criterion was that informed consent should be obtained from all patients undergoing colonoscopy in each center. Exclusion criteria included the following: age of 39 years or lower, inflammatory bowel disease, familial adenomatous polyposis, patients whose bowel preparation was inadequate, patients who had undergone colonoscopy in the past 6 months, patients who could not be observed up to the cecum although stenotic CRC did not exist, and patients who were already diagnosed with colorectal tumors but not treated. All procedures were performed or supervised by experienced endoscopists at each center. We collected data of age, sex, past history of CRC, family history of CRC, and reasons for colonoscopy. The relationship between the frequency of colonoscopy in the past 5 years and detection rate of advanced adenoma or CRC was analyzed. The Cochran-Armitage test was used to assess trends in the incidence proportions of outcomes. To adjust for risk factors, we used multiple logistic regression to estimate odds ratios. Results: A total of 5818 patients were finally analyzed. When the frequency of colonoscopy over the past 5 years increased, the detection rate of advanced adenoma or CRC tended to decrease (p 0.01). Table 1 shows the odds ratio of advanced adenoma. However, if the examination frequency over the past 5 years increased, the odds ratio did not decrease more than approximately 70%. Table 2 shows the odds ratio of CRC. CRC was not detected in most patients who underwent colonoscopy twice in the past 5 years. Conclusions: Our results indicate that undergoing colonoscopy twice in 5 years is recommended to prevent CRC.