ABSTRACT
AIM: Incomplete colonoscopy indicated for the detection of neoplasia occurs in 2-23% of patients, but there is little information on the long-term outcome of such patients. METHOD: All patients who underwent colonoscopy over 5 years at the Royal Liverpool University Hospital with a follow-up of up to 5 years were identified. RESULTS: The risk of colorectal cancer (CRC) was 2.9% (312/10 580) for all patients undergoing colonoscopy. For a failed colonoscopy, the risk was five-fold higher [14.3% (99/693)]. The mean age of the patients was 61 years and 58% were female. Following incomplete colonoscopy the risk of finding additional CRC, advanced colonic neoplasia and extracolonic neoplasia on subsequent investigation was 6.2%, 3.2% and 1.9%. The diagnostic yield on subsequent investigation for CRC or colonic polyps was 7% for repeat colonoscopy, 13.4% for computed tomography colonography, 10.3% for standard computed tomography and 1.8% for barium enema. In the 363 patients who were not offered a subsequent investigation, there was no further instance of CRC or CRC-related mortality over a 36-month period. CONCLUSION: Although the risk of CRC is higher in patients who have had a failed colonoscopy, a protocol approach of subsequent investigation should not replace clinical assessment on whether another test is necessary in the light of the good outcome of patients who were not subsequently investigated.