BACKGROUND

Colorectal cancer guidelines recommend screening colonoscopy every 10 years after a negative procedure. If risk reduction extends past 10 years, the recommended interval could be extended, reducing the burden on the individual and health-care system. We aimed to estimate the duration that patients remain at reduced risk of colorectal cancer incidence and mortality after a complete colonoscopy.

METHODS

We did a population-based cohort study of individuals aged 50-65 years between Jan 1, 1994, to Dec 31, 2017. We excluded individuals with previous exposure to colonoscopy or colorectal surgery, those previously diagnosed with colorectal cancer, or a history of hereditary or other bowel disorders. We followed up participants until Dec 31, 2018, and identified all colonoscopies performed in this time period. We used a 9-level time-varying measure of exposure, capturing time since last complete colonoscopy (no complete colonoscopy, ≤5 years, >5-10 years, >10-15 years, and >15 years) and whether an intervention was performed (biopsy or polypectomy). A Cox proportional hazards regression model adjusting for age, sex, comorbidity, residential income quintile, and immigration status was used to estimate the association between exposure to a complete colonoscopy and colorectal cancer incidence and mortality.

FINDINGS

5 298 033 individuals (2 609 060 [49·2%] female and 2 688 973 [50·8%] male; no data on ethnicity were available) were included in the cohort, with a median follow-up of 12·56 years (IQR 6·26-20·13). 90 532 (1·7%) individuals were diagnosed with colorectal cancer and 44 088 (0·8%) died from colorectal cancer. Compared with those who did not have a colonoscopy, the risk of colorectal cancer in those who had a complete negative colonoscopy was reduced at all timepoints, including when the procedure occurred more than 15 years earlier (hazard ratio [HR] 0·62 [95% CI 0·51-0·77] for female individuals and 0·57 [0·46-0·70] for male individuals. A similar finding was observed for colorectal cancer mortality, with lower risk at all timepoints, including when the procedure occurred more than 15 years earlier (HR 0·64 [95% CI 0·49-0·83] for female participants and 0·65 [0·50-0·83] for male participants). Those who had a colonoscopy with intervention had a significantly lower colorectal cancer incidence than those who did not undergo colonoscopy if the procedure occurred within 10 years for females (HR 0·70 [95% CI 0·63-0·77]) and up to 15 years for males (0·62 [(0·53-0·72]).

INTERPRETATION

Compared with those who do not receive colonoscopy, individuals who have a negative colonoscopy result remain at lower risk for colorectal cancer incidence and mortality more than 15 years after the procedure. The current recommendation of repeat screening at 10 years in these individuals should be reassessed.

FUNDING

Canadian Institutes of Health Research.