Abstract
Background & Aims:
Older adults with colorectal polyps undergo frequent surveillance colonoscopy. There is no specific guidance regarding when to stop surveillance. We aimed to characterize endoscopist recommendations regarding surveillance colonoscopy in older adults and identify patient, procedure, and endoscopist characteristics associated with recommendations to stop.
Methods:
This was a retrospective cohort study at a single academic medical center of adults aged ≥75 years who underwent colonoscopy for polyp surveillance or screening during which polyps were found. The primary outcome was a recommendation to stop surveillance. Predictors examined included patient age, sex, family history of colorectal cancer, polyp findings, and endoscopist sex and years in practice. Associations were evaluated using multilevel logistic regression.
Results:
Among 1,426 colonoscopies performed by 17 endoscopists, 34.6% contained a recommendation to stop and 52.3% to continue. Older patients were more likely to receive a recommendation to stop, including those 80-84 years (OR 7.7; 95% CI 4.8-12.3) and ≥85 years (OR 9.0; 95%CI 3.3-24.6), compared to those 75-79 years. Family history of colorectal cancer (OR 0.42; 95% CI 0.24-0.74) and a prior history of low-risk (OR 0.17; 95% CI 0.11 to 0.24) or high-risk (OR 0.02; 95% CI 0.01 to 0.04) polyps were inversely associated with recommendations to stop. The likelihood of a recommendation to stop varied significantly across endoscopists.
Conclusions:
Only 35% of adults ≥75 years of age are recommended to stop surveillance colonoscopy. The presence of polyps was strongly associated with fewer recommendations to stop. The variation in endoscopist recommendations highlights an opportunity to better standardize recommendations following colonoscopy in older adults.
Keywords: surveillance colonoscopy, older adults, polyp history, stopping surveillance
Introduction
Surveillance colonoscopy is the most common indication for colonoscopy in adults ≥75 years,1 with over 5.6 million older adults estimated to undergo surveillance colonoscopy annually by 20242. Increased participation in CRC screening and improved polyp detection, coupled with the rapidly aging US population, has heightened the importance of decision-making about surveillance colonoscopy in older adults. While the incidence of CRC increases with age, the precise benefit of colonoscopy in those with prior polypectomy, and in many cases protection from repeated prior colonoscopies, is unknown3. Moreover, the benefit of continued surveillance in some older adults is undoubtedly limited by advancing age and co-morbidities that decrease life expectancy while the potential harms of colonoscopy increase with age4,5.
There is a paucity of data regarding when to stop surveillance colonoscopy in older adults. The 2020 United States Multi-Society Task Force (USMSTF) on CRC calls for additional research on whether the benefits of surveillance outweigh risks for adults over 75 years or with multiple comorbidities6. Previously, the USMSTF recommended an individualized approach to continuing surveillance, considering of risk versus benefit and patient comorbidities7. In the absence of clear guidance on when, if ever, to stop surveillance, we hypothesized that there would be significant variation in endoscopist practice patterns regarding recommendations provided after colonoscopy in older adults.
The aim of this retrospective study was to characterize endoscopist practice patterns and predictors around stopping surveillance colonoscopy in older adults. We describe how often endoscopists at a single academic center recommended stopping surveillance and identify patient, procedure, and endoscopist level variables associated with recommendations to stop.
Methods:
Study Design and Population
We conducted a retrospective review of colonoscopies performed between 1/1/2021-12/31/2019 at a single tertiary referral academic center. We included adults aged ≥75 years undergoing colonoscopy for an indication of surveillance due to a personal history of neoplastic polyps or screening during which ≥ 1 neoplastic polyps (adenomas or serrated polyps) were detected. Patients with a personal history of inflammatory bowel disease, CRC, a genetic CRC syndrome such as Lynch syndrome or familial adenomatous polyposis syndrome were excluded.
Data Sources and Variables
Data from colonoscopy reports were extracted from the Provation® MD endoscopy database, (Provation; Minneapolis, MN): patient age, sex, family history of CRC, performing endoscopist, year of exam, bowel preparation adequacy (adequate versus inadequate), and description of polyp findings (number, size, and whether they were resected completely). An adequate bowel preparation was defined as a total Boston Bowel Preparation Score of ≥6 or a qualitative description of “adequate,” “good,” or “excellent”.8–9
From the electronic health record (EHR), Epic (Epic Systems Corporation; Verona, WI), we evaluated the pre-colonoscopy history and physical for American Society of Anesthesiologists (ASA) classification, pathology report for polyp histology, and any letters written to patients after review of pathology results from their recent colonoscopy. Findings were categorized as “none/hyperplastic polyp(s)”, “low risk” if 1-2 tubular adenomas ≤10 mm in size or serrated polyps ≤ 10 mm in size without dysplasia, and “high risk” if any advanced adenoma (i.e., adenoma ≥10 mm in size, or with presence of high-grade dysplasia or villous features), serrated polyp ≥10 mm in size or with dysplasia (e.g. traditional serrated adenoma) or multiple (≥3) small tubular adenomas or serrated polyps. We categorized multiple small polyps as high risk because the recommendations for follow up at the time of the study were identical to recommendations for advanced adenomas (i.e. 3 years)6.
Endoscopist sex and procedural experience, as measured by years in practice since completion of gastroenterology fellowship, were obtained from departmental records.
Recommendation and rationale for stopping colonoscopy
Recommendations following colonoscopy were abstracted from the colonoscopy report and follow-up letter (if any) written to the patient. If a discrepancy was noted between the colonoscopy report and the letter, the recommendation from the letter was used.
Because there was no existing, validated tool for abstraction of colonoscopy report recommendations, we conducted calibration and testing exercises to ensure that our data abstractors were reliable in assessing for the presence of a recommendation to stop surveillance (Supplemental Text 1). The kappa score was 1.0, which is considered perfect agreement10.
Outcomes
The primary outcome was the proportion of patients who received a recommendation to stop surveillance colonoscopy. The remaining recommendations were classified as either “continue surveillance colonoscopy” or “other.” Other included deferring the decision of colonoscopy to a later time (e.g. “Consider colonoscopy in 3 years pending health” or “Follow-up with primary care provider”), unclear recommendation that could not be classified, or the absence of a recommendation. Secondary outcomes included patient-, procedure-, and endoscopist-level predictors of a recommendation to stop surveillance and an assessment of endoscopist variability in recommending stopping surveillance.
Statistical Analysis
Descriptive statistics were used to summarize patient, procedure, and endoscopist characteristics. Continuous variables were reported as means and standard deviations (SD). Categorical variables were reported as frequencies with percentages.
To evaluate predictors ofendoscopist recommendations to stop surveillance colonoscopy, we conducted univariable analysis for the outcome of a recommendation to stop surveillance colonoscopy versus a recommendation to continue, excluding colonoscopies with “other” recommendations. A multilevel logistic regression model with endoscopists as random effects to account for clustering of colonoscopies by endoscopist was then fit using variables with p <0.20 in univariable analysis or a priori clinical relevance. Clinically meaningful interaction terms (i.e., patient age and sex, ASA score, colonoscopy findings, and family history; patient sex and family history of CRC, colonoscopy findings; ASA score and colonoscopy findings; family of CRC and colonoscopy findings; patient sex and endoscopist sex) were added to the model and retained if significant at the 0.05 level. We evaluated the magnitude of associations between model covariates and a recommendation to stop surveillance with adjusted odds ratios (OR) and 95% confidence intervals (CI). Significance of covariates was assessed using Wald Chi-Square tests at the 0.05 level.
We determined the mean predicted probabilities of stopping surveillance at each combination of age and pathology category using marginal effects estimation. To examine the significance of endoscopist-level variation, we conducted a likelihood ratio test comparing the multilevel model to a model without an endoscopist-level random effect. The discriminative ability of the model was determined by constructing receiver-operating curves (ROC) and calculating the C-statistic (area under ROC).
We used empirical Bayes prediction to calculate the mean probability of a recommendation to stop surveillance for each endoscopist, adjusting for patient (age, sex, ASA class (1/2 vs. 3), family history of CRC), procedure (bowel preparation, polyp, and pathology findings), and endoscopist (sex, experience) characteristics11. Endoscopists were ranked in order of mean probability of stopping surveillance. To quantify the variation in recommendation patterns between endoscopists, we calculated the intra-class correlation coefficient (ICC) using the latent response approach, with level 1 (patient-level) variance set to Π2/312,13 and median odds ratio (MOR) to determine the relative likelihood of an identical patient receiving a recommendation stop by one randomly selected endoscopist compared to another.
All statistical analyses were performed using SAS v9.4 (SAS Institute Inc., Cary, NC, USA) and Stata v15 (StataCorp, College Station, TX, USA).
Ethics
This study was approved by the Committee for Protection of Human Subjects at Dartmouth College as minimal risk on May 6, 2019.
Results:
Between 2012–2019, there were 1,936 colonoscopies performed in adults aged ≥75 years for surveillance or screening (Figure 1). After applying exclusions, a total of 1,426 colonoscopies were included for analysis, of which 1,191 were for surveillance and 235 for screening in which a neoplastic polyp was found. Exams were performed by 17 endoscopists, with a median experience of 21 years (IQR 11-28 years), of whom 5 (29.4%) were female. Each endoscopist contributed a median of 76 exams (IQR 40-112 exams).
Figure 1.
Flow diagram of colonoscopies performed in adults age ≥75 for surveillance of prior polyps or screening in which polyps were detected included in this study.
a Defined as any polypectomy resected in piecemeal fashion and/or requiring repeat colonoscopy at a sooner interval for evaluation or retreatment of post-polypectomy site.
b Defined as a recommendation to follow-up on the decision of surveillance colonoscopy at a later time or in consultation with another provider (e.g. PCP)
Sample characteristics
Table 1 describes the patient, procedure, and endoscopist characteristics of the cohort. Overall, 59.2% of patients were male with most between 75-79 years of age, with a mean (SD) age of 77.8 (± 2.8) years. Most patients were classified as ASA class 2, and 9.9% had a family history of CRC. Neoplastic polyps were found in 912 (64.1%) colonoscopies, of which 592 (41.6%) were categorized as low risk and 321 (35.2%) were high risk. The majority (57.3%) of the high risk polyps were due to having ≥3 small polyps, followed by size ≥1 cm (38.5%). Those that received a recommendation to stop were slightly older, more frequently female, had a lower prevalence of family history of CRC, and a lower prevalence of polyps on their current exam compared to those with a recommendation to continue (Supplemental Table 1). There were no significant differences between those receiving a recommendation to stop versus those receiving a recommendation to continue in bowel preparation quality or performing endoscopist characteristics.
Table 1.
Patient, procedure, and endoscopist characteristics among the 1,426 colonoscopies performed in older adults for surveillance or for screening in which neoplastic polyps were detected and removed
| Variable | Total, N (%) | Stop surveillance, N (%) | Continue surveillance, N (%) | Other,a N (%) |
|---|---|---|---|---|
| Total colonoscopies | 1,426 | 498 | 731 | 197 |
| Patient age, y | ||||
| 75-79 | 1,095 (76.8) | 323 (64.9) | 628 (85.9) | 144 (73.1) |
| 80-84 | 289 (20.3) | 149 (29.9) | 92 (12.6) | 48 (24.4) |
| ≥85 | 42 (2.9) | 26 (5.2) | 11 (1.5) | 5 (2.5) |
| Patient sex | ||||
| Male | 844 (59.2) | 276 (55.4) | 457 (62.5) | 111 (5.6) |
| Female | 582 (40.8) | 222 (44.6) | 274 (37.5) | 86 (4.4) |
| ASA Score b | ||||
| 1 | 112 (7.9) | 29 (5.8) | 77 (10.5) | 6 (3.0) |
| 2 | 1004 (70.4) | 373 (74.9) | 484 (66.2) | 15 (7.6) |
| 3 | 174 (12.2) | 53 (10.6) | 92 (12.6) | 29 (14.7) |
| Family history of colorectal cancer | ||||
| No | 1285 (90.1) | 457 (91.2) | 653 (89.3) | 175 (88.8) |
| Yes | 141 (9.9) | 41 (8.2) | 78 (10.7) | 22 (11.2) |
| Bowel preparation quality b,c | ||||
| Adequate | 1,323 (92.8) | 466 (93.6) | 676 (92.4) | 181 (91.9) |
| Inadequate | 57 (3.9) | 14 (0.2) | 32 (4.4) | 11 (5.6) |
| Pathology b,d | ||||
| None/hyperplastic polyp(s) | 511 (35.8) | 284 (57.0) | 145 (19.8) | 82 (41.6) |
| Low risk | 592 (41.5) | 177 (35.5) | 336 (46.0) | 79 (40.1) |
| High riske | 321 (22.5) | 37 (7.4) | 248 (33.9) | 36 (18.3) |
| Size ≥1 cm | 115 (8.1) | 11 (2.2) | 97 (13.3) | 7 (3.6) |
| High grade dysplasia | 5 (0.4) | 0 (0) | 5 (0.7) | 0 (0) |
| Villous features | 9 (0.6) | 1 (0.2) | 8 (1.1) | 0 (0) |
| Traditional serrated adenoma | 8 (0.6) | 2 (0.4) | 3 (0.4) | 3 (1.5) |
| ≥3 small adenomas and/or SSPs | 184 (1.3) | 23 (4.6) | 135 (18.5) | 26 (13.2) |
| Endoscopist sex | ||||
| Male (n=12) | 1105 (77.5) | 365 (73.3) | 602 (82.4) | 138 (70.1) |
| Female (n=5) | 321 (22.5) | 133 (26.7) | 129 (17.6) | 59 (29.9) |
| Endoscopist years in practice | ||||
| <10 (n=6) | 167 (11.7) | 45 (9.0) | 74 (10.1) | 48 (24.4) |
| 10-19 (n=4) | 227 (15.9) | 61 (12.2) | 108 (14.8) | 58 (29.4) |
| 20-29 (n=4) | 732 (51.3) | 309 (62.0) | 369 (50.5) | 54 (27.4) |
| ≥30 (n=3) | 300 (21.0) | 83 (16.7) | 180 (24.6) | 37 (18.8) |
ASA, American Society of Anesthesiologists; SSP, sessile serrated polyp
Other included deferring the decision of colonoscopy to a later time (e.g. “Consider colonoscopy in 3 years pending health” or Follow-up with the primary care provider”), unclear recommendation that could not be classified, or the absence of a recommendation.
Some values do not sum to total number of colonoscopies due to missing data
Adequate bowel preparation is defined as a Boston Bowel Preparation score of 6 or greater, or described qualitatively in endoscopy report as “adequate,” “good,” or “excellent.”
Pathology findings were categorized as “None/HP” if no polyps or only hyperplastic polyps were found, “low-risk” if 1-2 adenomas ≤ 10 mm in size were found, and “high risk” if meeting the following criteria: adenomas ≥10 mm in size, presence of high-grade dysplasia or villous features, or significant serrated polyp defined as size ≥10 mm or with dysplasia (e.g. traditional serrated adenoma).
Major reason for classifying as high risk listed below, with priority given (in order) to: Size ≥1 cm, high grade dysplasia or villous features, or traditional serrated adenoma among those with multiple features
Recommendations to stop future colonoscopy
A total of 498 (34.9%) records contained a recommendation to stop further surveillance colonoscopy and 731 (51.3%) to continue surveillance. Of the 197 (13.8%) remaining records, 113 were to defer a decision, 76 did not contain a recommendation, and 8 were unclear. The vast majority (N=108, 96.4%) of the 113 recommendations made to defer decision provided some type of rationale (e.g. age, risks versus benefits, prior polyp history).
Predictors of a recommendation to stop surveillance colonoscopy
Table 2 shows the crude and adjusted models predicting the likelihood of a recommendation to stop surveillance. In the adjusted model, when compared to the youngest group (age 75-79 years), individuals aged 80-84 years (OR=7.74; 95% CI, 4.86 to 12.33) and ≥85 years (OR=9.04; 95% CI, 3.33 to 24.58) had a higher odds of receiving a recommendation to stop surveillance. An ASA score of 3 (OR=2.04; 95% CI 1.17 to 3.57) was similarly associated with an increased likelihood of being recommended to stop surveillance compared to those with an ASA score of 1 or 2. In contrast, having a family history of CRC was associated with a reduced odds of being recommended to stop surveillance (OR=0.42; 95% CI 0.24 to 0.74).
Table 2.
Association of patient, procedure, and endoscopist factors with recommendations to stop surveillance colonoscopya,b,c,d
| Crude Odds Ratio (95% CI) |
p-value | Adjusted Odds Ratio (95% CI) |
p-valuee | |
|---|---|---|---|---|
|
|
|
|||
| Patient factors | ||||
|
| ||||
| Age, years | ||||
| 75-79 | Ref | Ref | ||
| 80-84 | 4.43 (3.15, 6.24) | <0.001 | 7.74 (4.86, 12.33) | <0.001 |
| ≥85 | 7.75 (3.32, 18.07) | 9.04 (3.33, 24.58) | ||
| Patient sex | ||||
| Male | Ref | Ref | ||
| Female | 1.36 (1.06, 1.76) | 0.02 | 1.30 (0.93, 1.83) | 0.12 |
| ASA score | ||||
| 1-2 | Ref | 0.08 | Ref | 0.01 |
| 3 | 1.47 (0.95, 2.28) | 2.04 (1.17, 3.57) | ||
| Family history of colorectal cancer | 0.56 (0.36, 0.88) | 0.01 | 0.42 (0.24, 0.74) | 0.003 |
|
| ||||
| Procedure factors | ||||
|
| ||||
| Adequate bowel preparation | 1.06 (0.51, 2.19) | 0.89 | 1.11 (0.44, 2.79) | 0.82 |
| Pathology | ||||
| None/HPs | Ref | <0.001 | Ref | <0.001 |
| Low risk | 0.21 (0.16, 0.29) | 0.17 (0.11,0.24) | ||
| High risk | 0.05 (0.03, 0.08) | 0.02 (0.01, 0.04) | ||
|
| ||||
| Endoscopist factors | ||||
|
| ||||
| Endoscopist sex | ||||
| Male | Reference | Reference | ||
| Female | 2.05 (0.55, 7.67) | 0.29 | 3.85 (0.62, 24.02) | 0.15 |
| Years in practice | ||||
| <10 | Reference | 0.56 | Reference | 0.54 |
| 10-19 | 2.70 (0.38, 19.08) | 3.45 (0.26, 45.06) | ||
| 20-29 | 3.50 (0.57, 21.70) | 4.36 (0.38, 50.01) | ||
| ≥30 | 1.80 (0.27, 12.13) | 1.32 (0.10, 16.68) | ||
ASA, American Society of Anesthesiologists; HP, hyperplastic polyp
Model was additionally adjusted for calendar year of colonoscopy
Multi-level mixed effects logistic regression, assuming endoscopist-level random effects
Interactions terms assessed (none found to be significant at α=0.05): age/sex, age/ASA score, age/pathology, sex/ASA score, age/family history, sex/family history, sex/pathology, ASA score/pathology, family history/pathology, sex/endoscopist sex
Likelihood ratio test comparing models with and without random intercept: p<0.001
Calculated using two-tailed Wald Chi-Square Test (α=0.05)
Older adults with polyps, including either low risk polyps (OR=0.17; 95% CI 0.11 to 0.24) or high risk polyps (OR=0.02; 95% CI 0.01 to 0.04), were significantly less likely to receive a recommendation to stop surveillance. Neither endoscopist sex nor experience were significantly associated with a recommendation to stop surveillance.
The predicted mean probabilities and 95% confidence intervals of endoscopists recommending stopping surveillance for each combination of age and polyp finding are shown in Figure 2. For example, the mean predicted probability of receiving a recommendation to stop surveillance in those 75-79 years without polyps or only hyperplastic polyps was 49.1% (95% CI 34.7% to 63.5%) compared to 6.3% (95% CI 0.6% to 12.1%) in those 75-79 years with high risk polyps. Overall, the likelihood of receiving a recommendation to stop surveillance colonoscopy increased with patient age; however, the presence of low or high risk polyps markedly decreased the likelihood of a recommendation to stop within each age category, even at the oldest ages (≥85). Recommendations were similar between individuals aged 80-84 years and ≥85 years across all pathology categories.
Figure 2.
Adjusted mean probability (95% CI) of endoscopists recommending to stop surveillance according to patient age and colonoscopy findings
Variation in endoscopists’ recommendations for stopping
The adjusted probability of individual endoscopists recommending stopping surveillance ranged from 0 to 61.8% (Figure 3). Across the 17 endoscopists, the average rate of stopping per endoscopist was 26.1% (95% CI 16.0 % to 36.3%). The predicted probabilities of stopping surveillance varied markedly across endoscopists, with the greatest variability seen in the 75-79 age group with none/hyperplastic polyp(s) and in the 80-84 or ≥85 age group with low risk polyps (Supplemental Figure 1). After accounting for patient age and colonoscopy findings, almost half of the variation in recommendation to stop surveillance was explained by unmeasured differences between endoscopists (ICC= 42.1%, 95% CI 25.7% to 70.5%; MOR=4.37, 95% CI 2.77 to 14.50).
Figure 3.
Probability (95% CI) of each endoscopist (N=17) recommending to stop surveillance colonoscopy after accounting for patient, procedure, and measurable endoscopist factors (i.e., sex, experience)
Discussion
In this retrospective study of older adults with a history of colon polyps, we found that only a minority were recommended to stop future surveillance colonoscopy. Increasing patient age was associated with a greater frequency of receiving recommendations to stop surveillance; however, this was strongly attenuated by a finding of neoplastic polyps on the current colonoscopy. Even small, low risk polyps strongly influenced recommendations to continue surveillance, when patients would be 80 or even 90 years old at the time of next colonoscopy. We found that individual endoscopist practice patterns explained about half of the variation in recommendations regarding surveillance colonoscopy, even when accounting for variation in patient, procedural, and known endoscopist characteristics.
The results of this study are consistent with our prior study that used data from a large national endoscopy database to evaluate utilization of surveillance colonoscopy in older adults.1 In that study, which included 3,976 endoscopists at 629 different sites, few older adults were told to stop surveillance colonoscopy. For example, in patients aged ≥90 years with only low risk findings, 37% were recommended to stop colonoscopy. As in the current study, having any type of polyp was associated with a decreased likelihood of being recommended to stop. Among 75-79 years olds with low risk polyps, 12.9% were recommended to stop colonoscopy in that database study compared to 22.9% in this study.
The current study has a few important advantages relative to the prior large database analysis, where information on recommendations may be less reliable. We had full access to individual patient-level data including letters written by the endoscopist post-procedure and had a standardized approach to determining our outcome of interest. Taken together with the prior work, our study suggests that surveillance colonoscopy is likely recommended too frequently among older adults. Additionally, endoscopists place a very high value on polyp findings even when they are low risk (1-2 small adenomas). There is growing evidence that patients with a history of only 1-2 small polyps are a very low risk population with a subsequent risk for advanced neoplasia or CRC that is similar to those with no prior neoplasia14–15. This type of evidence led to lengthening of surveillance intervals up to 10 years in those with 1-2 small adenomas6. In our analysis, patients with low risk findings received recommendations more like those with high risk findings than no findings. The results suggest a need for clear, detailed guidelines for older individuals regarding follow-up.
The 2012 USMSTF CRC surveillance guidelines broadly recommended that surveillance “should not be continued when risk may outweigh benefit” and that the “decision to continue surveillance should be individualized, based on an assessment of benefit, risk, and comorbidities”7 (Table 3). The updated 2020 USMSTF CRC guidelines did not directly address when to stop. In other countries, more specific guidance is provided. Table 3 summarizes recommendations from The British Society of Gastroenterology 16, European Union17, and the European Society of Gastrointestinal Endoscopy.18 Polyp surveillance guidelines that do not offer guidance on management of older adults with polyps should be updated to provide explicit recommendations for stopping surveillance by age or other factors.
Table 3.
Available Post-Polypectomy Surveillance Guidelines for Older Adults
| Society | Year | Surveillance Colonoscopy Recommendation |
|---|---|---|
| U.S. Multi-Society Task Force on Colorectal Cancer6–7 | 2020 | No recommendations provided |
| 2012 | The decision to continue surveillance should be individualized with consideration of benefit, risk, and comorbidities | |
|
| ||
| British Society of Gastroenterology (BSG)/ Association of Coloproctology of Great Britain and Ireland (ACPGBI)/Public Health England (PHE)16 | 2020 | Surveillance should only be performed in people whose life-expectancy is >10 years, and in general, not in people older than about 75 years |
|
| ||
| European Society of Gastrointestinal Endoscopy (ESGE)18 | 2020 | Stop post-polypectomy endoscopic surveillance at the age of 80 years, or earlier if life expectancy is thought to be limited by co-morbidities. |
|
| ||
| European Union24 | 2012 | The cut-off age for stopping surveillance is usually 75 years, but this should also depend upon patient wishes and comorbidity |
Operationalizing the language of potential guideline recommendations is challenging. A reasonable starting point might be to address the large proportion of older adults who have only ever had low risk findings, in whom the risk of future CRC is very low and approximates the risk among those without polyps.14–15 Conversely, guidelines could make recommendations for ongoing surveillance in populations more likely to benefit (e.g. older adults with a recent advanced adenoma). Guidelines must also convey that recommendations on stopping apply to surveillance exams in individuals without symptoms and that new symptoms (e.g., bleeding, change in bowel habits, etc.) may be warrant diagnostic colonoscopy.
The lack of specific recommendations to stop surveillance in the US likely contributes to the considerable variability in practice patterns we observed, even when looking at a single center. This variation, which persisted even when adjusting for age and pathology category, suggests the need for more primary data to guide best practices.19 However, developing such data will be difficult since it would require longitudinal studies of CRC incidence and mortality in this older, polyp-bearing population.
While advancing age and polyp findings were most strongly associated with our outcome of interest, we did identify a few other potentially important factors. Family history of CRC was associated with recommendations to continue surveillance. A greater co-morbidity burden was associated with recommendations to stop surveillance, suggesting that endoscopists are considering overall health and life expectancy in the formulation of their recommendations.20 Bowel preparation was not strongly associated with recommendations to stop surveillance colonoscopy. The low (<4%) prevalence of inadequate bowel preparation across the entire sample may, however, have limited our power to assess this factor. There was a trend towards female patients more frequently receiving a recommendation to stop surveillance, which did not remain significant in multivariable modelling (Supplemental Text 1).
The strengths of this study include the uniform abstraction process of colonoscopy reports to describe practice patterns and predictors of stopping. We were rigorous in ensuring reliability of interpretation of post-colonoscopy recommendations and had access to all post procedural documentation. Within the framework of a single center with a set number of endoscopists, we captured a relatively high number of cases that allowed for different permutations of patient and procedural characteristics. Finally, our findings on predictors of stopping generally have face validity and confirm prior similar work suggesting they are accurate.
Still, we acknowledge certain limitations. This was a single center study that may limit generalizability. While we captured a large number of colonoscopies, there were few patients in the oldest age strata (e.g. ≥85) and few patients in ASA class ≥3. While this limits our ability to comment on these sub-groups, it also reflects current practice patterns in those generally undergoing routine colonoscopy. While our study does not capture individual life expectancy for patients receiving a recommendation to continue surveillance colonoscopy, a prior study estimating life expectancy in a similar colonoscopy population found up to 35% of older adults undergoing surveillance colonoscopy had a life expectancy of <10 years and 4% had a life expectancy of <5 years.21 When we applied U.S. Life Table estimates of life expectancy to our study population based on age, we found that about 23% of older adults recommended to continue surveillance in our study had a life expectancy of <10 years and 0.1% had a life expectancy of <5 years.22 Our study is intended to be descriptive in nature and does not capture the nuances and rationales of why endoscopists choose to stop or continue surveillance. Also, we cannot account for the influence of prior colonoscopies and their findings on the recommendations evaluated in this study. While our study was performed in a single center, a significant proportion of patients in this age range are new to our system (e.g. moved for retirement) or travel in and out of our catchment area (e.g. live in our area part-time and receive only some of their medical care within our system), limiting reliable access to complete prior records. In general, however, high risk adenomas are relatively rare (~<10% of all polyps)23 and unlikely to account for our high percentage of recommendations to continue surveillance in older adults with current low risk findings. Qualitative work to understand factors influencing decision making around colonoscopy in older adults would be informative.
In summary, only 35% of adults 75 and older with a history of polyps are recommended to stop colonoscopy. Recommendations to stop were greatly reduced by the finding of any neoplastic polyp, even when the polyps were low risk. There was wide variation in practice among endoscopists, highlighting the need for more guidance in this area to help standardize practice of colonoscopy among older adults.
Supplementary Material
What you need to know.
Background
While older adults with colorectal polyps undergo frequent surveillance colonoscopy, there is no specific guidance regarding when to stop. The aim of this study was to characterize endoscopist recommendations regarding stopping surveillance in older adults.
Findings
The majority (52.3%) of older adults with polyps are recommended to continue surveillance, while only 34.6% are recommended to stop. This varied across endoscopists. Family history of colorectal cancer and the presence of high-risk polyps were associated with recommendations to continue.
Implications for patient care
Specific guidelines that address when to stop surveillance in older adults with a history of polyps are needed to help standardize clinical practice in this population.
Funding:
Dr. Calderwood is supported by The Dartmouth-Hitchcock Cancer Research Fellows Program, and by the NCI Cancer Center Support Grant 5P30CA023108 to the Dartmouth-Hitchcock Norris Cotton Cancer Center as well as The Dartmouth Clinical and Translational Science Institute, under award number UL1TR001086 from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH)
Disclosures:
Dr. Calderwood has funding from NCI 1R21CA227776-01A1. The contents of this manuscript do not represent the views of the U.S. Department of Veterans Affairs or the United States Government.
Abbreviations:
- ASA
American Society of Anesthesiologists
- BSG
British Society of Gastroenterology
- CI
confidence interval
- CRC
colorectal cancer
- EHR
electronic health record
- HP
hyperplastic polyp
- ICC
intra-class correlation coefficient
- IQR
interquartile range
- MOR
median odds ratio
- OR
odds ratio
- ROC
receiver-operating curves
- SD
standard deviation
- SSP
sessile serrated polyp
- USMSFT
United States Multi-Society Task Force
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Calderwood AH, Calderwood AH, Holub JL, Greenwald DA, Robertson DJ. Yield and practice patterns of surveillance colonoscopy among older adults: An analysis of the GI quality improvement consortium. Am J Gastroenterol 2019;114(11): 1811–9. [DOI] [PubMed] [Google Scholar]
- 2.Lieberman DA, Williams JL, Holub JL, et al. Colonoscopy utilization and outcomes 2000 to 2011. Gastrointest Endosc 2014; [DOI] [PubMed] [Google Scholar]
- 3.Brenner H, Kloor M, Pox CP. Colorectal cancer. Lancet 2014;383(9927):1490–502. [DOI] [PubMed] [Google Scholar]
- 4.Grossberg LB, Papamichael K, Leffler DA, Sawhney MS, Feuerstein JD. Patients over Age 75 Are at Increased Risk of Emergency Department Visit and Hospitalization Following Colonoscopy. Dig Dis Sci 2020;65(7): 1964–70. [DOI] [PubMed] [Google Scholar]
- 5.Day LW, Kwon A, Inadomi JM, Walter LC, Somsouk M. Adverse events in older patients undergoing colonoscopy: A systematic review and meta-analysis. Gastrointest Endosc 2011; 74(4): 885–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gupta S, Lieberman D, Anderson JC, et al. Recommendations for Follow-Up After Colonoscopy and Polypectomy: A Consensus Update by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2020;158(4):1131–1153.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lieberman DA, Rex DK, Winawer SJ, Giardiello FM, Johnson DA, Levin TR. Guidelines for colonoscopy surveillance after screening and polypectomy: A consensus update by the us multi-society task force on colorectal cancer. Gastroenterology 2012;143(3):844–57. [DOI] [PubMed] [Google Scholar]
- 8.Calderwood AH, Jacobson BC. Comprehensive validation of the Boston Bowel Preparation Scale. Gastrointest Endosc 2010;72(4):686–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Calderwood AH, Schroy PC, Lieberman DA, Logan JR, Zurfluh M, Jacobson BC. Boston Bowel Preparation Scale scores provide a standardized definition of adequate for describing bowel cleanliness. Gastrointest Endosc 2014;80(2):269–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Landis JR, Koch GG. The Measurement of Observer Agreement for Categorical Data. Biometrics 1977;33(1):159. [PubMed] [Google Scholar]
- 11.Schroeck FR, Kaufman SR, Jacobs BL, et al. Regional variation in quality of prostate cancer care. J Urol 2014;191(4):957–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Merlo J, Chaix B, Ohlsson H, et al. A brief conceptual tutorial of multilevel analysis in social epidemiology: Using measures of clustering in multilevel logistic regression to investigate contextual phenomena. J Epidemiol Community Health 2006;60(4):290–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sanagou M, Wolfe R, Forbes A, Reid CM. Hospital-level associations with 30-day patient mortality after cardiac surgery: A tutorial on the application and interpretation of marginal and multilevel logistic regression. BMC Med Res Methodol 2012; 12(28). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Click B, Pinsky PF, Hickey T, Doroudi M, Schoen RE. Association of colonoscopy adenoma findings with long-term colorectal cancer incidence. JAMA - J Am Med Assoc 2018;319(19):2021–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dubé C, Yakubu M, McCurdy BR, et al. Risk of Advanced Adenoma, Colorectal Cancer, and Colorectal Cancer Mortality in People With Low-Risk Adenomas at Baseline Colonoscopy: A Systematic Review and Meta-Analysis. Am J Gastroenterol [Internet] 2017. [cited 2020 Oct 29]; 112(12): 1790–801. Available from: https://pubmed.ncbi.nlm.nih.gov/29087393/ [DOI] [PubMed] [Google Scholar]
- 16.Rutter MD, East J, Rees CJ, et al. British Society of Gastroenterology/Association of Coloproctology of Great Britain and Ireland/Public Health England post-polypectomy and post-colorectal cancer resection surveillance guidelines. Gut 2020;69(2):201–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Steele RJC, Rey JF, Lambert R. European guidelines for quality assurance in colorectal cancer screening and diagnosis. First Edition Professional requirements and training. Endoscopy 2012;44(SUPPL3):SE106–15. [DOI] [PubMed] [Google Scholar]
- 18.Hassan C, Antonelli G, Dumonceau J-M, et al. Post-polypectomy colonoscopy surveillance: European Society of Gastrointestinal Endoscopy (ESGE) Guideline – Update 2020. Endoscopy [Internet] 2020. [cited 2021 Mar 26];52(08):687–700. Available from: http://www.thieme-connect.de/DOI/DOI?10.1055/a-1185-3109 [DOI] [PubMed] [Google Scholar]
- 19.Wennberg JE. Unwarranted variations in healthcare delivery: Implications for academic medical centres. Br Med J 2002;325(7370):961–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Vargo JJ, Holub JL, Faigel DO, Lieberman DA, Eisen GM. Risk factors for cardiopulmonary events during propofol-mediated upper endoscopy and colonoscopy. Aliment Pharmacol Ther 2006;24(6):955–63. [DOI] [PubMed] [Google Scholar]
- 21.Cazares K, O’Connor S CA. Life Expectancy Among Adults Undergoing Elective Colonoscopy: A Survey of Older Adults. Am J Gastroenterol 2018;113(S85). [Google Scholar]
- 22.Arias E, Xu J. National Vital Statistics Report [Internet], 2017. [cited 2021 Mar 26]. Available from: https://www.cdc.gov/nchs/products/index.htm.
- 23.Martínez ME, Baron JA, Lieberman DA, et al. A Pooled Analysis of Advanced Colorectal Neoplasia Diagnoses After Colonoscopic Polypectomy. Gastroenterology 2009; 136(3): 832–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Atkin WS, Valori R, Kuipers EJ, et al. European guidelines for quality assurance in colorectal cancer screening and diagnosis First Edition Colonoscopic surveillance following adenoma removal. Endoscopy 2012;44(SUPPL3):SE151–63. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.