Wednesday, May 26, 2021

For colorectal cancer screening, 45 is not the new 50

In 2018, when the American Cancer Society (ACS) recommended lowering the age to start screening for colorectal cancer from 50 to 45, I wrote a Medscape commentary explaining why I thought this decision was "premature at best, and at worst will cause more harm than good." I was not alone in my negative assessment of the new guideline. In the Annals of Internal Medicine, Dr. Michael Bretthauer and colleagues noted the absence of clinical study evidence to support the ACS's assumption "that screening will generate the same benefits for persons younger and older than 50 years," and that the 22% increase in incidence rates of colorectal cancer in persons in their 40s from 2000 to 2013 represented an absolute increase of only 1.3 cases per 100,000 person years. Consequently, he argued, "we need to exercise caution when using models rather than direct clinical data to guide health policy decisions affecting millions." In Gastroenterology, Dr. Peter Liang and colleagues observed that the intended consequences of earlier initiation of screening (reducing colorectal cancer morbidity and mortality in younger persons) could be accompanied by unintended consequences such as diverting resources to a lower-risk population, increasing screening disparities, increasing individual and societal costs, and making it more difficult to study the effectiveness of screening in a younger population.

Unlike the U.S. Preventive Services Task Force (USPSTF) guidelines, American Cancer Society guidelines are not directly linked to insurance coverage for preventive services. Nonetheless, a study found that past-year colorectal cancer screening rates in adults aged 45 to 49 years increased from 4.8% immediately prior to the ACS guideline to 11.7% several months after, suggesting that some physicians and patients had begun following the new guidelines. (My own family physician offered colorectal cancer screening after my 45th birthday, but I declined.)

Our understanding of the pathophysiology of colorectal cancer in younger adults is still evolving. A recent analysis of the rising incidence of early-onset colorectal cancer found that an increasing proportion of tumors are carcinoid tumors, which unlike the more common adenocarcinomas are unlikely to be affected or prevented by screening. As Dr. Bretthauer wrote in an accompanying editorial, "the majority of carcinoid tumors identified by screening represent incidental findings with little health benefit from detection. In fact, many may be characterized as overdiagnosed tumors, which by definition increase the burden and harms of screening without eh balance of additional benefit."

Nonetheless, last week the USPSTF formally recommended lowering the age to start routine colorectal cancer screening to 45, based on a modeling study that suggested 22 to 27 "additional life-years gained" per 1000 persons, assuming 100% adherence. I don't distrust the conclusions of all modeling studies; for example, I support the USPSTF's decision to lower the age to start lung cancer screening from 55 to 50 years based on a model, but in that case there were also significant supporting data from randomized controlled trials in the younger age group, which isn't the case for colorectal cancer and now may never be, at least in the United States. (See screening for prostate cancer for an example of why it's critical to not implement widespread screening in primary care before you study whether or not it actually works.)

The bottom line is that both the ACS and now the USPSTF recommendations to start colorectal cancer screening at age 45 are based on epidemiologic data and decision models with a lot of shaky assumptions - the most shaky being that cancers in younger adults behave the same way as those in older adults. If a young patient wants to get screened for colorectal cancer for individual reasons and/or if a health insurer decides that there are long-term benefits to paying for it, that's one thing, but mandating first-dollar coverage for this preventive service requires stronger evidence than we have.