Hyperplastic Polyps on Colonoscopy - Not so Benign After All

Abstract

Introduction: Hyperplastic polyps (HP) are the commonest colon polyps detected at colonoscopy. Small HPs are considered low risk colon lesions. Colorectal cancer screening and surveillance guidelines recommend an average risk colonoscopy screening interval of 10 years in individuals with <10mm rectosigmoid HPs, and offer no recommendation for < 10 mm HPs proximal to the sigmoid. We compared the risk of adenomas and serrated polyps between subjects with a normal baseline colonoscopy versus those with HPs (< 10 mm) within or proximal to the rectosigmoid colon. Methods: We obtained colonoscopy (CS) and pathology reports from patients undergoing >2 screening and surveillance colonoscopies between 2004-2014. We excluded inpatients, or patients with age 75 years, a family or personal history of CRC, IBD, colorectal surgery, or baseline colonoscopy with any adenoma, SSP, or HP >9mm. We compared the risk of metachronous polyps by baseline CS findings: normal, 1 HP, & ≥ 2 HPs. The risk of metachronous lesions is presented as odd ratios and 95% confidence interval (CI). Results: 1,798 patients were included. At index colonoscopy, 83% of patients had a normal exam, 11% had 1 < 1 cm HP and 6% had between 2 and 9 2HPs (56 mos [39.7,72.5]) vs 1 HP (62.4 mos [49.5,80.1]) vs those with a normal exam (63.7 mos [52.9,84.5]) (p<0.001). Patients with one or more HPs on index colonoscopy had an increased risk of metachronous HPs and SSPs throughout the colon (Table 1). When limited to patients with HPs in the rectosigmoid, the follow-up interval for colonoscopy was 54.2 mos [38.2,71.4] for patients with >2 HP, 61.9 mos [50.0,80.8] for patients with 1 HP, and 63.8 mos [52.9,84.5] (p<0.001) for patients with a normal exam. Patients with rectosigmoid HPs were not at an increased risk for metachronous SSPs or adenomas. Conclusion: Patients with <9mm HPs proximal to the sigmoid colon have an increased risk of metachronous HPs and SSPs. Patients with rectosigmoid HPs do not appear to have an increased risk carry the same predictive risk.274_A Figure 1. OR: odds ratio; CI: confidence interval; SSP: sessile serrated polyp a: lesions detected throughout the colon including sigmoid *Adjusted for age, indication for baseline scope and time between scopes Separate multivariable logistic regression models were built to look at each outcome. For each model, the finding of interest was modeled as the outcome with age, indication for baseline colonoscopy, findings at baseline colonoscopy (none vs. 1 HP vs. 2+ HPs) and time between colonoscopies as the independent variables274_B Figure 2. OR: odds ratio; CI: confidence interval; SSP: sessile serrated polyp a: lesions detected throughout the colon including sigmoid *Adjusted for age, indication for baseline scope and time between scopes Separate multivariable logistic regression models were built to look at each outcome. For each model, the finding of interest was modeled as the outcome with age, indication for baseline colonoscopy, findings at baseline colonoscopy (none vs. 1 HP in recto-sigmoid vs. 2+ HPs in recto-sigmoid) and time between colonoscopies as the independent variables