2 new colorectal cancer screening guidelines

Two recent guidelines sugggest changes in colorectal cancer (CRC) screening:

the International Agency for Research on Cancer (IARC) of the World Health Organization just published a brief updated summary of their recommendations for (see colon ca screening WHO nejm2018 in dropbox, or http://www.nejm.org/doi/full/10.1056/NEJMsr1714643?query=pfw&jwd=000101577173&jspc=GP ​. The full report will follow as volume 17 of the IARC Handbooks of Cancer Prevention, to be published soon. 

Details:

--stool-based tests for occult blood:

    --guaiac testing: 2 studies found sufficient evidence that screening every 2 years (without rehydrating the cards) reduces colorectal cancer (CRC) mortality, with relative risk reductions of 9-14%. but 3 RCTs found that screening every 2 years did not reduce the incidence of CRC 

    ​--FIT testing: no RCTs. observational studies were consistent: around a 10-40% CRC mortality reduction. sufficient evidence for screening be every 2 years (some of the impetus here is that guaiacs have pretty clear data showing benefit, and that FIT is a better test)

--endoscopic testing:

    --sigmoidoscopy: 4 large RCTs found relative risk reduction for CRC was 18-28%, and CRC mortality was 22-31% lower in 3 of the studies. in one of the studies there was 17 year follow-up, confirming a 26% reduction in CRC incidence and 30% reduction in CRC mortality (see http://gmodestmedblogs.blogspot.com/2017/04/one-time-flex-sig.html for a review of this study), and a recent meta-analysis of observational studies found a 50% risk reduction of both CRC incidence and mortality. There is sufficient evidence that a single screen reduces CRC incidence and mortality

    --colonoscopy: no RCTs (though 4 are underway); recent meta-analysis of observational studies found a 70% risk reduction of both CRC incidence and mortality, with effect more evident in distal than proximal colon. there is sufficient evidence that a single screen reduces CRC incidence and mortality. harm of colonoscopy: bleeding and perforation each at 0.01-0.05% of exams

    --CT colonography: limited evidence of effectiveness

--comparison of tests (basically done by network meta-analyses, a mathematical comparison of outcomes from different studies used when there are no head-to-head comparisons):

    --sigmoidoscopy is better than guaiacs in reducting CRC incidence but not mortality

    --colonoscopy is more effective than sigmoidoscopy and guaiacs in reducting CRC mortality (low quality evidence)

    ​--detection of advanced neoplasia by FIT testing every 2 years for 5 rounds was similar to one-time colonoscopy (see http://gmodestmedblogs.blogspot.com/2017/12/fit-testing-long-term-study.html for analysis of this study, as well as link to Canadian guidelines which suggest FIT every 2 years)

--------------------------------------------------------------------------------

The US Multi-Society Task Force (MSTF) on Colorectal Cancer screening recommendations, including several GI organizations, ranked colonoscopy every 10 years and FIT annually as the top-tier recommendations (see colon ca screening AGA2017 in dropbox, or http://gi.org/wp-content/uploads/2017/06/ajg2017174a.pdf​ ): for average risk patients, screening tests can be offered as one of  3 choices (they do not support one approach over others): the multiple options listed below, a sequenced approach (suggest a tier 1 approach, then go to tier 2 or 3 per patient wishes), or risk-stratified screening (preferentially recommending FIT in those with estimated low prevalence of advanced neoplasia, or colonoscopy in those at higher risk.)

Details:

--first-tier approaches

    --colonoscopy every 10 years. no RCTs, but has high sensitivity. one cohort study and 3 case-controlled studies found 80% reduction in CRC incidence and mortality in distal CRC and 40-60% reduction in proximal CRC in the US and Germany. and, even 1 or 2 negative exams seems to confer lifetime protection. But  colonoscopy is operator-dependent, needs thorough bowel cleaning, risks aspiration pneumonia, post-procedural bleeding (largely related to polypectomy), and a small risk of splenic injury: overall 0.5/1000 risk of perforation, 2.6/1000 risk of bleeding, and 2.9/100K risk of death. "colonoscopy is preferred approach to management of any benighn colorectal polyp regardlss of size or location".

         --if offering colonoscopy, patients should ask/know the colonoscopist's adenoma detection rate (should be >30% for men and >20% for women), cecal intubation rate (should be >90% overall, >95% for screening), and use of split-dose bowel preps (at least 1/2 of the prep should be ingested on the day of the procedure). and we clinicians should make sure to get pictures of the colon and that the quality of the bowel prep was described

    --FIT testing annually. noninvasive, 1-time sensitivity of 79% for cancer/30% for advanced adenomas, low cost ($20), but poor sensitivity for serrated class precursor lesions (though there is no evidence that cancers arising from these are less likely to bleed: ie, FIT testing may still work well). positive test results need to be followed by colonoscopy. They feel this is ideal test after a first negative colonoscopy, though "should likely always be one of the tests included in  a multple-options approach"

    ​--if following MSTF sequential approach (one of the 3 options): offer colonscopy first, with backup of FIT per patient preference (again, they give equal credence to all 3 approaches)

--second-tier approaches:

     --CT colonography every 5 years. lower risk of perforation than colonoscopy, sensitivity of 82-92% for adenomas >1cm. but, still need bowel prep to have adequate sensitivity, misses smaller polyps/flat lesions/serrated lesions, radiation exposure, no studies showing decreassed CRC incidence or mortality, and, unless there is a setup for same-day colonoscopy if positive findings on colonography, then another bowel prep later

     --FIT-fecal DNA test every 3 years: 1-time sensitivity of 92% for CRC, 40% sensitive for sessile serrated polyps, but much lower specificity (88%) vs FIT tests (96%), and higher cost

     --flex sig every 5-10 years: confirmed reduction of CRC incidence and mortality in the distal colon by 29-76%. lower cost and risk than colonoscopy, less aggressive bowel prep, no need for sedation. endoscopic screening is more effective in the left side of colon than right (raising the relative benefit of sigmoidoscopy relative to colonoscopy), so they suggest that 10-year interval is acceptible since it is for colonoscopy. they do note that patient satisfaction is lower with sigmoidoscopy since no sedation, though i would add that benefits also include lower risk from sedation, and that the patient can drive themselves home afterwards

 --third-tier approach:

     --capsule colonoscopy every 5 years: approved by FDA to image proximal colon when prior colonoscopy incomplete, though not approved for routine screening. detects adenomas >6mm (88% sensitivity) but not serrated lesions. involves even more extensive bowel prep. 

 --they suggest routine screening of average risk patients begin at age 50, but do support screening African-Americans at age 45 (limited evidence, though as been promoted by some guideline committes for awhile)

 --stop screening at age 75 or if have <10 years of life expectancy and negative prior screen (esp if done by colonoscopy); consider screening up to age 85 in people without prior screening, depending on their comorbidities

 --colonoscopy screening every 5 years if family history CRC or documented advanced adenoma in first-degree relative age <60 yo or 2 first-degree relatives at any age, beginning 10 years before the youngest affected relative or age 40 (whichever is earlier).  they do reinforce that he greatest relative risk of CRC seems to be in people <50yo if they have first-degree relative with CRC <50yo (hence the need to identifythem and screen more aggressively) and that if the patient reaches age 60 without abnormal colonoscopy findings, they can revert to normal screening intervals, per a recent study (at least they can be "offered the option of expanding the interval between exams")

    --some suggestion that colonoscopy is better than FIT (nonsignificant trend in comparative study, though they recommend colonoscopy despite "weak recommendation, very-low quality evidence"), so perhaps FIT should be offered particularly in those declining colonoscopy. they lump in this same group those with 1 or more first-degree relatives with documented advanced serrated lesion (>10mm, or with dysplasia). 

--if only 1 first-degree relative with CRC or advanced adenoma at age >60, should be offered screening at age 40, but the options for screening and recommended intervals are the same as for average-risk patients. 

Commentary:

--colorectal cancer is the 3rd most common cancer in men and 2nd most common one in women, 10% of the annual global cancer incidence

--increased risk is associated with increased levels of economic development (established risk factors being consumption of processed meas, alcohol, smoking, excess body fat; decreased risk with high dietary fiber, dairy products, and more physical activity)

--5-year survival rates are 60% in high-income countries, but 30% in low-income ones.

--my sense is that CRC screening seems to be the most effective screening test we have. already, at least partially attributable to increased CRC screening, the incidence of CRC has fallen 3-4% per yr in the US in persons >50yo (though increasing in those <50yo, but with low relative incidence).

--the MSTF does raise the untested suggestion that African-Americans begin screening at age 45 based on mathematical modeling (not recommended by USPSTF) because African-Americans have a higher incidence rates of CRC, earlier mean age at onset, worse survival and more late-stage presentation, and higher proportion of CRC in those <50yo (2 member societies endorse starting at age 45, the Am College of Physicians at age 40)

--in support of sigmoidoscopy (see blog referenced below), it is cheap and easy/less of a prep, and less risky. and, a huge observational study in the UK which followed patients for 17 years found overall CRC incidence reduced by 26% and mortality by 30% (distal CRC mortality by 46%, and up to 66% in the per-protocol analysis). no difference in proximal cancer mortality (which is beyond the sigmoidoscope), and about 20% of patients with proximal lesions have nothing visible distally (ie, would be considered to have had a normal sigmoidoscopoy). but, for reasons that escape me, and as alluded to above, several studies have found that excision of these right-sided lesions does not affect mortality (eg see Baxter NN. Ann Intern Med  2009; 150:1 ).

--the issue of family history is important. any family history of CRC increases the risk, though there is a risk gradient with a first-degree relative <60yo leading to increased and earlier screening.

so, these recommendations basically add a few potential changes to our approach to CRC screening.  MSTF provides clear delineation of preferred screening tests (colonoscopy and FIT testing, the latter being really easy and noninvasive, and the same sample can be used for a 1-time H pylori antigen screen, to boot), though they do say that pretty much any approach is okay. And the IARC suggests FIT testing every 2 years (based on a good observational study showing benefit), as has already been adopted in Canada. Though i do agree with the 2-yearly screening interval, i should add that since the USPSTF and MSTF both suggest annual screening with FIT, there may be some medicolegal risk of moving to the 2-yearly protocol. Also, there is now more consensus on screening African-American patients beginning at age 45, which does seem to make sense despite the lack of a rigorous study documenting benefit. and, given the increasing incidence of CRC in those <50yo, we clinicians should probably be inquiring more regularly about family history to see if the patient may be more than just average-risk.

other past blogs on CRC screening

-- http://gmodestmedblogs.blogspot.com/2017/07/racial-disparities-in-interval.html  which notes quite striking racial differences in interval cancers in those screened

-- http://gmodestmedblogs.blogspot.com/2016/10/colonoscopy-screening-in-elderly.html for a review of Medicare data on colonoscopy screening those 70-79yo

-- http://gmodestmedblogs.blogspot.com/2015/04/risks-of-colonoscopy-in-older-people.html, an older blg on the increased risk of colonoscopy in those up to age 80-84

-- http://gmodestmedblogs.blogspot.com/2019/04/uspstf-colorectal-cancer-screening.html for USPSTF recommendations

-- http://gmodestmedblogs.blogspot.com/2017/04/one-time-flex-sig.html for article supporting the 1-time sigmoidoscopy

​