COLORECTAL CANCER INCIDENCE AND MORTALITY AFTER NEGATIVE COLONOSCOPY SCREENING ACCORDING TO INDIVIDUALS’ RISK PROFILES IN 3 U.S POPULATION-BASED COHORTS

Background: The long-term risk of colorectal cancer (CRC) incidence and mortality after negative colonoscopy screening (NCS) according to individuals’ risk profiles remains unknown.
Method: We prospectively examined CRC incidence and mortality according to the NCS status in the Nurses’ Health Study I/II and Health Professionals Follow-up Study. Leveraging the repeatedly collected epidemiologic information, we further assessed CRC risk according to a risk score (range, 0-12) based on the major CRC risk factors: family history of CRC, age, sex, smoking status, body mass index, tall stature, physical activity, alcohol consumption, aspirin use, and diet. We used the time-varying Cox regression to calculate the hazard ratios (HRs) and 95% confidence intervals (CIs) to investigate the association between NCS and the risk score and CRC incidence and mortality.
Results: During a median follow-up of 14 years (interquartile range, 7-20), we documented 394 and 2 959 CRC cases, and 165 and 1 063 CRC deaths, among 81 151 individuals with NCS and 122 138 without NCS, respectively. Compared to individuals without NCS, individuals with NCS were consistently associated with lower risk of CRC incidence and mortality for 20 years after NCS, with the multivariable HR (95% CI) 0.67 (0.59, 0.75) and 0.62 (0.51, 0.75), respectively. When assessed by CRC subsites, the reduction in CRC incidence was more substantial for rectal cancer with a HR (95%CI) of 0.60 (0.44, 0.81) and distal colon cancer of 0.50 (0.37, 0.67) than proximal colon cancer of 0.79 (0.65, 0.95) (P_{heterogeneity}=0.02). The risk score showed a stronger association with CRC incidence in individuals without NCS than individuals with NCS, with the HR (95% CI) of 0.31 (0.28, 0.34) and 0.46 (0.34, 0.61), respectively, comparing the risk score of 0-5 to 8-12 (P_interaction=0.02). In the joint analysis, compared to individuals with no negative colonoscopy screening and a risk score of 8-12, those with no negative colonoscopy screening and a risk score of 6-7 and 0-5 had a HR (95%CI) of 0.58 (0.52, 0.64) and 0.31 (0.28, 0.35) for CRC incidence, respectively; those with a negative colonoscopy screening had an even lower risk of CRC incidence, with the HR (95%CI) of 0.54 (0.42, 0.68), 0.40 (0.34, 0.48), and 0.24 (0.19, 0.29) for the risk score of 8-12, 6-7, and 0-5, respectively (P-trend <0.0001). Among individuals with NCS, those with a score of 0-5 and 6-7 did not reach the 10-year cumulative incidence of CRC (~0.9%) in the 8-12 score group until ~25 and ~16 years, respectively, after the initial screening.
Conclusion:
The rescreening interval after a negative colonoscopy screening may be extended beyond the currently recommended 10 years, particularly for individuals with a low-risk profile.