Definition of Histopathologic Changes in Gastroesophageal Reflux Disease

Abstract

A series of 71 patients with multiple measured biopsies of the gastroesophageal junctional region permitting assessment of the presence and length of different glandular epithelial types is presented. All but nine of 53 patients in whom a 24-hour pH study was performed had abnormal reflux, suggesting that endoscopic recognition of an abnormal columnar mucosa at the gastroesophageal junction sufficient to precipitate multiple-level biopsies indicates a high probability of abnormal reflux. All patients had cardiac mucosa (CM) or oxyntocardiac mucosa (OCM). CM was present in 68 of 71 patients. The prevalence of intestinal metaplasia increased with increasing CM+OCM length, and was present in all 22 patients with a CM+OCM length >2 cm and in 20 of 49 patients with a CM+OCM length <2 cm. Patients with a CM+OCM length >2 cm had a markedly higher acid exposure than patients with a CM+OCM length <2 cm. The findings suggest that the presence of CM and OCM in the junctional region are predictive of abnormal acid exposure, and that increasing OCM+CM length correlates strongly with the amount of acid exposure. The histologic finding of CM and OCM represents a sensitive histologic criterion for gastroesophageal reflux rather than normal epithelia. These diagnostic criteria represent the first useful histologic definitions for assessing the presence and severity of reflux.

The occurrence of glandular mucosa in the lower esophagus has been recognized since the early 20th century. 1,2,7,13,19,25 After some controversy about whether this was a tubular stomach secondary to a congenitally short esophagus, 4,9 Barrett, 3 in 1957, designated this as the “columnar lined esophagus,” which bears his name. During the next several years, evidence was produced that Barrett's esophagus was an acquired condition resulting from gastroesophageal reflux disease, a concept that is now accepted universally. 6,20

In 1976, Paull et al. 22 described three types of glandular epithelium and established a histologic classification of Barrett's esophagus into junctional, fundic, and specialized types (Table 1). When it was later recognized that Barrett's adenocarcinoma arose only in the specialized columnar epithelium, 11,23 the definition of Barrett's esophagus was restricted to those cases that showed intestinal metaplasia.

Table 1 Image ToolsThe diagnosis of Barrett's esophagus is problematic because of the existing belief that columnar epithelium distinct from gastric oxyntic mucosa is normally present in the junctional region. This mucosa, called junctional and cardiac, consists of gastric-type surface epithelium and glands composed entirely of mucous cells. The normal extent of this cardiac mucosa (CM) has never been defined. The most influential of the definitions of the extent of normal CM is that of Hayward, 12 who asserted, without supporting data, that this mucosa normally lined the lower 2 cm of the tubular esophagus and extended into the proximal stomach.

Glandular mucosa in the lower esophagus is therefore considered abnormal only when it exceeds 2 cm. When >2 cm of glandular mucosa in the lower esophagus is shown to contain intestinal metaplasia, the diagnosis of long-segment Barrett's esophagus is made. This is the only definition of Barrett's esophagus that is accepted universally. The term “columnar lined esophagus of uncertain significance” has been suggested when there is no intestinal metaplasia in a glandular mucosa >2 cm. 24

Intestinal metaplasia has also been shown to occur frequently in patients who have glandular mucosa that is within Hayward's definition of normalcy. 12 When this occurs in the lower 2 cm of the tubular esophagus, the term “short-segment Barrett's esophagus” is used. When it occurs in the proximal stomach adjacent to the esophageal opening, the term “intestinal metaplasia of the gastric cardia” is used. The finding of glandular mucosa without intestinal metaplasia in the distal 2 cm of the tubular esophagus is currently regarded as normal.

These definitions are greatly dependent on endoscopic findings that make measurements from the gastroesophageal junction. It has been shown that endoscopic determination of the gastroesophageal junction is associated with considerable error when compared with manometry, 17 compromising further the veracity of these definitions.

Growing epidemiologic evidence suggests that adenocarcinoma arising in the junctional region and gastric cardia are identical to Barrett's adenocarcinoma of the lower esophagus. 5 A recent study also found an association between reflux and all these types of carcinomas. 18 The fact that proximal gastric cancer is associated with gastroesophageal reflux appears to confound all logic.

This study was undertaken to evaluate the relationship between the various histologic types of mucosae found in the junctional region and reflux, and to develop a hypothesis relating to the sequence of pathologic changes in reflux.