Background Surveillance colonoscopy after polyp removal is recommended to prevent subsequent colorectal cancer (CRC). It is known that advanced adenomas have a substantially higher risk than non-advanced ones, but optimal intervals for surveillance remain unclear.

Design We prospectively followed 156 699 participants who had undergone a colonoscopy from 2007 to 2017 in a large integrated healthcare system. Using multivariable Cox proportional hazards regression we estimated the subsequent risk of CRC and high-risk polyps, respectively, according to index colonoscopy polyps, colonoscopy quality measures, patient characteristics and the use of surveillance colonoscopy.

Results After a median follow-up of 5.3 years, we documented 309 CRC and 3053 high-risk polyp cases. Compared with participants with no polyps at index colonoscopy, those with high-risk adenomas and high-risk serrated polyps had a consistently higher risk of CRC during follow-up, with the highest risk observed at 3 years after polypectomy (multivariable HR 5.44 (95% CI 3.56 to 8.29) and 8.35 (95% CI 4.20 to 16.59), respectively). Recurrence of high-risk polyps showed a similar risk distribution. The use of surveillance colonoscopy was associated with lower risk of CRC, with an HR of 0.61 (95% CI 0.39 to 0.98) among patients with high-risk polyps and 0.57 (95% CI 0.35 to 0.92) among low-risk polyps. Among 1548 patients who had high-risk polyps at both index and surveillance colonoscopies, 65% had their index polyps in the proximal colon and 30% had index and interval polyps in the same segments.

Conclusion Patients with high-risk polyp findings were at higher risk of subsequent CRC and high-risk polyps and may benefit from early surveillance within 3 years. The subsite distribution of the index and recurrent high-risk polyps suggests the contribution of incomplete resection and missed lesions to the development of interval neoplasia.