In collagenous colitis, the literature is conflicting concerning where in the colon the lesions are most likely to be present and most severe. Conflicting data furthermore shed doubt on the sensitivity of the histological detection of the morphological abnormalities and the threshold criteria for diagnosis. We addressed these questions in 56 patients with collagenous colitis. Two hundred ninety-one coded biopsy specimens were analyzed according to six standardized sites from cecum to rectum. Subepithelial collagen deposits were subjectively graded in hematoxylin and eosin (H&E) sections and quantitatively measured in trichrome-stained sections, respectively. Semiquantitative grading was also done for inflammatory changes of the lamina propria and abnormalities of the surface and crypt epithelium. The transverse colon yielded the largest percentage of biopsy specimens (83%) interpreted as diagnostic of collagenous colitis and also had the largest percentage of biopsy specimens with inflammatory changes (98%). Biopsy specimens from both the rectosigmoid and the right colon (ascending and cecum) were significantly less likely to be diagnostic (P<.01). Only 66% of specimens obtained from the rectosigmoid were diagnostic, and 18% of these were interpreted as normal. Subepithelial collagen deposits proved to be significantly thicker in the transverse (median, 46.8 microm; range, 12 to 212.4) and descending (median, 49.2 microm; range, 6 to 230.4) than in the rectosigmoid (median, 33.6 microm; range, 9.6 to 178.8) and right colon (median, 35.4 microm; range, 6 to 140.4), respectively (P<.01). Almost all biopsy specimens (97%) had collagen deposits thicker than 10 microm. However, the subjective interpretation "diagnostic of collagenous colitis" proved to be most consistent with a threshold of 30 microm. Our results indicate that biopsy specimens from at least as proximal as the transverse colon should be obtained to definitely rule out collagenous colitis. Furthermore, it is evident that in a given biopsy specimen, markedly abnormal subepithelial collagen deposition had to be present for an unequivocal histological diagnosis of collagenous colitis.