A Second Opinion on Colonoscopy

The Second Opinion Guide to Colonoscopy — No. 3 in the series

In October 2022, the New England Journal of Medicine published the first randomized controlled trial of colonoscopy screening ever conducted. It had taken more than three decades to run. NordICC enrolled 84,585 healthy adults aged 55 to 64 from Norway, Sweden, and Poland. One third were invited to a single screening colonoscopy. Two thirds were assigned to usual care with no invitation. After ten years, the risk of dying from colorectal cancer was 0.28 percent in the invited group and 0.31 percent in the usual-care group. The difference did not reach statistical significance. The risk of dying from any cause was 11.03 percent in the invited group and 11.04 percent in the usual-care group. The intervention did not extend life.

That was the ten-year data. Some in the field argued ten years was too short. Bretthauer’s team returned in May 2026 with a 13-year follow-up, presented at Digestive Disease Week and published in The Lancet. The extra three years did not produce the mortality benefit the field had projected. It was not there.

This is the guide about what that means.

The only randomized trial

For decades, colonoscopy screening was recommended to hundreds of millions of adults on the basis of observational data, not trial data. Observational studies had projected that colonoscopy could reduce colorectal cancer mortality by around 60 percent. The largest pooled estimate was 62 percent. When the randomized trial was finally done and the mortality result came in below one sixth of that estimate, and did not reach statistical significance, the response inside American gastroenterology was informative. Douglas Rex publicly noted that the mortality finding was underwhelming. David Lieberman called the lack of impact on mortality “interesting.” That is the word practitioners use when a randomized trial produces a null result on the endpoint the intervention was designed to hit and they are looking for a way to keep practicing it.

To translate the incidence numbers into what they mean per person: out of every 1,000 people invited to screening, approximately 12 will be diagnosed with colorectal cancer whether they are invited or not, and approximately 10 will be diagnosed if they are invited. To prevent one case of colorectal cancer over ten years, 455 people have to be invited to the procedure. To prevent one death from it, the number is much larger and not statistically established, because the trial did not find one.

The critique that emerged in American gastroenterology focused on uptake. Only 42 percent of those invited underwent the procedure. If you restricted the analysis to people who actually had the colonoscopy, you got a 31 percent reduction in cancer incidence and something like a 50 percent reduction in cancer death. That is what the press releases pointed to. But that analysis is not truly randomized. The people who accepted the invitation differed systematically from those who declined, in the same ways observational studies were criticized for: healthier, wealthier, more engaged with medicine. Attributing their better outcomes to the procedure rather than to the traits that predicted their acceptance reintroduces the confounding that randomization was supposed to eliminate. The defense of the intervention rests on the analytical move that made the earlier evidence unreliable in the first place.

The polyp paradox

Between 32 and 50 percent of average-risk adults over 50 have polyps found during colonoscopy. Approximately 5 percent will develop colorectal cancer at some point in their lives. If every colon cancer came from a polyp, and polyps were found in 40 percent of screened adults, then only about one polyp in eight would have been destined to become cancer. Seven out of eight polyps removed would have posed no clinical threat. The person carrying them would have lived out her life without knowing they were there.

This is what medicine calls overdiagnosis. In an editorial published in Gastroenterology in September 2018, the field’s own leading researchers wrote plainly that although most colorectal cancers do arise from polyps, the proportion of polyps that would progress to clinically symptomatic cancer if not removed is low. Most polyp removals, they wrote, bring no clinical gain to the patient. These were the researchers with the best data in the field, writing in their own leading journal, using their own words. The word they chose was low.

If you were told before your colonoscopy that any polyp found would be removed to prevent cancer, and if the removal produced no clinical gain in seven out of eight cases, then the informed consent you gave was based on a description of the intervention that its own researchers would call misleading.

The blind spot

The 2018 editorial was titled “Overdiagnosis in Colorectal Cancer Screening: Time to Acknowledge a Blind Spot.” Its authors were Mette Kalager and Michael Bretthauer of the University of Oslo, Paulina Wieszczy and Michal Kaminski of Warsaw, Iris Lansdorp-Vogelaar of Rotterdam, and Douglas Corley of Kaiser Permanente. These are not marginal figures. Kalager and Bretthauer went on to lead the NordICC trial. Corley wrote the paper that established the metric now used to evaluate every American endoscopist. Lansdorp-Vogelaar leads one of the three simulation modeling programs that inform US screening guidelines.

They wrote that overdiagnosis in colorectal cancer screening is not mentioned in patient information or guidelines, and not addressed in scientific or public debate. They called it a blind spot. That was eight years ago. Patient information materials in the United States still do not mention overdiagnosis in colorectal cancer screening. The USPSTF guidelines still do not model it as a harm equivalent to what is documented for mammography and prostate-specific antigen testing. The public debate the editorial called for has not occurred.

Consider what this means at the point of clinical delivery. The letter you receive from your primary care office invites you to schedule a screening colonoscopy. It does not tell you that the only randomized trial found no mortality benefit at 13 years. It does not tell you that seven out of eight polyps removed would never have progressed to cancer. It does not tell you that non-invasive alternatives exist and are used as the first-line screen in most of the developed world. Your doctor does not tell you these things because her patient information materials do not tell her to. The field’s leading researchers publicly asked for the conversation. The field did not have it. Your doctor was not told to have it with you.

What the guide covers

The guide is 70 pages. Part I builds the case against routine screening entirely from establishment sources: the researchers who ran the trial, the editors of the leading gastroenterology journal, the population complication studies, the microbiome literature, and the physicians who worked inside the system and left. Part II presents the terrain alternative and the practical mechanics: how to think about colon health, what to do if you already have polyps, what non-invasive monitoring is available, and what the decision looks like once you have seen the material.

Contents

Part I. The Case Against Routine Screening

• Chapter 1: The Blind Spot

• Chapter 2: The Only Randomized Trial

• Chapter 3: The Polyp Paradox

• Chapter 4: What the Procedure Does to Your Body

• Chapter 5: The Business Model

• Chapter 6: The Self-Manufacturing Cycle

Part II. The Terrain Alternative

• Chapter 7: What Colon Health Actually Requires

• Chapter 8: If You Already Have Polyps

• Chapter 9: Monitoring Without the Procedure

• Chapter 10: The Decision

Appendices

• Appendix A: Questions for Your Doctor

• Appendix B: Further Reading

---

This is the third in the Second Opinion Guide series. The first, on the HPV vaccine, is permanently free and can be read here. This one, on colonoscopy, is below for Paid Subscribers.

The stakes are your health. The decision is yours.

---