Postpolypectomy colonoscopy surveillance guidelines: Predictive accuracy for advanced adenorna at 4 years

Background: Lack of confidence in postpolypectomy surveillance guidelines may be a factor in the observed low adherence rates among providers. Objective: To assess the 2006 postpolypectomy colonoscopy surveillance guidelines, which recommend 3-year follow-up colonoscopy for individuals with high-risk adenomas (defined as >= 3 adenomas or any advanced adenomas) and 5- to 10-year follow-up for patients with 2 or fewer nonadvanced adenomas, who are considered to be at low risk. Design: Analysis of prospective data from the Polyp Prevention Trial. Setting: United States. Participants: 1905 patients who had colorectal adenomas removed at baseline screening or diagnostic colonoscopy and completed the trial. Measurements: Baseline adenoma characteristics, risk-stratified according to definitions used in the guidelines, were examined as predictors for advanced adenoma recurrence. Results: 125 patients (6.6%) had advanced and 629 (33.0%) had nonadvanced adenoma recurrence; 1151 (60.4%) had no recurrence within 4 years of follow-up. The probability of advanced adenoma recurrence was 0.09 (95% Cl, 0.07 to 0.11) among patients with high-risk adenomas at baseline and 0.05 (Cl, 0.04 to 0.06) among those with low-risk adenomas at baseline. The relative risk for advanced adenoma recurrence for patients with high-risk adenomas versus those with low-risk adenomas at baseline was 1.68 (Cl, 1.19 to 2.38) when advanced adenoma recurrence was compared with no advanced adenoma recurrence and 1.76 (Cl, 1.26 to 2.46) when advanced adenoma recurrence was compared with no adenoma recurrence. The c-statistics for these 2 comparisons were 0.68 and 0.72, respectively. Limitation: Participants were self-selected and had restrictions on the degree of obesity. Conclusion: Although the risk for recurrence of advanced adenoma within 4 years is greater for patients with high-risk adenomas at baseline than for those with low-risk adenomas, the discrimination of this risk stratification scheme is relatively low.