LAUSR

BACKGROUND The effectiveness of screening for colorectal cancer (CRC) by sex and age in randomized trials is uncertain. OBJECTIVE To evaluate the 15-year effect of sigmoidoscopy screening on CRC incidence and mortality. DESIGN Pooled analysis of 4 large-scale randomized trials of sigmoidoscopy screening. SETTING Norway, the United States, the United Kingdom, and Italy. PARTICIPANTS Women and men aged 55 to 64 years at enrollment. INTERVENTION Sigmoidoscopy screening. MEASUREMENTS Primary end points were cumulative incidence rate ratio (IRR) and mortality rate ratio (MRR) and rate differences after 15 years of follow-up comparing screening versus usual care in intention-to-treat analyses. Stratified analyses were done by sex, cancer site, and age at screening. RESULTS Analyses comprised 274 952 persons (50.7% women), 137 493 in the screening and 137 459 in the usual care group. Screening attendance was 58% to 84%. After 15 years, the rate difference for CRC incidence was 0.51 cases (95% CI, 0.40 to 0.63 cases) per 100 persons and the IRR was 0.79 (CI, 0.75 to 0.83). The rate difference for CRC mortality was 0.13 deaths (CI, 0.07 to 0.19 deaths) per 100 persons, and the MRR was 0.80 (CI, 0.72 to 0.88). Women had less benefit from screening than men for CRC incidence (IRR for women, 0.84 [CI, 0.77 to 0.91]; IRR for men, 0.75 [CI, 0.70 to 0.81]; P = 0.032 for difference) and mortality (MRR for women, 0.91 [CI, 0.77 to 1.17]; MRR for men, 0.73 [CI, 0.64 to 0.83]; P = 0.025 for difference). There was no statistically significant difference in screening effect between persons aged 55 to 59 years and those aged 60 to 64 years. LIMITATION Data from the U.K. trial were less granular because of privacy regulations. CONCLUSION This pooled analysis of all large randomized trials of sigmoidoscopy screening demonstrates a significant and sustained effect of sigmoidoscopy on CRC incidence and mortality for 15 years. PRIMARY FUNDING SOURCE Health Fund of South-East Norway.