SAN FRANCISCO — For years, 50 years old has been the age at which screening for colorectal cancer (CRC) began in the United States, but recently one group lowered the starting age to 45.
This move by the American Cancer Society in 2018 was made in reaction to reports of an increase in the incidence of CRC in younger adults.
However, other groups have stayed with the benchmark 50. This includes the US Preventive Services Task Force and the National Comprehensive Cancer Network (NCCN).
Should the age be lowered in view of the mounting reports of an increase in CRC in younger adults? Experts argued both for and against the move here at the 2020 Gastrointestinal Cancers Symposium.
"We're having this debate because the health of more than 20 million Americans is in the balance," commented David Weinberg, MD, MSc, chairman of the Department of Medicine at Fox Chase Cancer Center in Philadelphia, Pennsylvania. "This is not just an academic discussion."
If the screening age shifts to 5 years earlier, the impact nationally would be about 30,000 colorectal cancers and 11,000 deaths averted.
"It will take about 11 million additional colonoscopies…and the overall bill would be $10 billion. That's not a small number, but if the country has the resources and we want to do this, I would say we can," argued Uri Ladabaum MD, director of the Gastrointestinal Cancer Prevention Program and the clinical chief of the Division of Gastroenterology and Hepatology at Stanford University in California.
Lower the Age
Ladabaum argued in favor of lowering the age to 45 years to start screening.
"In life, 60 may be the new 40, but for colorectal cancer screening, 45 is definitely the new 50," he said.
Anticipating arguments against such a move, Ladabaum focused on several points.
First, the magnitude of the problem is certainly not small, he noted, pointing to a 2017 study showing that colorectal cancer rates have increased by 1%-2.4% annually since the mid-1980s in persons aged 20 to 39 years and by 0.5%-1.3% since the mid-1990s in adults aged 40 to 54 years (J Natl Cancer Inst. 2017;109:djw322). Rectal cancer incidence has been increasing even more rapidly, at a rate of about 3.2% annually from 1974–2013 in adults aged 20 to 29 years.
Overall, people who were born around 1990 and later have double the risk of colon cancer (incidence rate ratio [IRR] = 2.40) and quadruple the risk of rectal cancer (IRR = 4.32) as compared to those born circa 1950.
"Thus, 45 to 49-year-olds are beginning to look like yesterday's 50 to 54-year-olds used to be," said Ladabaum.
One issue that has been raised is lead-time bias, with the burning question: Are the cancers found in adults in their 40s simply the same ones that would have eventually been detected in their 50s? Ladabaum argued that they are not, referencing a 2019 study showing that among persons aged 40 through 49 years, the disease was diagnosed at later stages (JAMA. 2019;321:1933-1934).
For 40- to 49-year-olds, there was a significant increase in incidence from 1995 through 2015. The proportion of distant cancers increased significantly (from 21.7% to 26.6%; P < .001), and the authors of the study had noted that this increase of 4.9% could not be explained by a decrease in unstaged cases.
"In the early '90s and mid-'90s, we began to see an increase in all stages," Ladabaum noted. "And the most important thing here is the distant cancers over time. They've gone up."
If the only explanation was lead time bias in 40- to 49-year-olds, then a person screened and diagnosed with cancer at age 48 would have earlier stage disease than if it had been found at age 51. "So is this all lead-time bias?" he said. "I think the answer is no."
Next, Ladabaum tackled the issue of whether benefit/risk ratio of CRC screening is different among younger vs older adults. This is difficult to tease out, he suggested, as the data are sparse and there are no controlled studies to date to address that. One study from Taiwan, which looked at the outcomes of fecal immunochemical testing (FIT), showed that in different age groups, the hazard ratio for detecting cancer in those with positive results is higher in younger people vs older ones (J Clin Gast. 2016;50:761).
"Indirect evidence shows that if we do a FIT test and it's positive, it probably means something," he said. "But is there something magical that at age 50 and older — that it becomes a screenable disease, and through age 49 it's not screenable? I would say no. Biology is not like this."
Finally, is it cost effective to start earlier? In a modeling projection published last summer by Ladabaum and colleagues, starting at 45 years would avert about 4 colorectal cancers and 2 colorectal cancer deaths per 1000 people, and the cohort would gain approximately 14 quality-adjusted life years (Gastroenterology. 2019;157:137-148).
"The incremental cost per quality-adjusted life year gained is highly acceptable," said Ladabaum. "This is well within the range of what's considered cost-effective in the United States — under $35,000 for colonoscopy, and under $8000 for fecal immunochemical tests."
Therefore, the answer is yes, it is cost effective, he concluded.
Not So Fast…
Arguing the case against a lowering of the starting age, Weinberg agreed with Ladabaum that colon cancer risk for younger people is rising.
That risk has increased from 5.9 per 100,000 to 7.2 per 100,000, which is a relative increase of 22%, he noted.
"That is what newspapers will want to put on their headlines to sensationalize it — that the risk of colon cancer for young people is up by over 20%," said Weinberg. "But that represents an absolute risk of just 1.3 more people per 100,000. Put in some context, 99.9% of people in their 40s will not develop colon cancer."
This observation was not to "make light of the remainder," he emphasized, but "the overwhelming majority are not going to get this disease at this age," he noted.
"It's also not entirely clear that starting screening at 45 is the right answer," he added.
"Taken to a somewhat ridiculous extreme," he continued, "why not start at 40? You'll catch more people that way, but no one is advocating that."
A better understanding of the factors contributing to the increased CRC incidence seen in younger adults is very important, he argued, and he suggested that some of it may be due to screening, along with factors such as obesity, diabetes, and childhood exposures.
Modeling has been done to calculate the risk and benefit of early screening, but while useful for decision making, models are "usually wrong and sit near the bottom of the evidence hierarchy."
"Models help inform decisions, but they don't define the standard of care," Weinberg said. "No one would vaccinate a population based on a model."
Weinberg also emphasized that the new recommendation from the American Cancer Society to start screening at age 45 years is qualified. This means that while "clear evidence of benefit exists," so does uncertainty about whether "the benefits really outweigh the harms."
The current evidence for reducing the screening age is not yet clear, he believes, and he questioned the premise that age should be the only criteria for cancer stratification.
Weinberg cited one study that looked at early onset colon cancer (ages 18-49 years) and compared it to two groups: patients diagnosed at 50 years and older and matched controls (Clin Gastroenterol Hepatol. 2019 Oct 14). Besides age, the study authors identified several nonmodifiable risk factors that were associated with early-onset disease, including gender, race, history of inflammatory bowel disease (IBD), and family history of colorectal cancer.
"Being male was a risk factor, and having a family history increased your risk by three times," said Weinberg. "And I would note that this study had removed people, at least as best one can, who were known to have syndromic risks of early onset colon cancer (such as familial adenomatous polyposis (FAP)."
Earlier screening (by age 40) should already be taking place in people with such syndromes, he commented, as well as those with a known family history and IBD. "Those recommendations were there before the ACS, and we don't necessarily need another one," he added.
To make things a little more confusing, a second recent study, using National Cancer Data Base data from 2004 to 2015, identified another set of factors associated with colon cancer in younger adults (Cancer. 2019;125:3828-3835). This study showed diagnosis under 50 years rose only in non-Hispanic white men, in Hispanic and non-Hispanic white women, and in those living in urban vs rural areas.
"But it gets more interesting," Weinberg pointed out. "Risk increased over time for people in the highest zip code income quartile and those with private insurance, and risk was lower for people with Medicaid and no insurance at all," he noted. "Well, that smacks of access to me," he commented.
Another issue is the possibility of lead-time bias. From 1975 through 2015, incidence rose over time, according to Surveillance, Epidemiology, and End Results (SEER) data (J Natl Cancer Inst. 2017;109:djw322). Screening of persons under the age of 49 years also more than doubled, from a low level in 2000 of about 6% to over 15% by 2010. As screening increases, the incidence increases, Weinberg pointed out. "But mortality doesn't change. And despite what Ladabaum said a moment ago about lead-time bias, that is textbook lead-time bias in any epidemiology study."
Finally, it is essential to carefully weigh the benefit against the risk, Weinberg said.
A core principle of population screening is to create more future health benefits than harms, and if the screening age is lowered, several million additional colonoscopies will be performed.
Colonoscopy reduces the colorectal cancer mortality risk by about 75%, and the incidence of the disease is 7.2 per 100,000 in the younger age group. "But colonoscopy-specific mortality — just having the test — is associated with a death rate of 7 per 100,000," Weinberg pointed out. "Let's not forget that there is a risk associated with this procedure." (Gastrointest Endosc. 2011;74:745-752.)
Weinberg emphasized that everyone wants to reduce the burden of cancer, and models are helpful for that purpose. "They're obviously thought-provoking, but they're not adequate to drive change without additional evidence of clinical and cost-effectiveness," he said. "These are important questions that need better data."
He added that without changing the current screening protocol, "we could certainly emphasize more than ever the impact of family history and IBD on colon cancer risk and colon cancer prevention."
"And certainly, there's plenty of evidence that patients with a known family history of colon cancer are not getting screened at the right age or with the right frequency," Weinberg concluded. "We can do better. All of us can do better."
Weinberg has disclosed relationships with Fujifilm and Exact Sciences. Ladabaum has disclosed relationships with Lean Medical, Universal Dx, Clinical Genomics, Medtronics, Modus GI, and Quorm Consulting
Gastrointestinal Cancers Symposium (GICS) 2020. Breakout: The Screening Age for Colonoscopy Should Be Reduced — A Debate.
Presented January 25, 2020.
