Commentary by CRC Screening Experts: Dr. Dennis Ahnen & Dr. Aasma Shaukat
This week, the Canadian Task Force on Preventive Health Care issued guidance on colorectal cancer screening in primary care which recommends against colonoscopy for routine screening of asymptomatic patients. These recommendations differ from those published by the U.S. Preventive Services Task Force (USPSTF) in 2008, which are currently being updated, that support the use of FOBT, flexible sigmoidoscopy, or colonoscopy for colorectal cancer screening in adults aged 50-75.
To put the Canadian guidelines in perspective, two physicians with significant interest and expertise in colorectal cancer prevention offer their personal views and share insights: Dr. Dennis Ahnen and Dr. Aasma Shaukat.
Dennis J. Ahnen, MD, FACG, University of Colorado
Dr. Ahnen is Professor of Medicine at the University of Colorado. He serves on the Steering Committee for the National Colorectal Cancer Round Table and as Co-Director of the University of Colorado Hereditary Cancer Center.
The Canadian Task Force focused on colorectal cancer mortality outcomes and only considered incidence of late-stage colorectal cancers (stage III and IV) in their analyses, so this minimized any benefits of screening in preventing early stage colorectal cancer. This approach may have led them to discount the evidence that has found that colonoscopy reduces colorectal cancer incidence. In my opinion, reducing incidence of early stage colorectal cancer is also an important outcome of screening.
We should recognize that the U.S. is a major outlier in the colorectal cancer screening world. The U.S. is a very resource-rich country, and one of only a couple of countries that even include colonoscopy as a screening recommendation. It is also the only country in which colonoscopy is the dominant screening strategy. We need to admit and accept that colonoscopy is very expensive and associated with risks; it is by far the most expensive and highest risk cancer-screening test ever recommended. We should not be surprised when other countries make recommendations that differ from ours. I think that an important goal for the U.S. GI community is to make colonoscopy better and less variable in quality, to minimize colonoscopy risk and to make colonoscopy less expensive –particularly by not adding cost that hasn’t been shown to be of benefit (anesthesia as opposed to conscious sedation) and by not performing colonoscopy more frequently than guideline-recommended surveillance intervals.
Have the Canadians relied on strong evidence in these new CRC guidelines?
Yes. The Canadians actually required stronger evidence of effectiveness that the USPSTF. The Canadian Task Force put very high emphasis on randomized controlled trials and less on economic modeling, whereas the USPSTF put substantial evidence on micro-simulation modeling.
The lack of randomized, controlled trials demonstrating effectiveness of colonoscopy on colorectal cancer mortality was a major reason cited by the CTF for not recommending colonoscopy. In addition they were concerned about what they perceived as a high level of uncertainty about the harms of colonoscopy when widely used in a population screening program. The USPSTF relied on modeling and its assumptions/estimates of benefit and risks to a substantial extent for its recommendations. Finally, the Canadian Task Force relied on relative resource requirements and availability of screening tests to a greater extent than the USPSTF.
How do these recommendations compare to the U.S. Preventive Services Task Force?
They differ substantially. The CTF included a recommendation for standard g-FOBT because the RCTs of FOBT have used them rather than FIT; they do note the increased sensitivity of FIT and high-sensitivity g-FOBT, and include them as possibly improved options. The USPSTF clearly recommends FIT and HS g-FOBT over standard FOBT.
The Canadian Task Force recommends FOBT every other year, whereas the USPSTF recommends annual FOBT. The rationale for the Canadian Task Force is that all but one of the randomized, controlled trials that have shown a benefit for FOBT have used every-other-year testing. The one trial that compared annual vs. every-other-year testing showed a higher impact on CRC mortality in the annual group, but did not show that the difference between the annual and every-other-year groups was statistically significant.
The Canadian Task Force recommends flexible sigmoidoscopy every 10 years, whereas the current USPSTF recommends every 5 years with an interval FOBT. The current draft of the USPSTF’s revised guidelines does include flexible sigmodoscopy every 10 years, but with an annual FOBT. Again, the USPSTF relied on modeling for their recommendations while the Canadian Task Force relied on the results of the randomized, controlled trials.
The main point of this commentary is that the Canadian Task Force did not recommend colonoscopy, whereas the USPSTF did include colonoscopy as a screening option. The rationale for the Canadian Task Force decision included the lack of direct evidence of colonoscopy efficacy by RCTs, the high level of resources required for colonoscopy screening, the long wait times in Canada for colonoscopy, and the low quality of evidence of the harms of routine colonoscopy in general and specifically in Canada.
Clearly, there are other relevant factors vis à vis Canadian vs. U.S. health system – how to account for those differences in terms of these recommendations?
I think this is correct. The U.S. does not have a National Health Service with a budget that limits the cost of overall health care expenditure as they do in Canada (as a result, the cost of health care is much lower in Canada and services are rationed for all rather than by circumstance, as it is in the U.S.). The Canadian Task Force is expected to take cost and resource availability into account in its recommendation, whereas the USPSTF does not focus on overall cost as much as cost effectiveness models. There are a lot of other differences between the countries that might play into this.
Aasma Shaukat, MD, MPH, FACG, University of Minnesota
Dr. Shaukat is the ACG Governor for Minnesota. She is the Section Chief, GI Section, Minneapolis VA Health Care System, an Associate Professor at the University of Minnesota, and a member of the Minnesota Evidence-based Practice Center.
Colorectal cancer is common and lethal. It is the third most common cancer in both men and women and the second most common cause of cancer deaths, accounting for about 8% of all cancer deaths in the United States. Our neighbors to the North share our burden of colorectal cancer. The incidence and mortality rates of colorectal cancer in Canada are similar to those in the U.S. (Incidence rate of 49 per 100,000 and mortality rates of 17 per 100,000). Screening for colorectal cancer in Canada is primarily FOBT based.
I read the new Canadian Task Force colorectal cancer screening guidelines with particular interest, and the first thing I was curious about was why their rating of the evidence, using similar literature and GRADE criteria, was different from U.S. Preventive Services Task Force.
How do these recommendations compare to the U.S. Preventive Services Task Force?
The Canadian Task Force (CTF) recommendations differ from USPSTF in that they require high-quality evidence (large, well-powered RCTs) for their recommendations. Lack of evidence is considered lack of efficacy in most cases, particularly for expensive invasive therapies. For example, if there is lack of evidence for a certain therapy, the USPSTF would rate it ‘insufficient,’ but the CTF would rate it a ‘D’, i.e., they would recommend against it.
The USPSTF will consider data from observational studies (all that we have for colonoscopy) if these are well done, while the CTF gives it much lower considerations or discounts it altogether.
The USPSTF funds modeling studies for their guideline development and use the results heavily in their recs. In both the 2008 and 2015 guidelines, CISNET models provided the framework for net benefit/harm analysis. The CTF does not use modeling studies for their guidelines.
Another difference is that the CTF recommends gFOBT while the USPSTF recommends high-sensitivity FOBT or FIT. gFOBT is the test used in all the RCTs of CRC screening, while high-sensitivity FOBT or FIT has better performance characteristics even though these have never been tested in an RCT.
The value framework of the two task forces are also different. The USPSTF reflects that we value aggressive strategies that may enhance early detection at the expense of more harms accrued by the invasive strategy. CTF reflects the value that they are comfortable that many more preventable cancers may develop, but there will be fewer harms from over-diagnosis or invasive testing.
Another difference is that the CTF ranks screening 60+ year-olds higher than screening 50–59-year-olds. The incidence and mortality rates for CRC are much lower in 50–59-year-olds, and hence the net benefit is smaller than 60+ year olds. The three screening studies the CTF considered for this recommendation (including long term follow up of the Minnesota FOBT trial) support greater benefit in this age group. However, the USPSTF sets the threshold of initiating screening much lower (at age 50) because of its perceived “value” of net benefit of screening in that age group.
Clearly, there are other relevant factors vis à vis Canadian vs. U.S. health system – how to account for those differences in terms of these recommendations?
For their recommendations, both task forces take the applicability of recommendations into account. With national healthcare and no infrastructure for screening colonoscopy, the applicability of colonoscopy as a primary screening modality is much lower for Canada.
Colonoscopy screening is resource intensive and opportunistic, i.e., requires individuals to come in contact with the healthcare system to get screened, while stool-based testing is programmatic, i.e., tests can be mass mailed and done in the comfort of one’s home. For a national healthcare system, programmatic approaches are far more appealing and effective at achieving high screening rates. In the United States, limited examples of systems that have achieved (and passed) the 80% benchmark for CRC screening, such as the VA and Kaiser system, use stool-based programmatic approaches heavily. Ultimately, both the U.S. and Canada have the same goal: to increase colorectal cancer screening rates. Regarding the modality, each healthcare system must make choices that work for their system, based on resources, and the values they assign to the risks versus benefits of each strategy.
The Canadian Task Force guidelines are very clear in their logic as to why they rated evidence the way they did and what it would take to change the recommendations. They state clearly that for colonoscopy to be a primary screening modality, and to justify its costs and harms, they would like to see evidence in which colonoscopy is directly compared to an effective, cheaper and less invasive strategy of FOBT and come out ahead. Four such trials are underway (including one among U.S. Veterans), but in the meantime, lack of evidence means lack of efficacy for the Canadians.