Surveillance Colonoscopy Findings in Older Adults With a History of Colorectal Adenomas

This cross-sectional study estimates advanced neoplasia yields during surveillance colonoscopy procedures among adults aged 70 years or older with a history of colorectal adenoma.


Introduction
5][16][17][18] However, guidelines provide little direction regarding the age at which colonoscopy surveillance is unlikely to be of substantial benefit and could be stopped. 18,19Given the increasing aging population in the US and that nearly 5.6 million adults older than 75 years will undergo surveillance annually by 2024, 20 estimating the yield of surveillance colonoscopy is important for understanding the balance between potential benefits and known risks of colonoscopy with advancing age.
The risks of colonoscopy increase with age, particularly among those aged 75 years or older, and include heart attack, stroke, sedation-related adverse events (eg, aspiration pneumonia), bleeding, infection, and perforation. 21In addition, the benefits of surveillance colonoscopy in older adults may be reduced because of a more limited life expectancy. 19Also, in many settings, colonoscopy demand exceeds capacity, and therefore, it is important to direct procedures to those for whom potential benefits will likely outweigh possible harms.These arguments against surveillance colonoscopy in older adults must be weighed against findings that rates of CRC increase with age, at least among unscreened individuals. 22 weighing the pros and cons of surveillance colonoscopy in older adults, information needed for shared decision-making between patients and clinicians includes the yields of CRC and advanced neoplasia at surveillance colonoscopy in this age group.[25][26][27][28][29] To address this knowledge gap, we evaluated the surveillance colonoscopy yields of CRC and advanced neoplasia in patients 70 to 85 years of age with a prior adenoma finding from a large, demographically diverse, community-based US health care system.Yields were estimated overall (all ages combined), by age group (70-74, 75-79, and 80-85 years), and by the combination of age group and prior adenoma finding (advanced adenoma vs nonadvanced adenoma).

Study Design
This cross-sectional study evaluated surveillance colonoscopy yields of CRC and advanced neoplasia in patients 70 to 85 years of age with a prior adenoma finding.This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.The study was approved by the Kaiser Permanente Northern California (KPNC) institutional review board with a waiver of informed consent because the research involved no more than minimal risk to participants and it could not practically be carried out without the requested waiver.

Study Setting
Study data were obtained from KPNC, and details of the population and screening practices have been described elsewhere. 30The KPNC membership is demographically diverse and similar in socioeconomic characteristics to the region's diverse census demographics, including the proportions of individuals with commercial insurance, Medicare, and Medicaid. 31Thus, studies within this setting approximate community-based research within a demographically diverse population. 32

Study Participants
KPNC health plan members were eligible for the study if they were 70 to 85 years of age; underwent a surveillance colonoscopy between January 1, 2017, and December 31, 2019; and had a prior colonoscopy with an adenoma detected (hereafter, "index colonoscopy") 12 or more months before their surveillance colonoscopy, colonoscopy and pathology reports available for each procedure, and at least 1 year of health plan enrollment prior to the surveillance procedure.Individuals were excluded if, prior to the surveillance colonoscopy, they had a diagnosis of CRC, hereditary CRC syndrome, and/or inflammatory bowel disease; had a prior colectomy; or their surveillance colonoscopy had an inadequate bowel preparation or was not complete to the cecum.The study sample included all patients who met the eligibility criteria.

Study Outcomes
The outcomes were CRC and advanced neoplasia (either CRC or advanced adenoma).In ascertaining outcome, the most advanced finding from the surveillance procedure was recorded (eg, for a patient diagnosed with both CRC and advanced adenoma, CRC was the recorded finding).Advanced adenoma diagnoses used pathology findings reported at or within 7 days after the procedure.To allow for additional diagnostic procedures for potentially inconclusive examinations, CRC diagnoses were ascertained at or within 180 days after the surveillance colonoscopy.

Data Sources and Definitions
Data from clinical and administrative databases, including electronic health records, were used to obtain information on cohort member demographic characteristics, diagnoses, pathology findings, and procedures.Race and ethnicity were included in the analysis because some racial and ethnic groups in the US experience inequities in access to and utilization and quality of CRC screening and treatment as well as higher CRC incidence and mortality. 33Race and ethnicity data were recorded as 1 of the following 8 categories as documented in the electronic health record: Hispanic; non-Hispanic Alaska Native or American Indian, Asian, Black, Pacific Islander, White, multiracial (reported multiple races), and unknown (race and ethnicity not reported).
Colonoscopies were identified using Current Procedural Terminology codes; International

Classification of Diseases, Ninth Revision and International Statistical Classification of Diseases and
Related Health Problems, Tenth Revision procedure codes; Healthcare Common Procedure Coding System codes; and site-specific codes.Colonoscopy indication (ie, screening, surveillance, diagnostic, and positive fecal immunochemical test result) was ascertained by a validated colonoscopy indication algorithm based on symptoms and conditions identified using electronic health records. 34,35Colonoscopy quality measures (ie, extent of the examination and bowel preparation quality) were ascertained from colonoscopy reports using commercial natural language processing software (Linguamatics I2E; Linguamatics).This approach has been validated in comparison with manual record review. 36enoma detection and histologic features were ascertained using Systematized Nomenclature of Medicine (SNOMED) coding in electronic pathology databases.Advanced adenoma was defined as a conventional adenoma with high-grade dysplasia or villous or tubulovillous histologic features or as any conventional adenoma 10 mm or greater in size; sessile serrated polyps, traditional serrated polyps, and hyperplastic polyps 10 mm or greater in size were not included in the definition.
Nonadvanced adenoma was defined as any conventional adenoma less than 10 mm in size and

Statistical Analysis
Summary statistics were used to describe the characteristics of patients who received a surveillance colonoscopy.Surveillance colonoscopy yields were calculated overall (for all ages), by age group (ie, 70-74, 75-79, and 80-85 years), and by both age group and prior adenoma finding (ie, advanced adenoma or nonadvanced adenoma).All surveillance procedures were considered in the yield calculations (ie, patients could contribute >1 procedure to the calculations).Differences in yield measures by age group were assessed using the χ 2 test of equal proportions.Trends in yields across age groups were evaluated using the Cochran-Armitage test for trend.In the primary analyses, yield measures were calculated using 12 months or longer as the time interval between the surveillance colonoscopy and the index colonoscopy.In sensitivity analyses, yield measures were calculated using 24 or more, 36 or more, 48 or more, and 60 or more months as the time interval to decrease the potential influence of higher-risk patients who may have been recommended to have a relatively early follow-up colonoscopy (ie, within the first few years).
Multivariable logistic regression was used to identify factors associated with advanced neoplasia detection at surveillance, and the odds ratio (OR) with 95% CI was used as an estimate of risk.The variables in the model were patient age (continuous, in years); sex (male or female); race and ethnicity, collapsed into 5 categories (Asian or Pacific Islander, Black, Hispanic, White, and remaining groups [Alaska Native or American Indian, multiracial, and unknown]); body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) of less than 25, 25 to 29.9, or 30 or greater, ascertained at the measurement date closest to the date of the surveillance colonoscopy; tobacco smoking history (ever vs never or unknown); Charlson Comorbidity Index score (0, 1, or Ն2), ascertained in the calendar year before the surveillance colonoscopy; diabetes diagnosis any time prior to the surveillance colonoscopy (yes or no); family history of CRC (yes or no for any relative with CRC); and adenoma findings at the index colonoscopy (advanced or nonadvanced adenoma).In a post hoc analysis, we also included the time interval between the surveillance and index colonoscopies (continuous, in years).Two-sided P < .05indicated statistical significance, and analyses were conducted from September 1, 2022, to February 22, 2024, using SAS, version 9.3 (SAS Institute Inc).

Cohort Characteristics
Among 9601 patients 70 to 85 years of age who had an adenoma detected and a follow-up colonoscopy performed 12 or more months after the examination at which the adenoma was detected, 9740 surveillance colonoscopy procedures were performed from 2017 to 2019 (

Surveillance Colonoscopy Yields Overall and by Age Group Stratified by Prior Adenoma Finding
At an interval of 12 or more months between colonoscopies, patients with a prior advanced adenoma had significantly higher yields of advanced neoplasia (380 of 2305 [16.5%]   3).The advanced neoplasia yields at the surveillance colonoscopy did not differ with increasing age for those with prior advanced adenoma (70-74 years: 2 of 580 [0.3%]; test for trend, P = .10)(Table 3).In sensitivity analyses, in surveillance colonoscopies both among patients with a prior advanced adenoma and patients with a prior nonadvanced adenoma, yield estimates did not differ substantially when the time interval between the surveillance colonoscopy and the index colonoscopy was restricted to 24 or more, 36 or more, 48 or more, and 60 or more months (Table 3).

Factors Associated With Advanced Neoplasia Detection at Surveillance Colonoscopy
In a multivariable analysis (

Discussion
In a large, integrated health care system, among 9740 surveillance colonoscopies in patients 70 to 85 years of age with a history of colorectal adenoma, detection of CRC or advanced neoplasia did not increase significantly with age.Overall, CRC detection was rare (0.3%), while detection of advanced neoplasia was more common (12.0%).Patients with a history of advanced adenoma vs nonadvanced adenoma were more likely to have CRC detected, though still rarely (0.5% vs 0.2%), and were more likely to have advanced neoplasia detected (16.5% vs 10.6%).These findings provide some of the first large-scale, community-based information on the yield of surveillance colonoscopy among older adults.
Prior studies of surveillance colonoscopy yield in older populations have reported conflicting results.4][25][26] For example, a study using the New Hampshire Colonoscopy Registry showed that CRC yield at surveillance colonoscopy was low but increased from 0.4% to 0.6% and 0.8% among older adults aged 70-74, 75-79, and 80-84 years, respectively. 238][29] The inconsistency in prior studies may stem from the wide range of age groups studied (ie, 50-100 years of age), date of publication (given the improvement in colonoscopy techniques and adenoma detection in the past 10-15 years), different cohort sizes (ie, 80-42 611), and varying colonoscopy indications (ie, diagnostic, screening, and/or surveillance) in the study samples.We included only patients 70 to 85 years of age undergoing a surveillance colonoscopy following a colonoscopy in which adenomatous polyps were removed, and the yield estimates represent a contemporary population with regard to colonoscopy quality.Our surveillance colonoscopy yield estimates among older adults with a history of polyps are similar to or slightly higher than rates reported in the literature among individuals younger than 70 years, as would be expected given the older population in our study. 37Based on a recent systematic review and metaanalysis, the yield of CRC among patients aged 50 to 70 years undergoing surveillance for a history of polyps ranged from 0.5% to 2.3%, with a pooled prevalence or yield of 1.4%. 37In addition, yield of advanced polyps among patients aged 50 to 70 years undergoing surveillance for a history of polyps ranged from 2.9% to 24.4%, with a pooled prevalence or yield of 8.2%. 37urrent US guidelines do not provide a recommendation for the age to stop surveillance but advocate for the decision to be individualized based on benefits, risks, patient health status, and patient preferences. 17,18The current study provides 2 key findings that can inform shared decisionmaking between patients and their clinicians.First, CRC detection at surveillance colonoscopy was rare among older adult patients with prior advanced or nonadvanced adenomas.Thus, for many older adults, particularly those with a prior nonadvanced adenoma, the low rate of CRC detection at b Includes non-Hispanic Alaska Native or American Indian, multiracial (reported as multiple races), and unknown (race and ethnicity not reported) as documented in electronic health records.

Table 1 .
Characteristics of Patients Who Underwent a Surveillance Colonoscopy in 2017-2019, Overall and by Age Group b Data are presented as number (percentage) of procedures unless otherwise indicated.JAMA Network Open | Geriatrics

Table 2 )
. Surveillance colonoscopy yields did not increase significantly with age (Figure and Table 2).Among those aged 70 to 74, 75 to 79, and 80 to 85 years, CRC yields were 0.2% Figure.Colorectal Cancer and Advanced Neoplasia Yields at Surveillance Colonoscopy by Age Group and Prior Adenoma Findings Yields were calculated using 12 or more months as the time interval between the surveillance colonoscopy and the index colonoscopy.AA indicates advanced adenoma; NAA, nonadvanced adenoma.

Table 2 .
Surveillance Colonoscopy Neoplasia Yields Overall and by Age Group was restricted to 24 or more, 36 or more, 48 or more, and 60 or more months with the exception that when restricted to 60 or more months, the increase in CRC yield with age was significant (Table2).

Table 3 .
Colorectal Cancer and Advanced Neoplasia Yield at Surveillance Colonoscopy by Prior Adenoma Finding Overall and by Age Group values for the comparison of yield proportions in patients with a prior NAA vs AA by χ 2 test of equal proportions.values for the χ 2 test for trend across age groups.
a Intervals were calculated as the time between the surveillance and the index colonoscopies.b P c P

Table 4
In a post hoc analysis, the time interval between colonoscopies was not associated with advanced neoplasia detection and did not affect risk estimates for other factors.

Table 4 .
Factors Associated With Advanced Neoplasia Detection at Surveillance Colonoscopy Adjusted for patient age, sex, race and ethnicity, BMI, tobacco smoking status, Charlson Comorbidity Index score, family history of colorectal cancer, and index colonoscopy finding.All covariates were ascertained at the time of the surveillance colonoscopy except weight and height for the BMI calculation (obtained at the measurement date closest to the date of the index colonoscopy) and Charlson Comorbidity Index score (ascertained in the calendar year before the surveillance colonoscopy). a