Older patients with Barrett's esophagus or a history of colorectal neoplasia should not automatically continue endoscopic surveillance, according to a new AGA clinical practice update, which advises weighing overall health, life expectancy, ability to undergo treatment, and patient preference.
Published in Clinical Gastroenterology and Hepatology, the expert review provides nine best practice recommendations on when to continue or stop surveillance. AGA commissioned the update, which underwent internal and external peer review. The recommendations are based on published research and expert opinion, rather than a systematic review, and were not formally rated for the quality of evidence or strength of the recommendations.
“The US population is aging, and with the numerous GI conditions for which people are under endoscopic surveillance, timing was right to evaluate our current state of practice and think about if, when, and how to stop surveillance based on balance of benefits/harms,” first author Audrey H. Calderwood, MD, MS, FACG, FASGE, professor of medicine at the Geisel School of Medicine and the Dartmouth Institute, Lebanon, New Hampshire, told GI & Hepatology News.
For physicians, the key is to consider a patient’s overall health and likely life expectancy, rather than age alone. Most cancer screening guidelines recommend routine screening through age 75, individualized decisions from ages 75 to 85, and stopping after 85, the authors noted. Dr. Calderwood and colleagues noted that patients generally need a life expectancy of at least 10 years to gain a survival benefit from cancer screening. Prognostic tools that account for age, frailty, physical function, and other health conditions can help guide these decisions.
Key points from the update include:
For patients with Barrett’s esophagus, decisions about continued surveillance should consider other health conditions, life expectancy, procedure risks and patient preferences. About 90% of patients with Barrett’s esophagus do not have dysplasia, and their risk of developing esophageal adenocarcinoma is about 0.1%-0.3% each year. In one study of patients successfully treated for Barrett’s-related neoplasia, the risk of dying from other causes was 40 times higher than the risk of dying from esophageal adenocarcinoma. A cost-effectiveness analysis suggested that the best age to stop surveillance may range from 69 years for women with severe comorbidities to 81 years for men without comorbidities.
Surveillance should stop at any age if a patient would not be able to undergo endoscopic therapy, surgery or cancer treatment if neoplasia were found. A life expectancy of less than five years and multiple health conditions may also support stopping surveillance. Risk assessment tools may help guide decisions, the authors wrote, although their role in practice remains unclear. One tissue-based test, for example, identified 38% of patients who progressed to high-grade dysplasia or cancer and correctly classified 94% of those who did not.
For colorectal cancer screening after age 75, physicians should weigh the potential benefits against the risks of colonoscopy. Among patients ages 76-85 who underwent screening, colorectal cancer was found in only about 0.2%-0.5%, while advanced adenomas were found in 6%-8%. Colonoscopy caused an estimated 14-28 complications per 1,000 older patients. Overall, the update estimated about 26 complications per 1,000 colonoscopies, compared with colorectal cancer found in only two to three patients per 1,000.
Polyp history can also help guide decisions. Among patients older than 75 with low-risk adenomas, the risk of developing colorectal cancer within five years was about 0.5%. For those with advanced adenomas on their most recent colonoscopy, the five-year risk was 0.8%. By comparison, the risk of dying from causes other than colorectal cancer was about 22% within five years and 48% within 10 years.
The authors recommend that physicians document discussions with patients about the benefits and risks of colonoscopy and avoid routinely scheduling colonoscopy without a consultation after age 75. However, evidence on the best ways to stop surveillance is limited, and more long-term research is needed on outcomes after repeated colonoscopies.
Koushik K. Das, MD, reported consulting for Olympus. Gary W. Falk, MD, MS, reported consulting relationships with Adare/Ellodi, Bristol Myers Squibb, EsoCap, Exact Sciences, CDx, Castle Bioscience, Uniquity, Regeneron/Sanofi, Takeda, Upstream Bio, and Lucid, and equity in Merck, Bristol Myers Squibb, Pfizer, and AbbVie. The other authors reported no conflicts.
Expert Insight
Dr. Calderwood spoke with GI & Hepatology News about what the guidance means for clinical practice.
Which life expectancy or frailty tool do you recommend physicians use routinely when deciding whether to stop surveillance?
Dr. Calderwood: I really like ePrognosis because it is easily accessible online and only takes a minute to input readily available data on the patient to get an estimate on life expectancy at 5, 10, and 14 years for ambulatory older adults. Super helpful!
How should the recent negative randomized trial influence Barrett’s esophagus surveillance in otherwise healthy adults over 75?
Dr. Calderwood: The Barrett’s Oesophagus Surveillance Versus Endoscopy at Need Study (BOSS) showed no differences in overall survival, cancer-specific survival, time to diagnosis of esophageal adenocarcinoma or cancer stage with every 2-year surveillance endoscopy vs. endoscopy for symptoms only in patients with Barrett’s esophagus. As we note in the clinical practice update, the study did have some limitations including inclusion of patients without intestinal metaplasia (~25% of the cohort) and higher frequency of endoscopy in the surveillance group than suggested by current guidelines, leading the 2025 AGA clinical practice guideline on Barrett’s esophagus to still recommend performing endoscopic surveillance over no surveillance. However, I do think the absence of benefit of surveillance here is useful when thinking about management of surveillance in adults 75 and older where competing mortality is increased. For patients age 75 and older without any esophageal dysplasia or in those with short-segment disease, it may very well be reasonable to stop Barrett’s surveillance or at least give patients the option to stop.
For adults over 75 with prior advanced adenomas, when would you continue surveillance colonoscopy versus discontinue it or consider FIT?
Dr. Calderwood: I would continue surveillance colonoscopy if their life expectancy is greater than 10 years and the advanced adenoma was on their most recent colonoscopy or if their most recent colonoscopy was not high quality (meaning adequate bowel prep and complete to the cecum with long enough withdrawal time).
For adults 75 and older with prior advanced adenomas, I would discontinue surveillance colonoscopy if their most recent colonoscopy was high quality, their life expectancy is less than 10 years and the advanced adenoma was in the distant past or if the patient has difficulty with colonoscopy, is not interested in surveillance or has other priorities to tend to.
There isn’t much data regarding FIT for use of surveillance in patients with recent advanced adenomas though I suspect more data will be forthcoming in the future. I would consider use of FIT in patients over 75 with non-advanced adenomas who are interested in surveillance and able to undergo and tolerate colonoscopy but who would prefer an alternate surveillance method (provided they would still be willing to have a colonoscopy if the FIT is positive). My team is conducting a RCT for older adults age 65-82 with recent low risk polyps comparing annual FIT vs. one time colonoscopy and hope to provide information on effectiveness and patient reported outcomes once the trial concludes.
As you and your coauthors put together this update, was there a topic, or perhaps more than one, that caused more deliberation than usual?
Dr. Calderwood: We spent a fair amount of time discussing scenarios in which surveillance could be stopped reasonably vs. scenarios in which we thought more discussion would be good to allow for individualized decision-making and preferences. It is challenging to make definitive BPAs in the absence of robust long term outcome data on older folk with Barrett’s accounting for participation in surveillance and older folks with a history of colon polyps who have been undergoing serial surveillance colonoscopy over their lifetime. I am happy that this clinical practice update will at a minimum raise awareness about stopping surveillance in these conditions and highlight factors to be considered that can be used in discussions with patients.