Colorectal cancer screening: which options exist and what to expect from a colonoscopy

The letter has been lying on the kitchen table for four days. Sender: your health insurer. Subject: colorectal cancer early detection. You opened it, skimmed it and put it down again, under the electricity bill.

It is not that you are against it. You just do not want to think about it right now. And somewhere you read that this examination supposedly does less than claimed. That was a good excuse, and it held.

Many people know that pile. In my consultations I hear almost always the same half sentence: I really should.

So let us not begin with an appeal, but with what sets this cancer apart from almost every other one. In breast cancer, in lung cancer, in prostate cancer, a screening examination can find something earlier that is already there. In colorectal cancer it can do more. It can remove the precursor in the same appointment in which it sees it.

The reason for that is banal and magnificent at the same time: the tool used to look is the same one used to remove. A working channel in the endoscope, a snare, a few seconds. Screening and procedure coincide.

What this text is meant to do: lay the options side by side, reproduce the current German rules cleanly, describe the procedure in a way that takes away its dread, and express the risks in numbers instead of adjectives. Numbers carry further than reassurance.

Why this cancer of all cancers suits screening so well

Picture a road that is built over twenty years. First a dirt track, then a gravel path, eventually a carriageway. Whoever comes by after five years and clears away the gravel has not prevented a road. They have stopped the construction.

That is roughly how the adenoma carcinoma sequence runs. Out of a harmless growth of the bowel lining, the adenoma, a carcinoma can develop over a very long time. In the microsimulation model MISCAN-COLON, calibrated against the data of the American National Polyp Study, this path from the progressing precursor to the clinical diagnosis takes on average around 20 years. Of these, 16.4 years fall on the time up to the preclinical carcinoma and 3.6 years on the phase in which the tumour is present but not yet causing symptoms.

This honest counter calculation belongs right next to it, otherwise the text loses credibility. A pathology review in the Journal of Pathology notes that the transition from adenoma to carcinoma takes place in only around 5 percent of cases. The widespread formula that the sequence always takes 10 to 20 years is almost entirely extrapolated from cross sectional observations, because the clean longitudinal study, meaning leaving adenomas in place and watching, is not ethically possible.

So: the vast majority of polyps removed during a colonoscopy would never have harmed anyone. That is not an argument against the examination. It is the reason it is done at all. Because nobody can tell by looking at a single adenoma which group it belongs to.

What the long term data show

The strongest evidence that removal changes something comes from a follow-up that runs over two decades.

A reanalysis of three randomised sigmoidoscopy trials separates early detection from prevention even more cleanly. That matters, because the usual incidence figures also contain carcinomas that were already present at the time of the examination and could not have been prevented at all.

The underlying British trial itself, the UK Flexible Sigmoidoscopy Screening Trial with 170,432 people, found 26 percent fewer cases and 30 percent fewer deaths in the invitation analysis after a median of 17.1 years. Among the 71 percent who actually came for the examination it was 35 and 41 percent (PMID: 28236467). Remember that gap between the two numbers. In the next section it decides everything.

A sigmoidoscopy only looks into the last segment of the large bowel. Why the complete colonoscopy can do more is shown by a large American cohort over 22 years: after an unremarkable colonoscopy the hazard ratio for colorectal cancer lay at 0.44, after an unremarkable sigmoidoscopy at 0.60. For the right, ascending colon the risk fell measurably only after colonoscopy (HR 0.73), and not at all after sigmoidoscopy (PMID: 24047059). That is cohort evidence, not randomisation, and people who go for screening differ from people who do not. The direction is plausible nonetheless and matches what one would anatomically expect.

And now you know why the ten years between two examinations are not slackness but fit the speed at which this cancer develops.

The trial that is constantly quoted wrongly

If you come across the sentence online that a large trial showed colonoscopy achieves hardly anything, then it almost always refers to the same piece of work. It is called NordICC, appeared in 2022 in the New England Journal of Medicine, and it is a good trial. It just answers a different question than most people believe.

Imagine someone wants to measure how good a blood pressure medication is. They send a prescription to 28,000 people and count after ten years how all 28,000 are doing. About 16,000 of them never filled the prescription. The result of that calculation is an honest statement. But it is a statement about postal delivery, not about pharmacology.

That is exactly what the invitation analysis measures. It is methodologically correct and for health policy even the more important number, because programmes have to reckon with real participation. For your personal question, what a colonoscopy might achieve in you, it is the wrong number.

That this gap is no arithmetic trick can be seen in the British sigmoidoscopy trial. There 71 instead of 42 percent came, and sure enough invitation and participation analyses lie much closer together: 26 versus 35 percent for the rate of disease, 30 versus 41 percent for mortality (PMID: 28236467). The more people actually come, the more similar the two numbers become. That is not coincidence, it is arithmetic.

A methodological review places observational and interventional studies of screening colonoscopy systematically side by side and concludes that both study types, despite seemingly different results, point to strong preventive effects (PMID: 41128479). Since the first publication of the National Polyp Study in 1993, incidence and mortality in the screening age group in the USA have fallen by about a third, while they are rising worryingly among younger people.

Two honest studies can therefore seem to say different things without either of them being poor. That is not a crisis of science. That is the difference between two questions.

The options compared, honestly sorted

The question I am asked most often is not whether, but with what. Is the stool test enough. Could I not swallow a capsule. What about the CT scan.

Before the table comes, a distinction that constantly goes wrong online: a microbiome stool test is something entirely different from a faecal immunochemical test for hidden blood. One describes bacterial groups, the other looks for haemoglobin. A microbiome profile, a PCR panel or a dysbiosis report never replaces colorectal cancer screening and is not intended for it. What these tests are actually about is described in the article on stool testing, PCR and dysbiosis diagnostics.

The German S3 guideline on colorectal carcinoma in version 3.2 of March 2026 names exactly three effective procedures in statement 4.2: colonoscopy every ten years, sigmoidoscopy every five years, faecal immunochemical test every one to two years. All other procedures should not be used in the symptom free population, which is recommendation 4.3 with grade of recommendation B, the grade that corresponds to the wording "should" rather than "shall".

Two things stand out in this table. First, the rejected procedures are not bad. The capsule has decent numbers, so does CT colonography. In the review for the US Preventive Services Task Force, colonoscopy and CT colonography sat closer together for lesions from 6 mm than one would expect, 0.89 versus 0.86. So the difference does not lie in the seeing. It lies in the fact that CT cannot remove anything, that it produces incidental findings outside the bowel, and that screening with ionising radiation is not permitted in Germany. Second, they all share the same problem: they see, but they do not act. Every abnormal finding leads to colonoscopy in the end anyway, and then you had two preparations instead of one.

The stool test is not a second class fallback

That colonoscopy can do more does not mean everything else is worthless. The classic guaiac test, the predecessor of today's immunochemical test, was examined in a randomised trial in Minnesota over 30 years: 46,551 people, colorectal cancer mortality 32 percent lower with annual testing and 22 percent lower with biennial testing (PMID: 24047060). Overall mortality did not change, and that sentence is readily turned against screening. It does not carry that weight. Colorectal cancer accounts for only a small share of all deaths, and no screening trial in the world is large enough to show a difference there.

One point belongs here without exception, and it is the practically most important one in this section. A positive stool test is not a result. It is an assignment. The S3 guideline says so in recommendation 4.4 with the highest grade of recommendation A: with a positive iFOBT, the entire large bowel should be examined endoscopically.

What you are entitled to in Germany, as of 2026

Precision pays off here, because many advice pages still reproduce the rules from before 2025. You may even know those better than the current ones, because they stood everywhere for years.

So if you are 52 and a woman, and read somewhere that you would have to wait another three years: that has not been true since April 2025.

When colorectal cancer has occurred in the family

Separate rules apply to this situation, and they come earlier. The S3 guideline recommends in 5.37 that first degree relatives be offered a first colonoscopy at an age that lies at the latest ten years before the age at which the affected person fell ill, at the latest at 40. If your father fell ill at 58, that would be at 48 for you. The same rule applies analogously if adenomas were found in the family before the age of 50 (recommendation 5.39).

What appears in hardly any German text is the conversion into absolute numbers. And that is exactly what you need, because a threefold increase sounds like a catastrophe without your knowing threefold of what.

And why the USA already starts at 45

This question comes up almost every time, and it deserves a proper answer instead of a shrug. Since 2021 the US Preventive Services Task Force has issued an A recommendation for ages 50 to 75 and a B recommendation for ages 45 to 49, meaning moderate net benefit with moderate certainty (PMID: 34003218). In the USA an estimated 10.5 percent of new cases occur in people under 50, and the rate of disease in 40 to 49 year olds rose by almost 15 percent between the periods 2000 to 2002 and 2014 to 2016.

The German S3 guideline knows exactly these data and still stays with 50. Its reasoning draws on the same simulation calculations: the estimated gain in life years when starting at 45 instead of 50 was mostly below 10 percent, while the number of examinations needed rose by considerably more than 10 percent.

What does not follow from this: that an earlier examination at your own expense would be generally necessary in order to be cleverer than the German programme. Whether it can make sense in an individual case, for instance with a family history, belongs in a medical conversation and not in a blanket recommendation. And what also does not follow: that symptoms before the age of 50 would be harmless. The point is a different one. Symptoms do not lead into screening anyway, they lead into a work-up, and that knows no age threshold.

And now you know why it is worth opening that insurer's letter again instead of relying on a number that may be three years old.

What happens step by step during a colonoscopy

When people tell me they are afraid of a colonoscopy, they usually do not mean the examination. They mean the night before, the drink, the loss of control. That is why this section comes in the order in which you will experience it, and the most unpleasant part gets the most space.

Why the second portion is the more important one

This is the part I most enjoy telling, because it turns an imposition into a contribution.

The European guideline on bowel preparation, the ESGE update of 2019, sets exactly these points as the standard: split dose, a defined interval between the last portion and the start of the examination, a low fibre diet on the day before instead of pure clear liquids, and low volume preparations as an alternative for people who tolerate large quantities poorly (PMID: 31295746). Those are Europe wide procedural standards. Which preparation you receive and in what rhythm is set by the practice examining you.

Two technical questions you are allowed to ask

The first point concerns the gas. For the camera to see anything, the bowel has to be unfolded. Room air can be used for that, or carbon dioxide. CO2 is taken up through the bowel wall and breathed out through the lungs; room air has to take the way down. That is exactly where the bloated feeling afterwards comes from.

The second point concerns the technique of advancing the scope. In the water exchange method, water is introduced instead of gas and suctioned off again. A network meta-analysis of 17 randomised trials with 10,350 people found a higher adenoma detection rate for it than with air (odds ratio 1.40) and better cleanliness, at a somewhat longer insertion time (PMID: 29981753). That is a well documented procedural advantage, but availability depends on the practice, and data on hard endpoints are missing. Asking about it can make sense. It is not a verdict on practices that do it differently.

On sedation itself, Cochrane updated the comparison between propofol and the classic combination in 2025, with endpoints ranging from completeness of the examination through recovery time to respiratory and circulatory events (PMID: 41147535). The short summary for you: with propofol a deeper sedation can as a rule be achieved, which has advantages and disadvantages, and this decision belongs in the conversation with the practice examining you.

What comes afterwards depends on the findings. If polyps were removed, their number, size and histological type determine when the next check makes sense. The European ESGE surveillance guideline governs this in detail (PMID: 32572858). I deliberately do not print a table of it here, because this assessment is a medical judgement and not an arithmetic exercise. Ask about it at the results conversation, and write the date down.

The risks, in orders of magnitude instead of adjectives

Most pages carry a sentence like: complications are rare. That may be true, but it has never taken anyone's fear away, because rare is not a number. So let us take numbers.

A meta-analysis of population based studies differentiates these numbers further, and that matters for honesty. Pooled across 21 studies, the perforation rate lay at 0.5 per 1,000, the bleeding rate at 2.6 per 1,000 and mortality at 2.9 per 100,000 colonoscopies. With polyp removal the rate of post-procedure bleeding rose to 9.8 per 1,000 (PMID: 27296945).

That means: someone who has a polyp removed carries a higher bleeding risk than someone in whom nothing was found. That is the price of prevention, and it belongs stated. Remarkable at the same time is that the complication rate in screening and surveillance examinations lay lower than in diagnostic examinations. People examined because of symptoms are simply sicker.

Why quality standards exist

A colonoscopy is not simply a colonoscopy, and that is not a call for mistrust but a measurable fact.

The European professional society formally defined these key figures in 2017: caecal intubation rate, adenoma detection rate, withdrawal time, quality of bowel preparation (PMID: 28268235). A good examination is therefore not a feeling. It is defined and it is measured.

And then there is the most honest number in this article. A meta-analysis following the standardised methodology of the World Endoscopy Organization analysed 220,106 carcinomas from eight countries. Result: 7.5 percent of all carcinomas detected in the West are ones that appeared within three years after a colonoscopy. The rate fell between 2006 and 2012 from 7.9 to 6.7 percent. It lay considerably higher in inflammatory bowel disease (29.3 percent), in diverticular disease (11.6 percent) and in the right colon (8.6 percent) (PMID: 39209191).

Who should not wait for screening

There is a distinction that blurs in everyday language and is decisive in medicine: the distinction between screening and work-up.

Screening is an offer to people who feel healthy. It follows a calendar, operates with probabilities and is therefore plannable, because by definition no symptom is pressing. Plannable does not mean postponable, though: an offered screening appointment belongs kept and not deferred. A work-up is something else. It begins with a symptom and follows no calendar but the symptom.

If you have blood in your stool, you do not belong in screening. You belong in a work-up, and promptly. The same applies to the other signs from the box further up: black tarry stool, unintended weight loss, symptoms that wake you at night, a new persistent change in bowel habit from around age 45 to 50, anaemia without explanation.

On anaemia a sentence of its own, because it is often overlooked. Iron deficiency without a recognisable reason is, in men and in women after the menopause, a finding that first needs an explanation and not just an iron tablet. That expressly does not mean that a medically prescribed iron therapy should be stopped or changed on your own. It only means that it does not replace the search for a cause. What else may lie behind it is described in the article on iron deficiency and absorption problems in the gut. The search for the source of bleeding still comes first, and it belongs in medical hands.

With diarrhoea that does not stop, the list of possible causes is long, and some of them are harmless. Which those may be and how to sort them is described in the article on chronic diarrhoea. But the order remains: first rule out the red flags, then the finer search for causes.

And one more group belongs here: people with an inflammatory bowel disease. Separate surveillance intervals apply to them, which have nothing to do with the general screening programme and belong in specialist hands. The 29.3 percent from the analysis just quoted show why that is no formality. How Crohn's disease and ulcerative colitis can be accompanied in an integrative way is described in the article on inflammatory bowel disease.

What counts outside screening, with honest effect sizes

Here is the point at which an integrative practice normally gets loud. I do the opposite, because the numbers demand it.

Lifestyle and screening are not alternatives. They play in different orders of magnitude. A factor that shifts the risk by 12 percent is something other than an examination that may prevent two out of three carcinomas in one bowel segment. Both together are better than one alone. But the one does not replace the other.

On the individual points there is deliberately no second explanation here, because they are covered in detail elsewhere. Why the debate about processed meat is more differentiated than headlines suggest is described in the article on meat, cancer and the difference between origin and processing. What the famous 30 grams of fibre are actually about is clarified in the article on fibre myths. And what else alcohol may do in the gut is described in the article on alcohol and the gut.

And now the humility alongside it

If I left this list standing without context, the article would be dishonest. Because most of the much discussed dietary associations do not hold up under strict scrutiny.

The guideline stays consistent on food supplements too. Recommendation 3.6 says that there are currently no secure data on prevention of colorectal carcinoma through micronutrients and that supplements should not be recommended for this purpose. Recommendation 3.7 applies to acetylsalicylic acid, COX-2 inhibitors, statins and hormones: likewise not for prevention.

Anyone who wants to work on the gut more broadly will find the larger frame in the overall concept of the gut reset. Only, that frame belongs alongside screening and never in its place. A gut restoration, a microbiome build-up or a change of diet are good things. None of them looks inside your large bowel.

This is a point at which I, as an integratively working physician, expressly go along with the guideline. There are fields in which I consider micronutrients underrated. Cancer prevention is not one of them. Whoever takes a capsule and leaves the letter lying is trading a documented route for an undocumented one.

The difference is an appointment

To close, the number that occupies me most of all, because it tells nothing about colorectal cancer and everything about people.

Between the 82.7 percent and the 43.3 percent lies no difference in character. There lies no lack of insight, no lack of reason and no lack of information. There lies a phone call.

That is why this article does not end with an appeal but with an observation. Whoever keeps the matter in their head postpones it. Whoever has an appointment in the calendar goes. And if the thought gives you a queasy feeling: that is normal, that is not an argument, and you are allowed to say it out loud in the preliminary talk.

And now you also know why I began this text with the image of a line in which one mark is missing. That is exactly the message. A development was interrupted, and afterwards time simply carries on quietly. That is all. And it is quite a lot.

Frequently asked questions

Where this topic connects to others

Screening does not stand on its own. It hangs on dietary patterns, on how red flags are handled, on the question of how a person deals with an unpleasant letter.