in children with TE fistula. J Pediatr Surg 1971:6:449-53. 19. Christie DL. O’Grady LR, Mack DV. Incompetent lower esophageal sphincter and gastroesophageal reflux in recurrent acute pulmonary disease of infancy and childhood J Pediatr 1978: 93:23-7. 20 Euler AR, Byrne WJ, Ament ME, et al. Recurrent pulmonar> disease in children: a complication of gastroesophageal reflux. Pediatrics 1979;63:47-51. 21 Rothstein FC, Halpin TC. High incidence of pulmonary symptoms in infants evaluated for esophageal disease. Ann Otol Rhino1 Laryngol 1980;89:450-3. 22 Goldman H, Antonioli DA. Mucosal biopsy of the esophagus. stomach, and proximal duodenum. Human Path01 1982; 13:423-48. con23. Bremner CG, Lynch VP, Ellis FH. Barrett’s esophagus: genital or acquired? An experimental study of esophageal mucosal regeneration in the dog. Surgery 1970;68:209-16. 24. Wong J. Finckh ES. Heterotopia and ectopia of gastric epithelium by mucosal wounding in the rat. Gastroenterolog! 1971;60:279-87. changes in the oesophageal epi25. Johns BAE. Developmental thelium in man. J Anat 1952:86:431-4%. ML. Aberrant mucosa in the esophagus 26. Rector LE, Connerley in infauts and children. Arch Pathol Lab Med 1941;31:285-94. heterotopias of the alimentary tract. 27. Taylor XL. The epithelial I Pathol Bacterial 1927:30:415-49.