Reviewed for medical accuracy by Dr Evelyne Jonniaux, Medical Advisor
Last updated 17 September 2026
Colonoscopy: 19% fewer cancers, mortality not yet demonstrated
That is what NordICC shows, the only randomised trial of screening colonoscopy, across 84,585 people followed for thirteen years (Kaminski, The Lancet, 2026). Colorectal cancer incidence falls from 1.80% to 1.46%, a relative risk of 0.81 with an interval from 0.71 to 0.90. Mortality falls from 0.47% to 0.41%, relative risk 0.88 with an interval from 0.68 to 1.08: not significant. The authors write that a single colonoscopy significantly reduces incidence but not mortality over thirteen years. These figures describe populations, not one patient. Noria Health arranges a full digestive work-up near Brussels within 24 to 48 hours.
What NordICC measured, and what it does not measure
NordICC is a pragmatic trial: people aged 55 to 64 were invited to a single colonoscopy and compared with those who received no invitation. The decisive point is that only 42% of those invited accepted.
| At 13 years | Invited group | No invitation | Effect |
|---|---|---|---|
| Colorectal cancer incidence | 1.46% | 1.80% | RR 0.81 (0.71 to 0.90) |
| Colorectal cancer mortality | 0.41% | 0.47% | RR 0.88 (0.68 to 1.08) |
| Incidence, per-protocol analysis | RR 0.55 (0.33 to 0.81) | ||
| Mortality, per-protocol analysis | RR 0.70 (0.26 to 1.25) | ||
| Right-sided colon cancer | 0.51% | 0.56% | RR 0.91 (0.71 to 1.09), not significant |
| Left-sided colon cancer | 0.87% | 1.11% | RR 0.79 (0.65 to 0.89) |
Two caveats belong with these figures, and they point in opposite directions.
First, the trial is underpowered for mortality. Mortality observed in the unscreened group, 0.47%, is far below what was anticipated at design, 0.82%. The absence of a signal may come from lack of power as much as from lack of effect.
Second, the figure of a 50% mortality reduction that circulates everywhere comes from the per-protocol analysis of the 2022 publication. The authors themselves reran it with a method less sensitive to residual confounding (Shi, Eur J Epidemiol, 2025) and obtain relative risks of 0.71 to 0.79, with intervals crossing 1 comfortably. At thirteen years the per-protocol analysis gives 0.70 with an interval from 0.26 to 1.25. That figure is not robust, and we do not repeat it.
What is solid is the effect on incidence: 19% in intention to screen, 35% to 45% among those who actually have the examination.
Sigmoidoscopy did show an effect on mortality
The pooled analysis of the four randomised sigmoidoscopy trials (Juul, Ann Intern Med, 2022), 274,952 people followed for fifteen years, gives a 20% mortality reduction, with a rate ratio of 0.80 and an interval from 0.72 to 0.88.
But the benefit is localised. In the American PLCO trial, mortality reduction is 50% for left-sided cancers and nil for right-sided ones, relative risk 0.97. And in the Norwegian NORCCAP trial, no benefit was observed in women: hazard ratio for mortality 1.01 against 0.63 in men.
The faecal test, and what nobody says
No randomised trial of the faecal immunochemical test against no screening has published a mortality result. The Spanish COLONPREV trial compares the faecal test with colonoscopy, not with nothing: at ten years mortality is 0.24% with the test and 0.22% with colonoscopy, a risk difference of -0.02 with an interval from -0.10 to 0.06. The American CONFIRM trial will not report before 2028.
The only mortality data against no screening come from the older guaiac tests.
| Trial | Number | Follow-up | Mortality reduction |
|---|---|---|---|
| Minnesota | 46,551 | 30 years | 32% annual screening, 22% biennial |
| Nottingham | 152,850 | 19.5 years | 13% (3% to 22%) |
| Funen | 61,933 | 9 rounds | 11% |
In the Minnesota trial, all-cause mortality does not move: relative risk 1.00 with an interval from 0.99 to 1.01. And in both Nottingham and Funen, cancer incidence did not fall.
The quality of the examination changes everything, and it varies enormously
This is the most actionable finding in the file, and the least known to patients. Corley’s study (NEJM, 2014), across 314,872 colonoscopies performed by 136 gastroenterologists, measured each one’s adenoma detection rate: it ranges from 7.4% to 52.5%.
| What is measured | The figure |
|---|---|
| Interval cancer, lowest detection quintile | 9.8 per 10,000 person-years |
| Interval cancer, highest quintile | 4.8 per 10,000 person-years |
| Hazard ratio, highest against lowest quintile | 0.52 (0.39 to 0.69) |
| Fatal interval cancer | 0.38 (0.22 to 0.65) |
| Effect of each additional point of detection rate | -3% risk, HR 0.97 (0.96 to 0.98) |
The relationship is continuous, with no threshold. The figure of 25% often presented as a standard is a consensus, not a biological threshold: the higher the operator’s detection rate, the lower the risk, across the whole observed range. A German confirmation across 822,715 people finds five-year cancer incidence 69% higher after polypectomy by doctors in the lowest detection quartile.
In NordICC itself, about 30% of endoscopists did not meet quality standards. That is a possible explanation of the result, and the authors mention it.
The complications, in numbers
| Complication at 30 days | Pooled rate |
|---|---|
| Perforation, all colonoscopies | 0.5 per 1,000 (0.4 to 0.7) |
| Perforation with polypectomy | 0.8 per 1,000 (0.6 to 1.0) |
| Bleeding after colonoscopy | 2.6 per 1,000 (1.7 to 3.7) |
| Bleeding after polypectomy | 9.8 per 1,000 (7.7 to 12.1) |
| Mortality | 2.9 per 100,000 (1.1 to 5.5) |
These come from a meta-analysis of 21 population studies (Reumkens, Am J Gastroenterol, 2016). Bleeding fell from 6.4 to 1.0 per 1,000 between 2001 and 2015; perforation and mortality stayed flat. In NordICC, across 11,843 colonoscopies, there were 15 major haemorrhages, no perforations and no screening-related deaths at 30 days.
One honest caveat: administrative registries underestimate these complications. A Norwegian chart review across 11,205 colonoscopies finds the sensitivity of billing data for detecting bleeding ranges from 36% to 89% depending on the centre.
Interval cancer, and a received idea to correct
Cancer arising between two colonoscopies is real. Applied strictly, the international reference methodology gives a rate of 7.5% with an interval from 6.4% to 8.7%, across 220,106 colorectal cancers and 18,148 interval cancers (Kader, Clin Gastroenterol Hepatol, 2025). That rate is falling over time.
The idea that these cancers are more aggressive, however, is refuted. A secondary analysis of the NORCCAP trial across 98,684 participants compares 163 interval cancers with 1,740 clinically detected cancers in unscreened people: mortality is identical, adjusted hazard ratio 0.98 with an interval from 0.72 to 1.35.
What the studies do not allow us to claim
That colonoscopy reduces mortality by 50% is not robust data, for the reasons set out above. Prefer the effect on incidence, which is.
That the faecal immunochemical test reduces mortality by 20% to 30% has no trial behind it. The figure is extrapolated from the guaiac trials, which give 11% to 32% depending on the trial and screening intensity. It has to be stated as an analogy, not a measurement.
That polypectomy prevents 76% to 90% of cancers does come from the National Polyp Study, but by comparison with three historical cohorts, without randomisation. The design has to be stated with the figure.
Finally, no single figure for incomplete colonoscopies is defensible: caecal intubation rate varies from 96% to 98% among the same operators depending on the definition used, and there is no harmonised method of calculation.
In practice, if this concerns you
Three questions structure the discussion. What is the adenoma detection rate of the endoscopist who will examine you, since that is the one modifiable parameter associated with interval cancer risk. Does your family history justify an examination earlier than population screening. And if a polyp is removed, what surveillance interval is proposed and on what argument.
Noria Health arranges a full digestive work-up near Brussels within 24 to 48 hours: consultation, colonoscopy with supervised preparation, histology of anything removed, and further investigation where the examination justifies it. The report is issued within 24 hours, follow-up is structured at D14, D30 and D90, and surgery is considered only around day 10 if it is needed.
What Noria Health does not do: present a colonoscopy as a guarantee. Heavy gastrointestinal bleeding, severe abdominal pain with fever, or a complete stop of stool and gas are matters for the emergency services in your country, immediately, not for an appointment.
Further reading
- Hernia, gallbladder, reflux: when to operate and when to wait
- Metabolic surgery, what the trials measure
- Heart: 44% of people with hypertension do not know they have it
Sources
- Kaminski MF, Bretthauer M, Loberg M, et al. Colonoscopy screening and colorectal cancer incidence and mortality, 13-year results of NordICC. The Lancet, 2026;407:1787-1795. https://pubmed.ncbi.nlm.nih.gov/42102826/
- Bretthauer M, Loberg M, Wieszczy P, et al. Effect of colonoscopy screening on risks of colorectal cancer and related death. N Engl J Med, 2022;387:1547-1556. https://pubmed.ncbi.nlm.nih.gov/36214590/
- Shi J, Loberg M, Kalager M, et al. Instrumental variable analysis of the NordICC trial. Eur J Epidemiol, 2025;40:419-425. https://pubmed.ncbi.nlm.nih.gov/40278966/
- Juul FE, Cross AJ, Schoen RE, et al. Effectiveness of colonoscopy screening vs sigmoidoscopy screening. Ann Intern Med, 2022;175:1525-1533. https://pubmed.ncbi.nlm.nih.gov/36215714/
- Atkin W, Wooldrage K, Parkin DM, et al. Long term effects of once-only flexible sigmoidoscopy screening. The Lancet, 2017;389:1299-1311. https://pubmed.ncbi.nlm.nih.gov/28236467/
- Schoen RE, Pinsky PF, Weissfeld JL, et al. Colorectal-cancer incidence and mortality with screening flexible sigmoidoscopy, PLCO. N Engl J Med, 2012;366:2345-2357. https://pubmed.ncbi.nlm.nih.gov/22612596/
- Holme O, Loberg M, Kalager M, et al. Long-term effectiveness of sigmoidoscopy screening, NORCCAP. Ann Intern Med, 2018;168:775-782. https://pubmed.ncbi.nlm.nih.gov/29710125/
- Castells A, Quintero E, Bujanda L, et al. Colonoscopy versus faecal immunochemical test in colorectal cancer screening, COLONPREV. The Lancet, 2025;405:1231-1239. https://pubmed.ncbi.nlm.nih.gov/40158525/
- Shaukat A, Mongin SJ, Geisser MS, et al. Long-term mortality after screening for colorectal cancer. N Engl J Med, 2013;369:1106-1114. https://pubmed.ncbi.nlm.nih.gov/24047060/
- Scholefield JH, Moss SM, Mangham CM, Whynes DK, Hardcastle JD. Nottingham trial of faecal occult blood testing, 20 year follow-up. Gut, 2012;61:1036-1040. https://pubmed.ncbi.nlm.nih.gov/22052062/
- Kronborg O, Jorgensen OD, Fenger C, Rasmussen M. Randomized study of biennial screening with a faecal occult blood test. Scand J Gastroenterol, 2004;39:846-851. https://pubmed.ncbi.nlm.nih.gov/15513382/
- Corley DA, Jensen CD, Marks AR, et al. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med, 2014;370:1298-1306. https://pubmed.ncbi.nlm.nih.gov/24693890/
- Schwarz S, Hornschuch M, Pox C, Haug U. Polypectomy rate and interval colorectal cancer. Int J Cancer, 2023;152:1547-1555. https://pubmed.ncbi.nlm.nih.gov/36444495/
- Reumkens A, Rondagh EJ, Bakker CM, Winkens B, Masclee AA, Sanduleanu S. Post-colonoscopy complications, a systematic review and meta-analysis. Am J Gastroenterol, 2016;111:1092-1101. https://pubmed.ncbi.nlm.nih.gov/27296945/
- Kader R, Cid-Mejias A, Brandao P, et al. Post-colonoscopy colorectal cancer rates using the World Endoscopy Organization methodology. Clin Gastroenterol Hepatol, 2025;23:519-530. https://pubmed.ncbi.nlm.nih.gov/39209191/
- Jodal HC, Loberg M, Holme O, et al. Mortality from post-screening colorectal cancer. Gastroenterology, 2018;155:1787-1794. https://pubmed.ncbi.nlm.nih.gov/30165051/
- Siau K, Hodson J, Valori RM, et al. Variability in caecal intubation rate reporting. Gastrointest Endosc, 2019;89:1026-1036. https://pubmed.ncbi.nlm.nih.gov/30659830/
See also: all our articles on digestive surgery.
Have your case reviewed. Noria Health arranges a specialist consultation near Brussels and confirms feasibility within 24 to 48 hours. Open a case.