What are the defining indices of the esophagogastric junction and the defining criteria for columnar lined esophagus (CLE)?
H.W. Boyce Jr. (Tampa)
Anatomic features of the normal esophagogastric junction
The esophagus in the average adult passes through the diaphragmatic hiatus at approximately 38 cm and joins the stomach near the squamocolumnar junction at about the 40 cm level. The normal level of the squamocolumnar mucosal junction may vary 1 to 2 cm depending on the body habitus, type of endoscope used and the care with which measurements are made. More accurate distance measurements are made on endoscope withdrawal than during insertion because the instrument is in a more straightened position. If minimal air insufflation of the proximal and mid-esophagus is utilized during endoscopy, the closed lower esophageal sphincter (LES) region may be readily demonstrated. At the point of closure of the proximal end of the sphincter, several (usually four to six) longitudinal symmetrical mucosal folds can be seen to disappear in the center of the closed lumen. This closure produces a rosette appearance with the lumen being precisely centered at the point where these longitudinal folds converge. The tone of the LES relaxes with primary or secondary peristalsis and also opens in response to gentle insufflation. As the closed normal esophageal sphincter is approached with the endoscope it will relax with gentle scope pressure and with passage through the sphincter there is little or no detectable resistance. As the high pressure sphincter zone relaxes, one can identify the
squamocolumnar mucosal junction 1 to 2 cm beyond and see into the short tubular cavity of the proximal stomach, the cardia. The proximal end of the LES is most easily demonstrated in patients with achalasia because of the dilation of the body of the esophagus and typical hypertonicity of the closed sphincter [1,2].
The level of the hiatal margin is not as readily demonstrated in normal patients as in those who have a hiatal hernia. However, it is possible in most instances to determine the level with relative precision. As the lumen is gently insufflated with air, the patient is asked either to sniff or inhale rapidly, at which time the diaphragmatic hiatal margin moves inferiorly either quickly or gradually depending upon the breathing maneuver used to demonstrate its location. The wall of the esophagogastric junction is slightly indented as the hiatal margin moves distally with a sniff or deep breath.
The squamous mucosa of the esophagus is pearly pink or pinkish-grey in color and contrasts sharply with the orange-red color of the gastric columnar epithelium. The esophageal mucosa is only slightly transparent and reflects light moderately. The gastric mucosa has a glistening surface because of the presence of mucus, but is more transparent and consequently absorbs a great deal of light. For this reason gastric mucosa requires more light for adequate photography.
The junction of the squamous and columnar epithelium appears after minimum inflation as a slightly irregular or undulating line, the so-called ora serrata or "Z" line. With continued inflation pressure, the irregular junction often becomes straight with only minimal serration. This junction normally is located in the distal half of the LES segment at or just below the esophageal diaphragmatic hiatus. This line of demarcation between the two types of mucosa is readily identifiable in the absence of pathologic changes. If there is uncertainty about the location of this mucosal junction, it can be dramatically demonstrated by application of several milliliters of 50% Lugol's solution through an endoscopic catheter . This will stain the glycogen in the esophageal mucosa in about 30 s. In addition to surface characteristics and color, the normal distal extent of the esophageal squamous epithelium is also clearly demarcated by the level of abrupt disappearance of multiple, linear, frequently branching, small blood vessels just proximal to the upper end of the gastric folds of the cardia at the junction of squamous and columnar epithelium.
After the endoscope is passed into the proximal stomach, a retroversion maneuver should be performed to view the cardia and fundus from below. In the normal setting the insertion tube of the endoscope can be seen coming through a snugly fitting intraabdominal segment of esophagus. The snug fit in this region is sustained throughout respiration and during moderate insufflation of the stomach, except that transient relaxation in response to primary or secondary peristalsis can be detected. The classic snug appearance of the region is nearly always demonstrated in patients with achalasia because of the increased tone in the LES segment. The angle of His on the greater curvature aspect demarcates the distal end of the sphincter region, also called the submerged or abdominal segment of the esophagus. The normally located squamocolumnar mucosal junction cannot be identified radiographically. However, in patients with herniation of the proximal stomach through the diaphragmatic hiatus, a lower esophageal ring may be demonstrated, the location of which corresponds with
the level of the intrathoracically displaced squamocolumnar mucosal junction.
Earlier radiographic and anatomic studies have caused much confusion about this area relative to the structures that are seen in normal patients and in patients with hiatal hernias. In an individual with normal anatomy there are no rings, no asymmetrical bulges, no bulbous contour of the distal esophagus, and no displacement of the squamocolumnar mucosal junction greater than 2 cm cephalad to the diaphragmatic hiatus. These alterations are seen only in patients with a hiatal hernia [1,2].
Several linear gastric folds normally are seen in the cardia, especially when a hiatal hernia is present. These folds normally terminate within 0.5 to 1.0 cm of the normal location of the squamocolumnar mucosal junction. Therefore, this termination point of the folds can be utilized endoscopically and radiographically as a marker for the approximate location of the normal squamocolumnar mucosal junction. The cephalad margins of these longitudinal gastric folds correspond to the level of the esophagogastric muscular junction as well. The proximal margins of these folds provide the best endoscopic landmark for the muscular junction between esophagus and stomach and as a marker for the normal location of the squamocolumnar junction. These relationships to the esophagogastric muscular junction can be demonstrated on surgical and autopsy specimens as well.
In some patients with a hiatal hernia and normal LES tone, the esophageal wall in the region of the esophageal sphincter will be closed snugly around the endoscope. Some relaxation may be apparent in relation to primary and secondary peristalsis or after greater degrees of air insufflation. In patients with reflux esophagitis, especially those with a columnar lined esophagus (CLE) who tend to have the lowest sphincter pressures, there is considerable free space around the endoscope in the region of the LES just proximal to the hernia pouch. Patients with a CLE nearly always have a widely patent sphincter region on antegrade and retrograde endoscopy.
When viewing the region of the squamocolumnar junction by retroversion in patients with hiatal hernia and normal sphincter pressure, the closure of the proximal end of the LES can be observed. The point of maximum closure in these cases appears about 1 cm above the squamocolumnar mucosal junction. This level of closure corresponds with the so-called esophageal A-ring or sphincter contraction ring, both in location and contour seen during radiography and antegrade endoscopy.
With a hiatal hernia of moderate or larger size the retroverted endoscope can be pulled back to the level of the diaphragmatic hiatus, or even a short distance into the hernia pouch, thus affording a close-up examination of the hernia pouch and the squamocolumnar mucosal junction from below.
It is important to observe carefully and record the characteristics of the distal esophagus and proximal stomach in all patients. The location of the diaphragmatic hiatus in relation to the proximal stomach, the level of the squamocolumnar mucosal junction and the proximal extent of the gastric mucosal folds in the hernia pouch are characteristics utilized in the precise endoscopic diagnosis of hiatal hernia and reflux sequelae, including the earliest stages of CLE. The levels of these landmarks should be recorded on every esophagoscopy report.
Endoscopic anatomy of the columnar-lined (Barrett's) esophagus
In the past few years so-called Barrett's or columnar lined esophagus (CLE) has been recognized as more common than previously believed [4-7]. Since metaplastic epithelium in CLE is considered premalignant, its early recognition is imperative. It is possible to recognize minimal degrees of columnar lined esophagus based on an understanding of the normal lumenal topographic relationships discussed above. Short segments of CLE appear to have premalignant potential similar to larger segments extending to the proximal esophagus.
CLE rarely, if ever, occurs without both a markedly hypotensive LES and a hiatal hernia. The recognition of this relationship is important for accurate diagnosis. Mucosal biopsies taken from a hernia pouch have resulted in overdiagnosis and confusion in surveillance reports. These problems result from the failure of the endoscopist to understand normal endoscopic anatomy, especially recognition of small sliding hiatus hernias.
The endoscopic features that aid in the diagnosis of CLE include: the location of the cephalad margin of the linear gastric mucosal folds in the ever-present hiatal hernia pouch, the displacement of the squamocolumnar junction greater than 1.0 cm cephalad from these folds, the level of disappearance of the linear esophageal vessels, the patulous LES region and the absence of a lower esophageal Schatzki or "B" ring
Identification of the squamocolumnar mucosal junction in normal patients and in those with the typical CLE is not usually difficult because of differences in mucosal color and texture. The mucosal color in the columnar lined segment is orange-red to red; the color usually is more reddish than normal gastric mucosa and is distinctly different from the pinkish-grey color of normal esophageal squamous mucosa. It may be impossible to precisely locate the squamocolumnar junction in the presence of severe esophagitis or a stricture. Close-up observation of the texture of metaplastic intestinalized columnar mucosa will reveal a pitted or villoid pattern. Several biopsies from areas of suspected columnar metaplasia usually will provide histologic proof of the diagnosis. The exact location of biopsies and the relation of this site to surrounding landmarks should be recorded when the samples are obtained. Normal squamous esophageal mucosa is characterized by a pink-grey smooth surface and many underlying small vessels oriented parallel to the long axis of the lumen. These small vessels disappear at the site of the normal muscular junction between esophagus and stomach, and are not observed when severe inflammation or neoplasm are present. The level where these vessels disappear is another reasonably accurate endoscopic landmark for the level at which the true esophagogastric muscular junction should lie. When the location of the junction remains uncertain, several milliliters of Lugol's solution can be instilled to stain squamous epithelium a brownish-black . Islands of residual squamous epithelium in the columnar-lined segment also are readily identified by this stain.
The squamocolumnar mucosal junction in CLE can vary widely in location and appearance. It often has an undulating appearance with orange-red, finger-like cephalad extensions of metaplastic epithelium between peninsulas of squamous
mucosa, the whole junction appearing asymmetrical and interrupted. There is a common tendency to attribute this appearance to esophagitis with erosion alone, when in some cases the orange-red proximally directed streaks are due to metaplastic columnar epithelium at the site of previous healed linear erosions. Areas of metaplastic columnar epithelium rarely contain adherent surface exudate as is typical for erosions. In some patients, islands of metaplastic columnar epithelium surrounded completely by squamous epithelium may be found proximal to the level of the main squamocolumnar junction.
When the squamocolumnar junction is displaced proximally 2 cm or more above the esophageal hiatus, a hiatal hernia is present and one can readily recognize the gastric folds of the gastric cardia that extend through the hiatus. These normal folds begin at a point 0.5 to 1 cm distal to the normal position of the squamocolumnar junction. Excessive air inflation usually obliterates these folds so that their cephalad margins appear to be more distally located than they really are. Since a measurement of 1 cm above the proximal end of the gastric folds indicates the expected site of a normally positioned squamocolumnar mucosal junction in patients with a hiatal hernia, it is easy to determine whether the junction is displaced more proximally as seen with CLE [1,2].
Others believe that CLE should be diagnosed only when the squamocolumnar mucosal junction is located more than 2 cm [8,9] or 3 cm [10,11] proximal to the cephalad margin of the gastric folds. This diagnosis can only be suspected by endoscopic examination, and biopsy confirmation of the presence of specialized epithelium (intestinal metaplasia) is required. Whether one uses the 1 cm, the 2 cm or the 3 cm distance of squamocolumnar mucosal junction location, biopsy proof of specialized columnar epithelium in this segment is required. It seems that a presumptive endoscopic diagnosis and a nondiagnostic biopsy using the 1 cm distance criterion is better than an unsuspected endoscopic diagnosis using the 2 or 3 cm distance criteria for separation of gastric fold margins from mucosal junction. The latter patient would never have had the potential benefit of surveillance. In general, the agreement between endoscopy and histology for diagnosis of CLE has been good. The sensitivity, specificity and accuracy of endoscopy have been shown to be 89%, 93% and 91% respectively .
With the use of these endoscopic definitions and landmarks, minimal degrees of metaplastic epithelial change in the distal esophagus can be suspected and confirmed by biopsy. It is, therefore, not necessary to await development of the classic later stages of CLE with squamocolumnar junction dislocation into midesophagus in order to diagnose this condition. CLE of short length often exists in patients with minimal symptoms and will not be diagnosed in such cases if considered only in relation to severe reflux symptoms, proximal esophageal stricture, an esophageal ulcer or a long segment of columnar-lined esophagus.
The endoscopic landmarks and topography described above should be used to avoid missing the diagnosis of CLE. Overdiagnosis occurs when the mucosa in a hiatal hernia pouch is interpreted as being in the tubular esophagus or the presence of an inlet patch in the cervical esophagus is interpreted as CLE. An adequate number of biopsy specimens (preferably with large forceps) and precise histologic interpreta-
tion usually prevents misdiagnosis. Underdiagnosis occurs with failure to recognize short segments of columnar mucosa lining the distal esophagus or failure to recognize the lining of most of the esophagus by columnar epithelium, when there is a markedly dislocated squamocolumnar junction as far proximal as the cervical esophagus. This error is more likely when proximal esophagitis and/or stricture are absent and the endoscopist is careless and fails to examine every millimeter of the esophageal mucosa with care.
4. Naef AP, Savary M, Ozzello L. Columnar-lined lower esophagus: an acquired lesion with malignant predisposition. Report on 140 cases of Barrett's esophagus with 12 adenocarcinomas. J Thorac Cardiovasc Surg 1975:70:826-835.
6. Herlihy KJ, Orlando RC, Bryson JC, Bozymski EM, Carney CM, Powell DW. Barrett's esophagus: clinical, endoscopic, histologic, manometric and electrical potential difference characteristics. Gastroenterology 1984:86:436-443.
12. Woolf GM, Riddell RH, Irvine EJ, Hunt RH. A study to examine agreement between endoscopy and histology for the diagnosis of columnar-lined (Barrett's) esophagus. Gastrointest Endosc 1989;35:541-544.