colonoscopy research
October 15, 2022

Preventing Chronic Disease

Confusion over results of a recent colonoscopy trial

A recently published NEJM paper appears to show that colonoscopy screenings do not reduce risk of colorectal cancer death. As is often the case, the results have been taken out of context and their implications grossly overestimated.

Peter Attia

Read Time 5 minutes

From the doctor’s office to television PSAs to billboards, recommendations for colonoscopy screenings are everywhere in the United States. Colonoscopy is the gold standard screening modality for colorectal cancer (CRC) – the second leading cause of cancer-related deaths after lung cancer – and the CDC and American Cancer Society recommend colonoscopy screenings every 10 years for those between the ages of 45 to 75.

So when a New England Journal of Medicine paper came out last weekend appearing to show that colonoscopy screenings do not reduce risk of colorectal cancer death, it immediately sparked a swarm of questions, commentary, and confusion. But don’t go tearing down the billboards just yet – as I explained briefly on social media earlier this week, the results of this study have been taken out of context and their implications grossly overestimated.

About the Study

The study, known as the Nordic-European Initiative on Colorectal Cancer (NordICC) trial, aimed to investigate the effect of colonoscopy screening on the risk of CRC diagnosis and death. Men and women ages 55-64 were enrolled in Poland, Norway, Sweden, and the Netherlands and were randomized in a 1:2 ratio into two groups: those in the “invited group” (n=28,220) were invited to have a one-time colonoscopy screening at no charge, while those in the “usual-care group” (n=56,365) received no such invitation.

The authors point out that, due to a lack of screening programs in these areas during the time of the study, participants would not have had the opportunity to undergo colonoscopy screening outside of the trial procedures. Median follow-up time was 10.0 years (interquartile range: 9.9-10.0 years); however, this recent publication marks only the first planned analysis point, and the study remains ongoing.

When considering the full intention-to-treat cohort, the investigators report that invitation to undergo colonoscopy screening was associated with an unexpectedly modest (18%) reduction in risk for CRC diagnosis relative to the usual-care group (RR=0.82; 95% CI: 0.70 to 0.93). Further, no significant difference was observed between the invited group and usual-care group in deaths attributable to colorectal cancer (RR=0.90; 95% CI: 0.64 to 1.16) or in deaths attributable to all causes (RR=0.99; 95% CI: 0.96 to 1.04).

These fairly dismal results have thus fueled heated debates over the value of colonoscopy as a screening tool, particularly given the costs associated with these procedures and screening programs. Parenthetically, of the 11,843 colonoscopies performed in this study there were zero perforations and 15 (0.127%) bleeding complications. If nothing else, this reinforces how safe colonoscopy is in the hands of experienced endoscopists. 

Screening vs. Invitation to Screen

The results mentioned above compare those who did not undergo colonoscopy with those who were invited to have the procedure – not with those who necessarily did. Indeed, the study authors report that only 42% of participants in the invited group actually followed through with screening. When restricting their analysis to this protocol-adherent sub-group (i.e., conducting a “per-protocol” analysis), the investigators found much larger disparities between groups.

Relative to the usual-care group, the invited group demonstrated significantly lower risk of both CRC diagnosis (1.22% incidence in usual-care vs. 0.84% in invited; RR=0.69; 95% CI: 0.55 to 0.83) and CRC mortality (0.30% vs. 0.15%; RR=0.50; 95% CI: 0.27 to 0.77) at 10 years follow-up.

Efficacy & Effectiveness: Prioritizing Populations vs. Individuals

Critics of colonoscopy screening programs argue that these per-protocol analyses are less relevant than intention-to-treat analyses – even an intervention that can prevent 100% of cancer deaths is useless if no one chooses to receive it – and from a public health perspective, they’re not wrong.

These critics are suggesting that efficacy – how well an intervention performs under ideal conditions (including perfect adherence) – is not, on its own, sufficient to inform decisions about the best ways to allocate resources for various public health agendas.

Effectiveness – how well an intervention performs in real-world conditions in which patients are free to make their own decisions – is also critical for understanding the overall impact an intervention will have on the population level. (For more on efficacy vs. effectiveness, take a look at my previous newsletter discussing these two metrics.)

That logic is reasonable when considering a whole population – the primary concern of policy-makers and epidemiologists, but it does little when considering any individual patient – the concern of healthcare providers and the individuals themselves. For any given person, the top priority is one’s own health – not optimal distribution of government resources – and the probable decisions of the rest of the population are irrelevant for one’s personal health choices (at least for non-communicable diseases such as cancer).

In other words, when it comes to a single patient’s decision on whether or not to undergo screening, efficacy is a far more important consideration than effectiveness.

Can a single trial test efficacy and effectiveness?

Clinical trials are generally suited to test for efficacy or effectiveness, but not both. The NordICC trial was designed to test effectiveness, and while per-protocol analyses can hint at efficacy, it’s important to recognize that such estimates are very crude due to inherent flaws in such secondary analyses.

As an effectiveness study, the NordICC trial followed a fairly robust design: a randomized trial with a large cohort in which the independent variable under study (invitation to undergo colonoscopy vs. no invitation) was entirely within the control of the investigators. But as an efficacy trial, virtually all of these strengths are lost. For per-protocol analyses, the variable under study is not the invitation to undergo screening but rather screening itself, and whether or not participants in the invited group chose to undergo colonoscopy was up to the participants themselves. This is to say, the independent variable was no longer fully controlled by the investigators and was not randomly assigned.

The usual-care group remained barred from colonoscopy screenings, but within the invited group, the decision to undergo the procedure was likely subject to considerable confounder effects, as those who chose to get screened may have had higher-than-average risk of colorectal cancer in the first place. Screening compliance rates were higher among males than females and among the 60-64 age group than the 55-59 age group – both reflecting established risk factors for colorectal cancer (for example, CRC rates in men are roughly 30% higher than in women, according to the American Cancer Society).

And as Drs. Jason Dominitz and Douglas Roberston point out in their commentary on the study, those in the invited group who did not undergo screening exhibited lower cancer detection rates at follow-up than those in the usual-care group. These observations suggest that those who did choose to be screened may have done so out of greater concern about risk – perhaps due to mild symptoms or family history of CRC – and if this is the case, the per-protocol cohort would have likely displayed higher CRC rates than would have been observed if screening had truly been randomly assigned. In other words, the per-protocol analyses would have underestimated the true efficacy of colonoscopy screenings.

So what should we take from these results?

The NordICC trial was designed to generate information on effectiveness, not efficacy. If you happen to work for the CDC or other public health organizations, the trial results are perhaps worth taking into consideration for determining the optimal allocation of funds to various health initiatives.

But if instead you are an individual concerned above all with maintaining your own health and that of your loved ones – or if you are a healthcare provider responsible for advising individuals on their own health – these data have very little relevance, and if anything, they provide encouraging (albeit crude) evidence that colonoscopy screenings do significantly reduce risk of CRC diagnosis and death, and they are very safe in experienced hands.

So keep scheduling those regular screenings –  even if the billboards and PSAs become fewer and farther between.

For a list of all previous weekly emails, click here

Podcast | Website | AMA

Disclaimer: This blog is for general informational purposes only and does not constitute the practice of medicine, nursing or other professional health care services, including the giving of medical advice, and no doctor/patient relationship is formed. The use of information on this blog or materials linked from this blog is at the user's own risk. The content of this blog is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Users should not disregard, or delay in obtaining, medical advice for any medical condition they may have, and should seek the assistance of their health care professionals for any such conditions.
Facebook icon Twitter icon Instagram icon Pinterest icon Google+ icon YouTube icon LinkedIn icon Contact icon