Predictors of metachronous colorectal neoplasms in sporadic adenoma patients

Our objective was to assess the overall risk of subsequent colorectal neoplasms (cancer or adenoma) in relation with the various characteristics of the index lesion in a cohort of patients who underwent endoscopic polypectomies of colorectal adenomas. A total of 1086 patients with adenomas of the large bowel were reported between 1979 and 1999 at the National Cancer Institute of Milan during a screening program for colorectal carcinoma. Data on patients who had colonoscopic examinations and treatments were collected prospectively. The relation between colorectal cancer (CRC) and adenoma features was assessed by computing the hazard ratio (HR) values and corresponding confidence intervals (95% CI) according to Cox proportional hazard models. Of the 1086 eligible patients (487 females, 579 males), 736 had single adenomas (67.7%) and 350 had multiple adenomas (32.3%). Histologic examination revealed 772 cases of tubular adenoma (73%), 205 cases of tubulovillous adenoma and 80 cases of villous adenoma (7.5%). Severe dysplasia was found in 3.3% of the cases. During the 11393 person-years of follow-up, with an average time of surveillance of 10.5 years, colorectal carcinomas developed in 10 patients (0.8%) and a new adenoma in 323 patients (29%). Multivariate analysis showed that male gender (HR 1.6; 95% CI 1.3-2.0), multiple polyps (HR 1.6; 95% CI 1.3-2.0), polyps larger than 2 cm (HR 1.5; 95% CI 1.1-2.1), tubulovillous and villous histology (HR 1.3; 95% CI 1.0-1.6 and HR 1.8; 95% CI 1.2-2.6, respectively) at index polypectomy were statistically significant risk factors for developing metachronous adenomatous polyps. The standardized incidence rates (SIR) for CRC was 0.52 (95% CI 0.25-0.95). The SIR was increased in subjects with severe dysplasia (2.8; 95% CI 0.34-1.02). Some features of large bowel adenomas are strongly correlated with an increased risk of metachronous adenomas and colorectal cancer. However, the endoscopic polypectomy is able to reduce by 50% the incidence of CRC in patients with large bowel adenomas.