Adults aged at least 75 years with adenomas detected at prior colonoscopy had higher subsequent risks of colorectal cancer (CRC) and CRC-specific death than those without adenomas, although cumulative risks were low and substantially outweighed by competing risks of non-CRC death, according to a retrospective cohort study published in JAMA.
Decisions about continuing CRC surveillance in older adults are increasingly complex, as current guidelines do not provide clear age- or frailty-based recommendations for deimplementation. This gap persists despite procedural risks and uncertain benefit in those with prior adenomas, for whom surveillance is typically performed for early detection and prevention of cancer.
“Knowing when to stop CRC surveillance in elderly patients is challenging in part because of the concern of missing clinically significant pre-cancerous lesions,” Ziad F. Gellad, MD, MPH, AGAF, told GI & Hepatology News.
The 10-year cumulative incidence of non-CRC death ranged from 46.9% to 48.4% and was thus found to exceed that of CRC (1.1%) and CRC-specific death (0.5%) in patients with prior adenomas, compared with 0.7% and 0.4%, respectively, in those without adenomas.
“Older adults may consider deprioritizing surveillance colonoscopy relative to other health concerns,” lead author Samir Gupta, MD, MSCS, of Veterans Affairs San Diego Healthcare System and the University of California, San Diego, and colleagues wrote.
Study details
The researchers focused on 91,952 older adults (median age, 71 years at last colonoscopy; 98% male) who underwent colonoscopy between January 1, 2006, and December 31, 2019, and prior to 75 years of age within the U.S. Department of Veterans Affairs. Of this population, 25,538 (27.8%) had prior adenomas and 66,414 (72.2%) did not.
The study estimated the cumulative incidence of CRC, CRC-specific death, non-CRC death, and all-cause mortality among patients with and without adenomas detected at prior colonoscopy. Comparisons of CRC incidence and associated death between groups were performed using Gray's test; in patients with prior adenomas, these outcomes were further stratified by Veterans Affairs Frailty Index categories (nonfrail [≤ 0.10]; prefrail [0.11 to 0.20]; mild frailty [0.21 to 0.30]; moderate frailty [0.31 to 0.40]; and severe frailty [> 0.40]), reflecting increasing risk of all-cause mortality.
10-year outcomes
After 10 years of follow-up, the cumulative incidence of CRC was higher in patients with prior adenomas than in those without (1.1% vs. 0.7%; Gray's test P < .001).
The cumulative incidence of CRC-specific death was 0.5% in the adenoma group vs. 0.4% in the nonadenoma group (Gray's test P = .005).
Non-CRC death had a cumulative incidence ranging from 46.9% to 48.4%. In patients with prior adenomas, the cumulative incidence of non-CRC death was reported to substantially exceed the incidence of CRC across all frailty strata, ranging from 34.2% in the nonfrail group to 82.0% in the severe frailty group.
“The study is an important addition to the literature because it highlights the low risk of colorectal cancer in this previously screened population, even in those with prior adenomas,” Gellad said. “The study also elegantly contextualizes that risk by comparing it to the much higher risk of non-CRC death.”
“These findings raise major questions about the clinical relevance of surveillance colonoscopy in older adults with prior adenomas after reaching 75 years,” the authors concluded. “Collectively, our results also highlight the importance of an ongoing clinical trial comparing de-escalation of surveillance with a strategy of annual fecal immunochemical testing vs. usual care surveillance colonoscopy among older adults with prior history of polypectomy.”
Limitations noted by the authors include the predominantly male U.S. veteran population, which may limit generalizability, particularly given sex differences in CRC risk and life expectancy. Exploratory analyses showed no significant differences in follow-up colonoscopy frequency between patients with prior adenomas who did vs. did not develop incident CRC and could not disentangle risk reduction attributable to surveillance vs. diagnostic colonoscopy. The authors also wrote that more granular analyses using manually abstracted colonoscopy data are planned for future work.
Gellad shared additional commentary about the study's scope. “There remain some important limitations to this study including that the analysis was limited to males and thus may not be representative of the general population,” he said. “It also excluded those with sessile serrated adenomas which are clinically relevant pre-cancerous lesion that may have a different impact on cancer risk than adenoma.”
“These limitations notwithstanding, these findings can provide some reassurance to gastroenterologists when they decide to stop surveillance in elderly patients,” Gellad said.
Full disclosures of the study authors are available in the published study.