Incident rates of colonic neoplasia in older patients: When should we stop screening?

• Published in: Journal of gastroenterology and hepatology 2006-06, Vol.21 (6), p.1021-1025
• Main Authors: Harewood, Gavin C, Lawlor, Garrett O, Larson, Mark V
• Format: Article
• Language: English
• Subjects: Aged; Aged, 80 and over; Biological and medical sciences; Colonic Neoplasms - diagnosis; Colonic Neoplasms - epidemiology; Colonoscopy; colorectal cancer; Digestive system. Abdomen; Endoscopy; Female; Gastroenterology. Liver. Pancreas. Abdomen; Humans; Incidence; Investigative techniques, diagnostic techniques (general aspects); Male; Mass Screening; Medical sciences; screening; Stomach. Duodenum. Small intestine. Colon. Rectum. Anus; Tumors
• ISSN: 0815-9319; 1440-1746
• DOI: 10.1111/j.1440-1746.2006.04218.x

Introduction:  Current guidelines endorse colon cancer screening every 5–10 years in patients over 50 years of age. However, there is no consensus regarding what age is appropriate to stop screening. The aim of this study was to characterize neoplasia occurrence/recurrence in a large cohort of patients ≥70 years of age undergoing colonoscopy. Methods:  The Mayo Rochester endoscopic database was reviewed to determine the incidence of colonic neoplasia in patients ≥70 years undergoing two colonoscopies at least 12 months apart between January 1996 and December 2000. Patients were classified based on (i) age: 70–74, 75–79, ≥80 years; and (ii) polyp detection on initial examination, that is, subsequent examination for screening or surveillance. Results:  Overall, 1353 patients underwent two colonoscopies at least 12 months apart (median interval 140 weeks) with removal of polyp on initial examination in 726 (53.7%) patients (surveillance cohort). On subsequent endoscopy, polyps ≥10 mm were detected in 54 (4.0%) and cancer in 13 (1.0%) patients. All age groups were well matched with respect to detection of neoplasia on index examination (P = 0.9) and polyp size on initial colonoscopy among the surveillance group (P = 0.9). Using a Cox proportional hazards model, adjusted hazard ratios (95% confidence interval [CI]) for neoplasia (polyps ≥10 mm) were: 2.0 (1.50–2.73, P