Gastric Cardia Intestinal Metaplasia: Biopsy Follow-Up of 85 Patients Neal S
Total Page:16
File Type:pdf, Size:1020Kb
Gastric Cardia Intestinal Metaplasia: Biopsy Follow-Up of 85 Patients Neal S. Goldstein, M.D. Department of Anatomic Pathology, William Beaumont Hospital, Royal Oak, Michigan flammation may allow the metaplastic mucosa to Background: Gastric cardia intestinal metaplasia revert to normal. (CIM), denoted by goblet cells is common. The fre- quency of persistent CIM is unknown. Methods: 85 KEY WORDS: Barrett’s esophagus, Cardia, Dyspla- patients with CIM and follow-up endoscopies were sia, Esophagus, Inflammation, Intestinal prospectively identified during the time period of metaplasia. 10/6/94–12/21/97. The presence of goblet cells was Mod Pathol 2000;13(10):1072–1079 the defining feature of CIM, other metaplastic cell types were not evaluated. All 85 patients initially Gastric cardia intestinal metaplasia (CIM), defined had biopsies that straddled the squamocolumnar as goblet cells within cardia mucosa, is a recently junction (SCJ) showed CIM, an otherwise normal described pathologic process that occurs in 6 to proximal stomach, lower esophagus, and squamo- 57% of patients with normal squamocolumnar columnar junction. The SCJ lay within the 2 cm of junction regions (1–24). Although the columnar ep- mucosa immediately proximal to the uppermost ithelium of Barrett’s esophagus is thought not to gastric fold and overlaid the junction of the tubular significantly regress, it is unknown whether CIM esophagus and the saccular dilatation of the stom- represents a permanent or reversible metaplasia. ach in all patients. The patients underwent endos- Cardia intestinal metaplasia may have the potential copy for many reasons. They were randomly iden- to undergo metaplasia back to its original nonin- tified based on the absence of a hiatal hernia and testinalized epithelium. the presence of CIM. Results: Ten of the 85 patients This study examines histology of follow-up cardia had CIM on repeat biopsy. Among patients with no biopsies of 85 patients who had CIM on initial en- CIM in the first repeat endoscopy, the degree of doscopy to address these questions. cardia inflammation decreased between the initial and first repeat endoscopy, whereas there was no MATERIALS AND METHODS change in the amount of inflammation among pa- tients who had CIM in the first repeat endoscopy. Eighty-five patients with normal gastroesopha- The changes in mean inflammation score was sig- geal anatomy and a gastric cardia biopsy that con- -tained intestinal metaplasia, defined as the pres ؍ nificantly different between the two groups (P .024). Twenty-two patients underwent a second re- ence of goblet cells, were prospectively identified peat endoscopy and five had a third repeat endos- during the time period of October 6, 1994 to De- copy. Including all follow-up biopsies, six of the 85 cember 21, 1997 from the file of the William Beau- patients (7%) had CIM. Four patients who did not mont Hospital Surgical Pathology Department. have CIM on initial repeat endoscopy had CIM on Only goblet cells were evaluated; no other meta- their second repeat endoscopy, probably reflecting plastic cell types that have been described in Bar- sampling issues. None of the biopsies had dysplasia. rett’s esophagus were studied. All patients under- Conclusions: Cardia inflammation is a stimulus for went upper endoscopy for upper gastrointestinal cardia intestinal metaplasia, and a reduction in in- symptoms, had endoscopically normal gastro- esophageal anatomy, and had biopsies of the an- trum, cardia, and lower esophageal squamous mu- Copyright © 2000 by The United States and Canadian Academy of cosa. All patients came from the practices of eight Pathology, Inc. gastroenterologists. Patients who had prior gastric VOL. 13, NO. 10, P. 1072, 2000 Printed in the U.S.A. Date of acceptance: May 11, 2000. or esophageal surgery were excluded. Address reprint requests to: Neal Goldstein, M.D., Department of Ana- The anatomy of the proximal stomach and lower tomic Pathology, William Beaumont Hospital, 3601 West Thirteen Mile Road, Royal Oak, MI 48073; e-mail: [email protected]; fax: 248- esophagus was extracted from endoscopy reports 551-9054. and discussions with gastroenterologists, who, if 1072 necessary, reviewed the endoscopic photographs of Amount of gastric cardia intestinal metaplasia the gastroesophageal junction. Definitions of nor- Intestinal metaplasia was defined as the presence mal gastroesophageal anatomy were strictly and of goblet cells. All biopsies had Alcian blue, pH 2.5 rigidly applied. All patients included in the study stains to confirm the presence of goblet cells. The initially had a normal proximal stomach, lower amount of intestinal metaplasia was classified as esophagus, and squamocolumnar junction (Z-line). focal (goblet cells within one foveola or gland), The squamocolumnar junction lay within the 2 cm moderate (goblet cells within 2 to 5 foveola or of mucosa immediately proximal to the uppermost glands), or extensive (goblet cells within more than (proximal) gastric fold and overlaid the junction of 5 glands). the tubular end of the esophagus and the saccular Almost all of the patients initially had tissue pro- dilatation of the stomach. No patients had a prom- cured from esophageal squamous mucosa that ac- inent or exaggerated squamocolumnar junction in companied gastric cardia mucosa biopsies in the which a columnar mucosa tongue greater than 1 or same container and from the gastric antrum that 2 mm extended into the lower esophagus. The pa- was placed in a separate specimen container. The tients underwent endoscopy for many reasons, and amount of esophageal squamous GERD changes, were randomly identified based on the absence of a cardia Helicobacter pylori infection, and gastric an- hiatal hernia and the presence of CIM. Inclusion in trum inflammation and Helicobacter pylori infec- the study required at least one repeat endoscopy. tion was quantified in these biopsies, and are re- Patients were also excluded from the study if the ported in a prior study (26). Too few patients had presence or absence of specific anatomic features repeat esophageal squamous or gastric antrum tis- and relationships were not mentioned in the pro- sue biopsies to allow for meaningful comparisons cedure notes or were not clear from the retroflexed- and analyses. Therefore, this study was restricted to positioned photographs. In order to maintain a the changes within cardia and SCJ mucosa. Thirty- cautious and conservative approach to patient in- five patients (41%) had Helicobacter pylori identi- clusion in the study, patients were excluded if there fied in their antral biopsies. was possibility that a short tongue of columnar All of the biopsies were procured with standard, mucosa extended above the junction of the saccular 2.2-mm (closed) cup, standard biopsy forceps. The stomach and tubular esophagus for a distance initial endoscopic biopsies were randomly reviewed greater than 2 mm. Specific symptom information as a single group, and the repeat endoscopic biop- was not recorded. Excluding the specific descrip- sies (from all of the endoscopies) were randomly tions of the gastroesophageal junction, endoscopic reviewed as a second group. The histology of the appearances in relation to categorizing the nature patient’s initial endoscopic biopsies was unknown of any inflammation of the lower esophagus and when the repeat endoscopic biopsies were being stomach were not used in the study. reviewed. Which of the repeat endoscopies was also All patients initially had a biopsy that straddled unknown when the review of the repeat endoscopic the squamocolumnar junction composed of squa- biopsies was being performed. mous and columnar mucosa via retroflexed endo- scope. Most patients also had an additional pure squamous mucosa biopsy procured from the lower RESULTS 2 cm of the esophagus, and a biopsy fragment com- posed only of columnar mucosa from the cardia Initial Endoscopy region within 1 cm below the squamocolumnar The median and mean ages at the time of the junction. Tissue was submitted in Hollande’s solu- initial endoscopy were 58 and 55 years, respectively tion. All the blocks were cut in an identical manner. (range, 29 to 74; standard deviation, 10.8 years). Three levels were cut, with approximately 200 m Fifty-eight patients (68%) were male. The median between each level. number of cardia tissue fragments examined per The following histologic features within the car- patient was two (range, one to four). dia mucosa were examined: All patients, by definition of patient inclusion in the study, had goblet cells in the gastric cardia in Gastric cardia inflammation the initial biopsy. Fifty-six (66%) patients had focal Classifications were noted as mild, moderate, or gastric cardia goblet cells, 27 (32%) had moderate marked using the written criteria and color draw- goblet cells, and two (2%) had extensive goblet cells. ings of the visual analogue scales provided within Cardia intestinal metaplasia was present on one the updated Sydney Gastritis Classification system biopsy tissue fragment in all of the cases with focal (25). For comparison and statistical purposes, mild and moderate numbers of goblet cells. Goblet cells inflammation was given a value of 1, moderate in- were present on two biopsy tissue fragments in one flammation was given a value of 2, and marked of the two patients with extensive goblet cells, and inflammation was given a value of 3. on three biopsy tissue fragments in the other. The Gastric Cardia Intestinal Metaplasia (N.S. Goldstein) 1073 amount of cardia