Lynch syndrome is the most common inherited cause of colorectal cancer (CRC). Contemporary and mutation-specific estimates of CRC-risk in patients undergoing colonoscopy would optimize surveillance strategies. We performed a prospective national cohort study, using data from New Zealand, to assess overall and mutation-specific risk of CRC in patients with Lynch syndrome undergoing surveillance.We performed a prospective study of 381 persons with Lynch syndrome in New Zealand (98 with Lynch-syndrome associated variants in MLH1, 159 in MSH2, 103 in MSH6, and 21 in PMS2). Participants were offered annual colonoscopy starting at age 25 y, and those who underwent 2 or more colonoscopies before December 31, 2017 were included in the final analysis. Patients with previous colonic resection, history of CRC or diagnosis of CRC at index colonoscopy were excluded.Study participants underwent 2061 colonoscopies during 2296 person-y; the median observation-period was 4.43 y and mean-age at enrollment was 43 y. Eighteen patients developed CRC (8 with variants in MLH1, 8 in MSH2, and 2 in MSH6) after a median follow-up period of 6.5 y (range 1-16 y). Eighty-three percent of patients had a surveillance colonoscopy in preceding 24 months before diagnosis of CRC; 94% were diagnosed with stage 0-II CRC and there was no CRC-related mortality. The overall-risk of developing CRC in the 5 y after first surveillance colonoscopy was 2.49% (95% CI, 1.18-5.23); cumulative risk for CRC in patients with Lynch syndrome-associated variants in MLH1, MSH2, or MSH6 by age 70 y were 17.7%, 17.8%, and 8.5%, respectively. Age-adjusted CRC-risk in patients with variants in MSH6 was lower than in MLH1 (hazard ratio, 0.2; 95% CI, 0.04-0.94; P=.02). Of patients with CRC, 33% had an adenomatous polyp resected from same segment in which a colorectal tumor later developed.The risk of CRC in patients with Lynch syndrome-associated mutations in MSH6 or PMS2 was significantly lower than in patients with mutations in MLH1. Incomplete adenomatous polyp resection might be responsible for one third of surveillance-detected CRCs.