Flexible sigmoidoscopy following a positive faecal occult blood test within a bowel screening programme may reduce the detection of neoplasia.

Abstract

Colorectal cancer screening using the faecal occult blood test (FOBt) detects a disproportionate number of left-sided tumours. This study aims to examine the theoretical impact on neoplasia detection rates of a sigmoidoscopy-first protocol in FOBt-positive patients undergoing colonoscopy. From retrieved endoscopy/pathology reports, pathology up to and including the splenic flexure was assumed detectable by sigmoidoscopy. High-risk polyps prompting subsequent colonoscopy were classed as three or more polyps, one polyp of ≥ 1 cm, villous or tubulovillous components or the presence of high-grade dysplasia. Between April 2009 and April 2011, 4631 patients underwent colonoscopy as a result of a positive FOBt in Greater Glasgow and Clyde. Cancer was detected in 398 (9%) and adenomas were detected in 1985 (47%) of which 1323 (67%) were deemed significant according to British Society of Gastroenterology guidelines. Applying the flexible sigmoidoscopy-first model, cancer would have been detected in 329 (8%) patients and adenomas in 1640 (39%), of which 1140 (70%) would have been significant. In total, 1546 (37%) patients would have required subsequent colonoscopy, following which 21 patients would have a new diagnosis of cancer. The positive predictive values (PPVs) for neoplasia (47 vs 57%, P < 0.001), significant neoplasia (35 vs 41%, P < 0.001) and cancer (8 vs 9%, P = 0.007) were all lower in the sigmoidoscopy-first model. A significant reduction in the detection of both adenomas and cancers would be seen if the sigmoidoscopy-first protocol were to be used following a positive FOBt. Furthermore, a significant proportion of patients would be subjected to two procedures, with considerable implications for both the patient and cost.