Colonoscopy Screening Lowers Incidence But Not Mortality After 13 Years

Key Takeaways

• Invitation to colonoscopy was associated with lower colorectal cancer incidence at 13 years in this randomized population-based comparison.
• Incidence estimates were stronger in per-protocol analysis and appeared more pronounced for distal than proximal disease.
• Colorectal cancer mortality was not significantly reduced over the reported follow-up, and interpretation included lower-than-expected mortality in the no-screening group.

After 13 years, invitation to colonoscopy screening was associated with lower colorectal cancer incidence, with an intention-to-screen RR of 0.81 and risks of 1.46% versus 1.80%. Fewer colorectal cancers accrued in the invited group over long follow-up in this randomized comparison across three countries. Colorectal cancer mortality was not significantly different between the invited and no-screening groups at this follow-up. The findings reflect screening invitation rather than universal procedure completion. At 13 years, lower incidence and no significant mortality difference remained the central pattern.

The multicountry, population-based randomized controlled trial enrolled 84,583 men and women aged 55 to 64 years in Norway, Poland, and Sweden. Participants were randomly assigned in a 1:2 ratio to invitation to colonoscopy or to no screening. Colorectal cancer incidence and mortality were the primary outcomes over a planned 10-to-15-year follow-up window, with a first analysis after 10 years and later analyses at longer intervals. These results reflect 13 years.

In the intention-to-screen analysis, investigators recorded 375 colorectal cancers among 28,217 invited participants and 912 among 56,366 participants without screening invitation. Those counts corresponded to a risk ratio of 0.81 (95% CI 0.71-0.90). A separate per-protocol analysis showed an incidence RR of 0.55 (95% CI 0.33-0.81). The two analyses used different analytic frames, but both indicated lower colorectal cancer incidence after invitation.

Subgroup patterns suggested a clearer incidence reduction for distal than proximal disease, although proximal cancer rates were 129 (0.51%) versus 283 (0.56%), RR 0.91 (95% CI 0.71-1.09). Distal cancer rates were 224 (0.87%) versus 563 (1.11%), RR 0.79 (95% CI 0.65-0.89), with interaction p<0.0001. Among men, incidence was 214 of 14,154 versus 541 of 28,247, with RR 0.77 (95% CI 0.64-0.88). Among women, incidence was 161 of 14,063 versus 371 of 28,119, with RR 0.87 (95% CI 0.70-1.02) and interaction p<0.0001. These subgroup estimates indicate that incidence differences varied by tumor site and sex.

For colorectal cancer mortality, investigators recorded 106 deaths (0.41%) among 28,217 invited participants and 236 deaths (0.47%) among 56,366 participants without screening invitation. The intention-to-screen mortality RR was 0.88 (95% CI 0.68-1.08), and the per-protocol mortality RR was 0.70 (95% CI 0.26-1.25). Mortality in the no-screening group was 0.47%, compared with 0.82% expected when the trial was designed, a factor investigators cited in interpreting the absence of a clearly demonstrated mortality reduction at 13 years. The 13-year update showed reduced colorectal cancer incidence after invitation to colonoscopy without a statistically significant mortality difference.