First surveillance colonoscopy at five years is non-inferior to three after high-risk adenomas

Design

Multinational, randomised non-inferiority trial (interim analysis)

Population

10,799 patients after removal of high-risk adenomas

Primary outcome

Cumulative colorectal cancer incidence (10-year primary; 5-year interim)

Effect

0.77% vs 0.82% at 5 years; upper bound 0.68 within 0.7-point margin

EPoS II randomised 10,799 patients across eight European countries who had high-risk adenomas removed — at least one adenoma of 10 mm or more, high-grade dysplasia or villous features, or three to ten adenomas — to a first surveillance colonoscopy at 5 years or at 3 years (the current guideline interval).

In this planned interim analysis at 5.5 years, cumulative colorectal cancer incidence was 0.77% in the 5-year group and 0.82% in the 3-year group (difference −0.05 percentage points). The upper bound of the confidence interval was 0.68, inside the non-inferiority margin of 0.7. Stage at diagnosis looked similar, and colorectal cancer deaths were few (3 vs 2).

If confirmed at 10 years, this frees a large share of colonoscopy capacity, which matters most in settings where endoscopy lists are long. The interim result already supports not pulling these patients forward. Final guideline changes will wait for the 10-year analysis.

• For patients with high-risk adenomas completely removed, a first surveillance colonoscopy at 5 years was as safe as at 3 years.
• Do not bring surveillance forward beyond guideline intervals out of caution — it adds procedures without detected benefit.
• Ensure the index colonoscopy was complete and of good quality; the trial assumes clearance.
• Patients with new symptoms need investigation regardless of the surveillance schedule.
• Guidelines have not yet changed; this is an interim analysis of a 10-year trial.

Why it matters

It challenges the default three-year interval and could release a large volume of colonoscopy capacity.

Don't overread it

This is an interim analysis at 5.5 years; the primary end point is 10-year incidence and guidelines have not changed.

The statistics, in plain English

Non-inferiority asks whether the longer interval is 'not meaningfully worse'. The margin was 0.7 percentage points of cancer incidence; the worst-case estimate was 0.68, just inside it. Because this is an interim look, a stricter confidence level (99.12%) was used to avoid false reassurance. The final 10-year result could still move.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.