Diminutive polyps among black and Latino populations undergoing screening colonoscopy: evidence supporting a resect and discard approach

Background: A resect and discard strategy for diminutive (<= 5 mm) colon polyps has been proposed to save costs of screening colonoscopy (SC). Prior studies on neoplasia prevalence based on polyp size have involved mostly white patients. Objective: To determine the prevalence of adenomas and advanced histologic features by size among primarily black and Latino patients enrolled in a prospective SC study. Design: Retrospective analysis of data from a prospective clinical trial. Setting: Urban academic medical center. Patients: Average risk, asymptomatic, minority patients aged >= 50 years undergoing SC. Interventions: Screening colonoscopy. Main Outcome Measurements: Rates of neoplasia and advanced histologic features (villous histology, high-grade dysplasia, or cancer) by polyp size and location. Results: A total of 566 polyps from 295 patients were analyzed. Diminutive polyps and small (6-9 mm) polyps had lower prevalence of >= 1 advanced feature compared with large (>= 10 mm) polyps (0.9% and 2.7%, respectively, vs 13.6%; P < .001 for both comparisons). Distal polyps were less likely to be neoplastic (31.7% vs 61.4%; P < .001) than proximal polyps in all size categories (P < .001 for all comparisons). After adjusting for sex, ethnicity, age, and location, large polyps were more likely to have >= 1 advanced feature than diminutive polyps (adjusted odds ratio [OR] 19.5; 95% CI, 4.4-85.6) or small polyps (adjusted OR 6.1; 95% CI, 2.2-16.9). Limitations: Use of pathology reports for polyp size. Conclusion: Among a cohort of minority patients, advanced histologic features were very rare in diminutive polyps. Distal polyps were less likely to be neoplastic than proximal polyps in all size categories. This supports a resect and discard strategy for diminutive polyps, especially in the distal colon.