CASE REPORT Collagenous Gastritis: A Case Report, Morphologic Evaluation, and Review Z. Vesoulis, M.D., Gerard Lozanski, M.D., Pars Ravichandran, M.D., Edward Esber, M.D. Departments of Pathology (ZV, GL, PR) and Gastroenterology (EE), Summa Health Systems, Akron, Ohio dently identified by using analogous defined fea- Collagenous gastritis is rare; there are only four tures of collagenous colitis: subepithelial collagen previous case reports. Histologic features seem to more than 10 ␮ in a patchy distribution, lamina overlap with the other “collagenous enterocoliti- propria lymphoplasmacytic infiltrates, intraepithe- des”; however, pathologic criteria are not yet estab- lial lymphocytes, and surface epithelial damage. lished for the diagnosis of collagenous gastritis. We Collagenous gastritis also seems to have the same describe an additional case of ostensible collage- spectrum of associated clinical findings as collage- nous gastritis in a patient who initially presented nous colitis, including frequent coexistence of celiac with celiac sprue and subsequently developed co- sprue, watery diarrhea syndrome, and female pre- lonic manifestations of mucosal ulcerative colitis. dominance. Endoscopic biopsies of the stomach revealed depo- sition of patchy, very thick bandlike subepithelial KEY WORDS: Celiac sprue, Collagenous colitis, Col- collagen in gastric antral mucosa, focal superficial lagenous gastritis, Lymphocytic colitis, Lympho- epithelial degeneration, numerous intraepithelial cytic gastritis, Mucosal ulcerative colitis. lymphocytes, and a dense lamina propria lymphop- Mod Pathol 2000;13(5):591–596 lasmacytic infiltrate. Image analysis evaluation of gastric antral biopsies demonstrated a mean thick- Collagenous colitis, first described by Lindstrom in ness of subepithelial collagen of 27.07 ␮. Morpho- 1976 (1), is an inflammatory bowel lesion; diagnosis logic comparison was made with age-matched con- requires the clinical presentation of chronic watery trol groups of 10 patients who had normal gastric diarrhea, an endoscopically normal or “near nor- mucosal biopsies and 10 patients who had “chron- mal” mucosa and colon biopsy with subepithelial ic” gastritis, which revealed mean subepithelial col- collagen deposition, chronic inflammation, and ep- lagen measures of 1.37 ␮ and 1.19 ␮, respectively. ithelial degeneration. Collagenous colitis is proba- We compared these morphologic findings with bly an autoimmune disorder (2–6), although re- those of all previous case reports of collagenous cently other potential causes have been proposed gastritis and propose a pathologic definition based and include bacterial toxins, cytotoxins, nonsteroi- on the limited combined data. It seems that subep- dal anti-inflammatory drugs, infectious agents, and ithelial collagen is dramatically thickened in re- antibiotics (7–10). ported cases of collagenous gastritis, with a cumu- “Collagenous” sprue was originally described as a lative mean measure of 36.9 ␮. ⌱t is also apparent variant of typical sprue with villous atrophy, prom- from this and previous reports that the thickened inent subepithelial collagen deposition, and refrac- subepithelial collagen is accompanied by a chronic toriness to a gluten-free diet (11). Some degree of or chronic active gastritis and sometimes intraepi- collagen deposition in sprue may be a common and thelial lymphocytes and surface epithelial damage. relatively nonspecific finding that occurs without Recently described associations of lymphocytic gas- alteration of the typical clinical course (12). tritis, sprue, and lymphocytic colitis as well as col- Collagen deposition in gastric mucosa has rarely lagenous and lymphocytic colitis suggest a common been described; there are four case reports to date, pathogenesis that empirically may include collage- and no detailed pathologic descriptions or diagnos- nous gastritis in the same disease spectrum. We tic criteria have been elucidated (13–16). We de- propose that collagenous gastritis can be confi- scribe a case of collagenous gastritis that developed in association with celiac disease and ulcerative colitis. Careful morphometric comparison of this Copyright © 2000 by The United States and Canadian Academy of Pathology, Inc. case and control groups of normal gastric antral VOL. 13, NO. 5, P. 591, 2000 Printed in the U.S.A. biopsies and “chronic gastritis” cases illustrates the Date of acceptance: November 2, 1999. Address reprint requests to: Dr. Vesoulis, Department of Pathology, striking and overt subepithelial collagen deposition Summa Health Systems, 525 E. Market Street, Akron, OH 44304. that delineates this unusual lesion. We also review 591 and compare the clinical and pathologic features of strict emphasis on maintaining a gluten-free diet, other reported cases of collagenous gastritis and his diarrhea and hematochezia resolved. Antiglia- attempt to derive unifying features and histopatho- din and antiendomysial antibodies were negative in logic criteria that characterize this lesion. The rela- follow-up studies. tionship and frequent co-occurrence of collagenous gastritis, collagenous and lymphocytic colitis, and sprue is also examined. MATERIALS AND METHODS Five endoscopic biopsies from the stomach, 6 CASE REPORT biopsies from the small intestine, and 15 biopsies from the colon were received in 10% buffered for- A 57-year-old Caucasian man presented with a malin and processed for histologic evaluation in the hypochromic, microcytic anemia and loose stools typical manner. All sections were stained with he- for several months. He experienceda5lbweight matoxylin and eosin, and sections of the stomach loss and denied abdominal pain. His medical his- were additionally stained with Masson’s trichrome, tory was remarkable for coronary artery disease and Giemsa, and Warthin-Starry silver stains. Immu- lower extremity vascular surgery. There was no his- nohistochemical staining with a selected panel tory of food allergies, arthritis, or autoimmune dis- of antibodies, including CD3, UCHL-1(CD45ro), ease. His initial workup included an unremarkable L-26 (CD20), and ␬ and ␭, were performed on colonoscopy, but 4 months later, after persistence all biopsies. Immunostaining incorporated the of loose stools, continued weight loss, and anemia, streptavidin-biotin-peroxidase method after over- an upper gastrointestinal and small bowel series night incubation in humidity chambers at 4° C. A demonstrated small bowel dilation. Repeat esopha- microwave antigen retrieval method was used with gogastric duodenoscopy (EGD) with biopsy of the all antibodies. Slides were immersed in 200 mL of small intestine revealed a severe mucosal villous APK (Ventana Corp., Tucson, AZ) microwave buffer abnormality. Antigliadin IgG and IgA was elevated and heated in an 800 W microwave oven using a (Ͼ15 EIA units) and antiendomysial IgA was ele- high setting for 5 min. Fifty milliliters of deionized vated (Ͼ1:5). The patient was treated with a gluten- water was added to replace evaporated water from free diet; symptoms resolved, but intermittent heating, and slides were microwaved for an addi- flares of diarrhea continued. Strict adherence to a tional 5 min before automated immunostaining on gluten-free diet was considered questionable for the Ventana Nexes (Ventana Corp.). Control tissues this patient. Repeat upper endoscopy with biopsies with established reactivity were stained concomi- revealed a persistent mild villous abnormality with- tantly. Morphometric measurements of subepithe- out significant collagen deposition. The patient was lial collagen in the hematoxylin and eosin and not rechallenged with gluten. Another colonoscopy trichrome sections of the stomach were performed was performed after the patient once again pre- on the CAS 200 image analyzer (Becton-Dickinson, sented with diarrhea. Stool cultures were negative San Jose, CA). Mean measures for each gastric frag- for bacterial infection, ova, and parasites, and C. ment were determined by two independent observ- difficile toxin assays were negative. Colon biopsy ers (GL and PR) using Ocular Micrometer v.1 soft- revealed a “nonspecific” increase in the inflamma- ware (Cell Analysis Systems, Elmhurst, IL). Only tory content of the lamina propria. The patient biopsy fragments with optimal orientation were continued to have intermittent bouts of moderate used in measurements to avoid the artifactual diarrhea and eventually developed bloody diarrhea. thickening introduced by tangential sections. Each Colonoscopy performed at this time revealed of the five optimally oriented gastric tissue frag- marked erythema with granularity and ulceration of ments was measured in at least three separate areas the entire colon. Biopsies were taken from proxi- representing the thickest visually determined zones mal, transverse, and distal colon and rectum. Pe- of subepithelial collagen. Each measurement con- rinuclear antineutrophilic cytoplasmic antibody sisted of 10 independent readings in each of the (pANCA) was positive (Ͼ1:160), and cytoplasmic three selected zones. Deposition of excessive colla- antineutrophilic cytoplasmic antibody (cANCA) gen at each point of measurement was confirmed was negative (Ͻ1:20). Serum IgA was elevated at by comparison with Masson’s trichrome stain of 682 mg/dL (range, 89 to 446 mg/dl). and IgG, IgM, parallel sections. Because there is no standard for complement C3, C4, antinuclear antibody, antithy- excessive thickness of subepithelial collagen in gas- roglobulin, and antimicrosomal antibodies were tric