Predictors of advanced proximal neoplasia in persons with abnormal screening flexible sigmoidoscopy.

BACKGROUND & AIMS: The relationship between distal and proximal colonic findings is uncertain. Thus, there is no consensus on which findings on screening flexible sigmoidoscopy should trigger colonoscopy. METHODS: We analyzed data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial to assess the relationship between distal and proximal colonic findings. RESULTS: A total of 8802 subjects had an abnormal baseline sigmoidoscopy and colonoscopy follow-up. Subjects with <10-mm single or multiple tubular adenomas had similar risks for advanced proximal neoplasia as subjects with hyperplastic polyps or other benign lesions (3%-5%). Subjects with large (>or=10 mm), villous, or severely dysplastic distal adenomas had similarly elevated risks for advanced proximal neoplasia (11%-12%). Multivariate logistic modeling showed a significantly increased risk for advanced proximal neoplasia associated with the presence of a large tubular (odds ratio OR], 2.6; 95% confidence interval CI], 2.0-3.4) or villous distal adenoma (OR, 2.7; 95% CI, 2.1-3.5) but not with the presence of one (OR, 1.05; 95% CI, 0.8-1.3) or multiple (OR, 0.8; 95% CI, 0.5-1.2) <10-mm tubular distal adenomas. CONCLUSIONS: Among subjects with a polypoid lesion on screening flexible sigmoidoscopy, those with small tubular distal adenomas are at similar risk for advanced proximal neoplasia as those without distal adenomas. Subjects with a large, villous, or dysplastic distal adenoma are at increased risk. A strategy that encourages individuals with small tubular adenomas on sigmoidoscopy to undergo follow-up colonoscopy and excludes those with nonadenomatous lesions is of questionable validity, because both groups are at similar risk for advanced proximal neoplasia.