Collagenous Gastritis: A Morphologic and Immunohistochemical Study of 27 Cases
Thomas Arnason, Ian Brown, Blake O'Brien, Claire Wilson, Harland Winter, Gregory Y Lauwers. Massachusetts General Hospital, Boston, MA; Royal Brisbane and Women's Hospital, Brisbane, Qld, Australia; Sullivan Nicolaides Pathology, Brisbane, Qld, Australia
Background: Collagenous gastritis (CG) is defined by the histologic finding of a superficial subepithelial collagen layer. Disease associations include celiac disease (CD), collagenous sprue (CS) and collagenous colitis (CC). The cause of increased collagen deposition remains undetermined. This study aims to further characterize the morphologic spectrum of CG through histologic study of the largest series of cases to date and investigate two possible pathogenetic mechanisms related to eotaxin and IgG4.
Design: All diagnoses of CG, including consults, were identified at three hospitals. Slides were reviewed for features including collagen thickness, intraepithelial lymphocyte (IEL) count and eosinophil count. Immunohistochemical stains for eotaxin and IgG4/IgG were applied to a subset. Clinical records were reviewed for demographics and medical history.
Results: 27 cases of CG (19 female, 8 male) were identified with a mean age at onset of 33 years (range 3-89 years). 56% of patients (15/27) presented between ages 12-35 years. 4 patients had associated CD and 6 had CS and/or CC (1 patient had all 3). Histologic review confirmed subepithelial collagen (mean thickness 49µm, range 15-145µm) and at least focal surface epithelial sloughing in all cases. Two distinct inflammatory patterns were observed, eosinophil rich (>30 eosinophils/HPF) and lymphocyte rich (>25 surface IELs/100 epithelial cells) inflammation. The eosinophil rich pattern was observed in 12/27 (44%) cases and was more frequent in younger patients, with mean age 19 years vs. 41 years for those without this pattern (p=0.022). There were increased surface IELs in 4 cases, all of which had associated CD, CS, and/or CC, which was a more frequent rate of associated disease than in the remainder of the population (p<0.01). No increased eotaxin expression was identified in 9 cases tested. 1 of 11 cases tested had increased IgG4 positive cells (100/HPF) with an IgG4/IgG ratio of 55%. The other 10 cases had a mean IgG4/IgG ratio of 18%. No patients had evidence of systemic IgG4 related disease.
Conclusions: This study identified two distinct inflammatory patterns in CG. A lymphocytic pattern (increased surface IELs) was more common in the setting of CD, CS, and CC. An eosinophilic pattern was identified in younger patients, but the absence of increased eotaxin expression did not suggest an allergic/atopic mechanism. The finding of an increased IgG4/IgG ratio was too infrequent to implicate IgG4 as a major pathway in the pathogenesis of CG.
Wednesday, March 6, 2013 1:00 PM
Poster Session VI # 76, Wednesday Afternoon