Diagnostic value of distal colonic polyps for prediction of advanced proximal neoplasia in an average-risk population undergoing screening colonoscopy.

BACKGROUND For colorectal cancer screening, the predictive value of distal findings in the ascertainment of proximal lesions is not fully established. The aims of this study were to assess distal findings as predictors of advanced proximal neoplasia and to compare the predictive value of endoscopy alone vs. combined endoscopic and histopathologic data. METHODS Primary colonoscopy screening was performed in 2210 consecutive, average-risk adults. Age, gender, endoscopic (size, number of polyps), and histopathologic distal findings were used as potential predictors of advanced proximal neoplasms (i.e., any adenoma > or =1 cm in size, and/or with villous histology, and/or with severe dysplasia or invasive cancer). Polyps were defined as distal if located in the descending colon, the sigmoid colon, or the rectum. Those in other locations were designated proximal. RESULTS Neoplastic lesions, including 11 invasive cancers, were found in 617 (27.9%) patients. Advanced proximal neoplasms without any distal adenoma were present in 1.3% of patients. Of the advanced proximal lesions, 39% were not associated with any distal polyp. Older age, male gender, and distal adenoma were independent predictors of advanced proximal neoplasms. The predictive ability of a model with endoscopic data alone did not improve after inclusion of histopathologic data. In multivariate logistic regression analysis, the predictive ability of models that use age, gender, and any combination of distal findings was relatively low. The proportion of advanced proximal neoplasms identified if any distal polyp was an indication for colonoscopy was only 62%. CONCLUSIONS A strategy in which colonoscopy is performed solely in patients with distal colonic findings is not effective screening for the detection of advanced proximal neoplasms in an average-risk population.