- Email: [email protected]
Adenoma detection rate influences the risk of metachronous advanced colorectal neoplasia in low-risk patients
Accepted Manuscript Adenoma detection rate influences the risk of metachronous advanced colorectal neoplasia in low-risk patients Tae Jun Kim, MD, Eun Ran Kim, MD, Sung Noh Hong, MD, PhD, Young-Ho Kim, MD, PhD, Sun-Young Baek, MS, Soohyun Ahn, PhD, Dong Kyung Chang, MD, PhD PII:
S0016-5107(17)32328-3
DOI:
10.1016/j.gie.2017.09.028
Reference:
YMGE 10760
To appear in:
Gastrointestinal Endoscopy
Received Date: 3 April 2017 Accepted Date: 25 September 2017
Please cite this article as: Kim TJ, Kim ER, Hong SN, Kim Y-H, Baek S-Y, Ahn S, Chang DK, Adenoma detection rate influences the risk of metachronous advanced colorectal neoplasia in low-risk patients, Gastrointestinal Endoscopy (2017), doi: 10.1016/j.gie.2017.09.028. 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.
ACCEPTED MANUSCRIPT Title: Adenoma detection rate influences the risk of metachronous advanced colorectal neoplasia in low-risk patients
RI PT
Running title: ADR and metachronous advanced neoplasia
Tae Jun Kim, MD1, Eun Ran Kim, MD1, Sung Noh Hong, MD, PhD1, Young-Ho Kim, MD,
1
SC
PhD1, Sun-Young Baek, MS2, Soohyun Ahn, PhD2, and Dong Kyung Chang, MD, PhD1
Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of
2
M AN U
Medicine, Seoul, Republic of Korea
Statistics and Data Center, Samsung Medical Center, Sungkyunkwan University School of
Medicine, Seoul, Republic of Korea
TE D
Corresponding author Dong Kyung Chang, MD, PhD
Division of Gastroenterology, Department of Medicine, Samsung Medical Center,
Republic of Korea
EP
Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul, 06351,
AC C
Tel: +822-3410-3409, Fax: +822-3410-6983, E-mail: [email protected]
Author contributions:
Study concept and design: Tae Jun Kim and Dong Kyung Chang. Acquisition, analysis, or interpretation of data: Tae Jun Kim. Writing and Drafting of the manuscript: Tae Jun Kim. Critical revision of the manuscript for important intellectual content: Eun Ran Kim, Sung Noh Hong, Young-Ho Kim, and Dong Kyung Chang.
ACCEPTED MANUSCRIPT Statistical analysis: Sun-Young Baek and Soo Hyun Ahn. Study supervision: Dong Kyung Chang.
RI PT
All authors approved the final submission.
Abbreviations used in this paper: BMI, body mass index; CRC, colorectal cancer; AN,
SC
advanced neoplasia; OR, odds ratio; CI, confidence intervals.
AC C
EP
TE D
M AN U
DISCLOSURE: The authors have no conflicts of interest to declare.
ACCEPTED MANUSCRIPT ABSTRACT Background and Aims: In individuals with either no or 1 to 2 nonadvanced adenomas, future risks of advanced colorectal neoplasia (AN) vary according to the clinical risk factors.
RI PT
However, little is known about the association between the adenoma detection rate (ADR) and the risk for metachronous AN in patients with low-risk adenomas.
Methods: We identified 7,171 participants with no or 1-2 non-advanced adenomas at first-
SC
time screening colonoscopy. The risk of metachronous AN was investigated at surveillance
M AN U
colonoscopy, according to clinical characteristics and the ADR.
Results: In multivariate analysis, the risk for metachronous AN was strongly associated with increasing age, male sex, increasing number of adenoma and the ADR of the endoscopist. With the ADR modeled as a continuous variable, each 1.0% increase in the rate of ADR predicted a 3.0% decrease in the risk of metachronous AN (adjusted odds ratio [aOR], 0.97;
TE D
95% confidence CI [CI], 0.95-0.99). With the ADR modeled using a binary cut-off (32%), the risk of metachronous AN was reduced in patients of endoscopists with an ADR ≥32% (aOR, 0.53; 95% CI, 0.35-0.83). Moreover, the risk of metachronous AN was reduced (aOR,
EP
0.66; 95% CI, 0.46-0.95) in patients of endoscopist with an ADR in the highest tertile,
AC C
compared to patients of endoscopists with ADR in the lowest tertile. The impact of ADR on metachronous AN was significant for patients with low-risk adenomas rather than patients with no adenoma.
Conclusions: In patients with low-risk adenomas, the ADR of the endoscopist was inversely associated with the risk of metachronous AN. Key words: Colonoscopy; Surveillance; Adenoma detection rate; Metachronous advanced neoplasia.
ACCEPTED MANUSCRIPT INTRODUCTION Because the majority of colorectal cancers (CRCs) develop from benign adenomas, colonoscopic screening and surveillance have been very effective in preventing CRC.1-14
RI PT
Despite the benefits of screening and surveillance, uncertainty remains with respect to surveillance policies for patients with low-risk adenomas, defined as one or two small (<1 cm) adenomas. Current guidelines recommend surveillance colonoscopy after 10 years or surveillance in 5 to 10 years for individuals with a history of no or 1 to 2 non-advanced
SC
adenomas.15-18 Korean guidelines recommend surveillance after 5 years in low-risk patients.17
M AN U
Recent studies have focused on the time interval between screening and surveillance colonoscopy, with the interval for low-risk patients having been lengthened.7,9,19,20 There is increasing evidence that most low-risk patients are also at low risk for subsequent CRC incidence and mortality, as well as being at low risk for advanced colorectal neoplasia (AN) at follow-up.9,12,20,21 In practice, however, endoscopists do not routinely follow guidelines and
TE D
many recommend shorter colonoscopy surveillance intervals for low-risk individuals.22 The principal reason to recommend a shorter interval might be the concern of interval CRC.
EP
Interval CRC is defined as CRC diagnosed after a previous colonoscopy and before the time of the next recommended examination.23 The majority of interval CRCs are thought to
AC C
result primarily from missed lesions.24-26 Therefore, high-quality colonoscopy is important for preventing interval colorectal neoplasia. Current guidelines for screening and surveillance colonoscopy also mention that high-quality colonoscopy should be the standard.16 Several studies have reported that future risk for AN can vary according to patient- and adenomarelated factors among low-risk patients.21,22,27 However, these studies did not consider colonoscopy quality-related factors, such as the ADR, bowel preparation and withdrawal time.21,27 To the best of our knowledge, there has been no prior large-scale study to evaluate the
ACCEPTED MANUSCRIPT risk for occurrence of metachronous AN among low-risk patients using patient- and adenoma-related characteristics and the colonoscopy quality indicators. The aims of this study were to determine the risk for metachronous AN among individuals with no or low-risk
RI PT
adenomas, and to evaluate whether the ADR determined by endoscopists influences the risk of metachronous AN.
SC
METHODS Study population
M AN U
This retrospective cohort study included healthy adults, aged >45 years old, who participated in a comprehensive health-screening exam at the Center for Health Promotion of the Samsung Medical Center, South Korea, from January 2005 to December 2011 (Figure 1). Because our objective was to evaluate the risk of metachronous colorectal neoplasia
TE D
according to baseline patient characteristics, adenoma characteristics and index colonoscopy quality, we included participants who underwent two colonoscopies, namely a first-time colonoscopy for screening and a second colonoscopy for surveillance (n = 9,547). We
EP
excluded 2,376 participants who met any of the following criteria: poor bowel preparation (n = 509), incomplete colonoscopy (n = 45), history of malignancy (n = 150), history of
AC C
colorectal surgery (n = 9), inflammatory bowel disease (n = 5), and other findings, except low-risk findings, such as no or 1-2 small (<1 cm) low-risk adenomas at index colonoscopy (n = 1,719). Finally, 7,171 participants with no or low-risk colorectal adenomas were included in this study.
This study was approved by the Institutional Review Board of the Samsung Medical Center, and the requirement for informed consent was waived because only de-identified data routinely collected during health screening visits were used.
ACCEPTED MANUSCRIPT Data collection The comprehensive health-screening program included demographic characteristics, anthropometric measurements, detailed physical examination, and a self-administered health
RI PT
questionnaire on smoking, alcohol intake, current medication use, family history of CRC and personal medical history including malignancy, previous colorectal polyps, and surgical history.28 Smoking status was categorized as light smoker (<20 pack years) and heavy
SC
smoker (≥20 pack years). Alcohol consumption was divided into mild (≤10 g/day) and modest (>10 g/day). Height and weight were measured to the nearest 0.1 cm and 0.1 kg,
M AN U
respectively. Each participant’s weight and height was measured with light clothing and in bare feet. Body mass index (BMI) was calculated as the weight (kg) / height (m2).
Colonoscopies
TE D
Thirty-one experienced, board-certified, gastroenterologists performed the colonoscopies after bowel preparation with 4 L of polyethylene glycol. Median year of graduation from medical school of 31 gastroenterologists is 1999 (range, 1989-2001). In this 7,171 study
EP
sample, median number of exams per gastroenterologist was 312 (range, 111-779). The location of the colorectal neoplasm was assessed by the endoscopists, and the size of each
AC C
lesion was estimated using an open biopsy forceps. The following information was recorded in the electronic medical record after the colonoscopy: number of polyps, location and size of polyps, time and results of the last colonoscopy, family history of CRC, bowel preparation (excellent/good/fair/poor), and completeness of the colonoscopy.28
Variables The primary outcome was incidence of metachronous advanced neoplasia (AN) at surveillance
colonoscopy
according
to
baseline
patient
characteristics,
adenoma
ACCEPTED MANUSCRIPT characteristics, and quality of index colonoscopy. AN was defined as any adenoma ≥1 cm, any adenoma with a villous component, high-grade dysplasia, or invasive cancer.18 Low-risk adenoma was defined as 1 or 2 adenomas, of <1 cm and without a villous component or high-
RI PT
grade dysplasia.18 Patients-related information used in the analysis included age, sex, BMI, smoking status, alcohol intake, regular use of aspirin, family history of CRC, bowel preparation, and follow-up interval. Adenoma-related characteristics included in the analysis
SC
were as follows; adenoma number, size, and location. Withdrawal time and ADR was used as quality indicators of the index colonoscopy. The ADR, determined by each endoscopist, was
M AN U
calculated from data of individuals ≥50 years old who underwent a first-time screening colonoscopy, except cases with poor bowel preparation.
Statistical analysis
TE D
Baseline characteristics of patients were summarized using conventional statistics. Continuous variables were reported as means ± standard deviation, whereas categorical variables were presented as percentages. Continuous variables were compared between
EP
groups using the Student t-test, whereas categorical variables were compared using the chisquared test. The main outcome was metachronous colorectal neoplasia at surveillance
AC C
colonoscopy. Baseline factors with a potential prognostic effect on metachronous colorectal neoplasia were initially analyzed by univariate and multivariate logistic regression analyzes. We conducted logistic regression models to estimate the adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for metachronous colorectal neoplasia associated with baseline patient, adenoma characteristics and the adenoma detection rate. Multivariate logistic regression, with step wise variable selection, was conducted by including all variables that were significant in univariate analysis. The candidate factors included age, sex, BMI, smoking status, alcohol consumption, regular aspirin use, family history of CRC, follow-up
ACCEPTED MANUSCRIPT interval, bowel preparation, number and size of baseline adenomas, withdrawal time, and the ADR of the endoscopist. We further included the ADR of the endoscopists, who performed the follow-up colonoscopy, as a potential confounding factor. We modeled the ADR as a
RI PT
continuous variable, as well as considering binary and tertile cutoffs. The Youden index, sensitivity, specificity, and positive and negative predictive values were calculated. From the highest Youden index, we selected an optimal cutoff point of ADR for predicting
SC
metachronous colorectal neoplasia.
Finally, we built a classification tree to predict the ‘hidden’ risk of metachronous AN
M AN U
among the low-risk patients. We split the root node, or each of the child nodes of the root node, by a single variable selected by recursive partitioning, a process that minimizes the sum of the node impurities in the two child nodes. In this model, age, sex, number of adenomas, and the ADR were considered inputs. The P value at each node measured the association
TE D
between the selected input variable and the response (incidence of AN). The terminal nodes were grouped according to the response (risk of metachronous AN), and the results were presented as the final set of prognostic factors. A P value < 0.05 was considered statistically
EP
significant. Statistical analyzes were performed using R version 3.3.1 (The R Foundation for
AC C
Statistical Computing, Vienna, Austria) and SAS version 9.4 (SAS Institute, Cary, NC).
RESULTS
Baseline characteristics of the study participants The baseline characteristics of the 7,171 study participants with no or low-risk adenomas are shown in Table 1. The mean (SD) age of participants was 52.7 (6.1) years, with 61.4% being male. Participants with low-risk adenomas were more likely to be older, male, have a higher BMI, and to be heavy smoker and modest alcohol consumers, compared to those without adenomas. However, no between-group difference in aspirin use and family history of CRC
ACCEPTED MANUSCRIPT was observed.
The risk of any metachronous colorectal adenomas
RI PT
Among the 7,171 participants with no or low-risk colorectal adenomas, any colorectal adenomas developed in 2,291 participants at 3 to 5 years after screening colonoscopy, with an incidence rate of 31.9%. In univariate logistic regression analysis, increasing age, male sex,
SC
increasing BMI, heavy smoking, modest alcohol intake, family history of CRC, longer duration of follow-up interval, increasing number of adenoma, larger size of adenoma,
M AN U
withdrawal time, the ADR of the endoscopists and the ADR of the endoscopist performing the follow-up endoscopy were significant predictors for any metachronous colorectal adenoma at surveillance colonoscopy. In the multivariate analysis, age (OR, 1.04; 95% CI, 1.03-1.05), male sex (OR, 1.65; 95% CI, 1.38-1.98), heavy smoking (OR, 1.23; 95% CI,
TE D
1.04-1.45), family history of CRC (OR, 1.36; 95% CI, 1.09-1.69), longer duration of followup interval (OR, 1.01; 95% CI, 1.00-1.01), increasing number of adenoma (OR, 1.68; 95% CI, 1.54-1.83), withdrawal time (OR, 0.97; 95% CI, 0.95-0.99), ADR of the endoscopist (OR,
EP
0.97; 95% CI, 0.96-0.99), and ADR of the endoscopist who performed the follow-up colonoscopy (OR, 1.03; 95% CI, 1.02-1.05) were significant predictors of any metachronous
AC C
colorectal adenoma (Table 2).
The risk of metachronous advanced neoplasia Among the 7,171 participants, AN developed in 154 participants, an incidence rate of 2.1%. In the univariate analysis, increasing age, male sex, heavy smoking, modest alcohol intake, increasing number of adenoma, proximal location of adenoma, withdrawal time, ADR of the endoscopist, and ADR of the endoscopist who performed the follow-up colonoscopy were significantly associated with the risk of metachronous AN. In multivariate analysis,
ACCEPTED MANUSCRIPT increasing age (OR, 1.05; 95% CI, 1.02-1.07), male sex (OR, 1.57; 95% CI, 1.13-2.20), increasing number of adenoma (OR, 1.90; 95% CI, 1.56-2.31), ADR of the endoscopist (OR, 0.97; 95% CI, 0.95-0.99), and ADR of the endoscopist who performed the follow-up
RI PT
colonoscopy (OR, 1.03; 95% CI, 1.01-1.06) were retained as independent predictors of metachronous AN (Table 3).
Recursive partitioning analysis of prognostic factors demonstrated that the risk of
SC
metachronous AN at follow-up was stratified based on number of adenoma, age, and the ADR of the endoscopist. These variables were also retained as significant factors for
M AN U
metachronous AN in the multivariate logistic regression model. Among the terminal nodes with different risk of metachronous AN, the highest-risk node had an 8.1% risk of incident AN, with all other nodes have a risk of AN of <5%. The highest-risk node was classified according to number of adenoma and the ADR of the endoscopist, and included patients with
TE D
2 adenomas who were examined by endoscopists with a low ADR (<32%). Aggregating the remaining nodes, the risk of metachronous AN at follow-up was 1.7%. The difference in risk of metachronous AN between the 2 groups was significant (OR, 3.98; 95% CI, 2.76-5.74; P <
EP
0.001).
AC C
Adenoma detection rate and risk of metachronous colorectal neoplasia We calculated the ADR for each of the 31 endoscopists in this study. The ADRs of the endoscopists were distributed from 18.6% to 42.0%, with a median ADR of 30.6% (Supplementary Figure 1). Mean withdrawal time of the normal screening colonoscopy was significantly related to the ADR (coefficient = 0.597, standard error = 0.027, P < 0.001). With the ADR modeled as a continuous variable, each 1.0% increase in the rate predicted a 3.0% decrease in the risk of metachronous AN (adjusted odds ratio [aOR], 0.97; 95% confidence CI [CI], 0.95-0.99) (Table 4). We selected an optimal cutoff point of ADR for predicting
ACCEPTED MANUSCRIPT metachronous colorectal neoplasia, according to the Youden index. The cutoff point of ADR was 29% for any metachronous colorectal neoplasia and 32% for metachronous AN, respectively. We then dichotomized the ADR into a low and high ADR from the optimal
RI PT
cutoff point for predicting metachronous colorectal neoplasia. With the ADR modeled by binary cut-off (32%), the risk of metachronous AN was reduced in patients of endoscopists with an ADR ≥32% (aOR, 0.53; 95% CI, 0.35-0.83). When the endoscopist ADR was
SC
considered in tertiles, the risk for metachronous AN was lower for patients of endoscopists with an ADR in the upper tertile than in patients of endoscopists with an ADR in the lowest
M AN U
tertile (aOR, 0.66; 95% CI, 0.46-0.95). Similarly, the risk was lower for patients of endoscopists with an ADR in the middle compared to lowest tertile (aOR, 0.68; 95% CI, 0.47-0.99). In subgroup analyzes, the ADR was not a significant predictor for metachronous AN among patients with no adenoma; however, the ADR was a significant predictor among
TE D
patients with low-risk adenoma (Table 5). The subgroup analysis, however, did not reveal a significant interaction between baseline adenomas (no adenoma versus low-risk adenoma; P
DISCUSSION
EP
for interaction = 0.998).
AC C
To the best of our knowledge, this is the first large-scale study to investigate prognostic factors for occurrence of metachronous AN in low-risk patients using a number of readily available clinical characteristics and ADR as a colonoscopy quality indicator. Current guidelines for surveillance colonoscopy of low-risk individuals at screening colonoscopy are heterogeneous. For individuals with 1 to 2 low-risk adenomas, recommendations for surveillance range from routine screening after 10 years,15,16 to performance of surveillance at 5 to 10 years,18 to a 5-year surveillance.17 Heterogeneity among the guidelines may be due to uncertainty of risk for metachronous colorectal neoplasia among the large and heterogeneous
ACCEPTED MANUSCRIPT group of low-risk patients. Indeed, we found that the risk for metachronous colorectal neoplasia varied according to patient characteristics, adenoma characteristics, and quality of index colonoscopy. Results from multivariate analysis demonstrated that increasing age, male
RI PT
sex, increasing number of adenoma, and the ADR of the endoscopist were independent predictors of metachronous any adenomas or AN at follow- up colonoscopy. Considering the ADR as a continuous, binary or ternary variable, a low ADR was a significant risk factor for
SC
metachronous AN. From the results of recursive partitioning analysis, patients with 2 small adenomas who were screened by endoscopists with a low ADR were identified as a hidden-
M AN U
risk group. Our results reinforced the importance of the quality of the screening colonoscopy for preventing the incidence of metachronous AN.
Several previous studies also investigated risk stratification for surveillance in low-risk patients, according to clinical characteristics.21,27 The studies revealed that several patient or
TE D
adenoma characteristics were significantly associated with incidence of metachronous AN. However, the studies did not consider quality factors of the index colonoscopy, such as bowel preparation, withdrawal time, and ADR of the endoscopists who performed the screening
EP
colonoscopy. The recommended surveillance guideline applies to individuals who underwent high-quality colonoscopy, such as complete colonoscopy with a meticulous inspection of
AC C
adequately cleaned colorectal mucosa.16 In the multivariate analysis, we included potential confounding factors. Additionally, we considered the ADR of the endoscopists, which is the most important quality indicator of interval colorectal neoplasia. Indeed, a low ADR was an independent predictor for both any metachronous colorectal neoplasia and metachronous AN. Even among low-risk patients, a low ADR was significantly associated with risk of metachronous AN. These results underline how variation in the performance of colonoscopy
ACCEPTED MANUSCRIPT may influence its effectiveness with respect to detection and prevention of colorectal neoplasia. Our results have potential clinical implications. The risk stratification may be due to
RI PT
variations in risk for metachronous colorectal neoplasia according to patient-related and adenoma-related characteristics, and quality of the index colonoscopy, even in the low-risk group. The results of our recursive partitioning analysis demonstrated that presence of 2
SC
adenomas and performance by an endoscopist with a low ADR increased the risk for metachronous AN. We defined the group with a higher risk for metachronous AN as the
M AN U
hidden-risk group among low-risk patients. Our results underlined the importance of the ADR as a prognostic factor for incident AN, to the same extent as the number, size and histology of adenomas at baseline.
The impact of ADR on metachronous advanced colorectal neoplasia was significant for
TE D
patients with low-risk adenomas rather than patients with no adenoma. This is probably due to the very low risk of metachronous advanced neoplasia in individuals without any adenomas. Indeed, in this study, the incidence of advanced colorectal neoplasia was less than
EP
1.7% in individuals with no adenoma at baseline. Therefore, the risk of metachronous
AC C
advanced neoplasia was absolutely low that this group would not have been affected by ADR. The ADR benchmark of 32% in this study was higher than the recommended ADR benchmark of 25% for all patients (30% for men and 20% for women).29 This discrepancy may be due to differences in the prevalence of colorectal adenoma according to geographic area. Recent data from United States reported a prevalence rate of colorectal adenomas of 20% to 23% among individuals 50- to 59-years-old.30 A recent Korean study reported an overall prevalence rate of 37.8% among individuals 50- to 59-years-old.31 In areas with high
ACCEPTED MANUSCRIPT prevalence rates of colorectal adenomas, the ADR benchmark may also increase accordingly. Future studies will be required to investigate these findings. Several limitations should be considered in interpreting our findings. First,
RI PT
misclassification might have occurred during the measurement of polyp size and histological diagnosis. However, as endoscopists and pathologists were unaware of the study objectives, these measurement errors would be non-differential and could possibly result in an
SC
underestimation of the association between clinical characteristics and metachronous colorectal neoplasia. Second, our study was performed at a single center and focused on
M AN U
asymptomatic Korean men and women; therefore, it may be difficult to generalize our findings to other populations. Third, several co-variates, such as smoking habit, alcohol intake and medical history, might be inaccurate based on self-report in questionnaire survey or medical records. Finally, the main outcome was metachronous lesions at 3 to 5 years. In
TE D
the current guidelines, surveillance colonoscopy was recommended at 5 to 10 years among low-risk individuals. Therefore, it is unclear whether there is a substantial difference in the risk of metachronous AN related to differences in surveillance intervals.
EP
This study also has considerable strengths, including its cohort design, large sample size,
AC C
inclusion of first-time colonoscopies, and consideration of colonoscopic quality indicators, such as bowel preparation, completeness of colonoscopy, indications for colonoscopy, and ADR of the endoscopists. Comprehensive risk variables, including patient-related data, adenoma-related data, and the quality indicators of colonoscopy, resulted in well-powered evaluations on the rare incidence of metachronous AN and more detailed analysis through multivariate adjustment. Additional strengths include the use of high quality, standardized clinical methods and a standardized self-administered health questionnaire completed by all participants.
ACCEPTED MANUSCRIPT In conclusion, there is substantial variation in the risk of metachronous AN according to readily available clinical characteristics and the ADR of the endoscopists among low-risk patients. An increasing number of adenomas, increasing patient age, male sex, and
RI PT
performing by an endoscopist with a low ADR were independent predictors for metachronous AN. The ADR was inversely associated with the risk of metachronous colorectal neoplasia, including any adenomas and advanced neoplasia in patients with low-risk adenomas. This
SC
association was consistently observed, even when the ADR was modeled either as a continuous, binary or ternary variable. Our findings support the validity of the ADR as a
M AN U
quality indicator of the performance of a screening colonoscopy even in low-risk patients. For low-risk patients who have undergone screening colonoscopy by an endoscopist with an appropriate ADR, long surveillance interval is likely to be sufficient. REFERENCES
4. 5.
6.
7. 8.
9. 10. 11.
TE D
3.
EP
2.
Brenner H, Chang-Claude J, Seiler CM, et al. Protection from colorectal cancer after colonoscopy: a population-based, case-control study. Ann Intern Med 2011;154:22-30. Kahi CJ, Imperiale TF, Juliar BE, et al. Effect of screening colonoscopy on colorectal cancer incidence and mortality. Clin Gastroenterol Hepatol 2009;7:770-5; quiz 11. Nishihara R, Wu K, Lochhead P, et al. Long-term colorectal-cancer incidence and mortality after lower endoscopy. N Engl J Med 2013;369:1095-105. Shaukat A, Mongin SJ, Geisser MS, et al. Long-term mortality after screening for colorectal cancer. N Engl J Med 2013;369:1106-14. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;116:544-73. Atkin WS, Edwards R, Kralj-Hans I, et al. Once-only flexible sigmoidoscopy screening in prevention of colorectal cancer: a multicentre randomised controlled trial. Lancet 2010;375:1624-33. Brenner H, Chang-Claude J, Seiler CM, et al. Long-term risk of colorectal cancer after negative colonoscopy. J Clin Oncol 2011;29:3761-7. Holme O, Loberg M, Kalager M, et al. Effect of flexible sigmoidoscopy screening on colorectal cancer incidence and mortality: a randomized clinical trial. JAMA 2014;312:606-15. Imperiale TF, Glowinski EA, Lin-Cooper C, et al. Five-year risk of colorectal neoplasia after negative screening colonoscopy. N Engl J Med 2008;359:1218-24. Schoen RE, Pinsky PF, Weissfeld JL, et al. Colorectal-cancer incidence and mortality with screening flexible sigmoidoscopy. N Engl J Med 2012;366:2345-57. Segnan N, Armaroli P, Bonelli L, et al. Once-only sigmoidoscopy in colorectal cancer screening: follow-up findings of the Italian Randomized Controlled Trial--SCORE. J
AC C
1.
ACCEPTED MANUSCRIPT
17. 18.
19.
20.
21.
22.
23. 24. 25.
26.
27. 28. 29.
RI PT
SC
16.
M AN U
15.
TE D
14.
EP
13.
AC C
12.
Natl Cancer Inst 2011;103:1310-22. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale-Update based on new evidence. Gastroenterology 2003;124:544-60. Winawer SJ, Zauber AG, O'Brien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. The National Polyp Study Workgroup. N Engl J Med 1993;328:901-6. Zauber AG, Lansdorp-Vogelaar I, Knudsen AB, et al. Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force. Ann Intern Med 2008;149:659-69. 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 Suppl 3:SE151-63. Hassan C, Quintero E, Dumonceau JM, et al. Post-polypectomy colonoscopy surveillance: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2013;45:842-51. Hong SN, Yang DH, Kim YH, et al. [Korean guidelines for post-polypectomy colonoscopic surveillance]. Korean J Gastroenterol 2012;59:99-117. Lieberman DA, Rex DK, Winawer SJ, et al. Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2012;143:844-57. Noshirwani KC, van Stolk RU, Rybicki LA, et al. Adenoma size and number are predictive of adenoma recurrence: implications for surveillance colonoscopy. Gastrointest Endosc 2000;51:433-7. Ponugoti PL, Rex DK. Yield of a second screening colonoscopy 10 years after an initial negative examination in average-risk individuals. Gastrointest Endosc 2017;85:221-4. Chung SJ, Kim YS, Yang SY, et al. Five-year risk for advanced colorectal neoplasia after initial colonoscopy according to the baseline risk stratification: a prospective study in 2452 asymptomatic Koreans. Gut 2011;60:1537-43. Anderson JC, Baron JA, Ahnen DJ, et al. Factors Associated With Shorter Colonoscopy Surveillance Intervals for Patients With Low-Risk Colorectal Adenomas and Effects on Outcome. Gastroenterology 2017;152:1933-43 e5. Brenner H, Chang-Claude J, Seiler CM, et al. Interval cancers after negative colonoscopy: population-based case-control study. Gut 2012;61:1576-82. le Clercq CM, Bouwens MW, Rondagh EJ, et al. Postcolonoscopy colorectal cancers are preventable: a population-based study. Gut 2014;63:957-63. Bressler B, Paszat LF, Chen Z, et al. Rates of new or missed colorectal cancers after colonoscopy and their risk factors: a population-based analysis. Gastroenterology 2007;132:96-102. Singh H, Nugent Z, Demers AA, et al. Rate and predictors of early/missed colorectal cancers after colonoscopy in Manitoba: a population-based study. Am J Gastroenterol 2010;105:2588-96. Gupta S, Jacobs ET, Baron JA, et al. Risk stratification of individuals with low-risk colorectal adenomas using clinical characteristics: a pooled analysis. Gut 2015. Rampal S, Yang MH, Sung J, et al. Association between markers of glucose metabolism and risk of colorectal adenoma. Gastroenterology 2014;147:78-87 e3. Rex DK, Schoenfeld PS, Cohen J, et al. Quality indicators for colonoscopy. Gastrointest Endosc 2015;81:31-53.
ACCEPTED MANUSCRIPT
31.
Corley DA, Jensen CD, Marks AR, et al. Variation of adenoma prevalence by age, sex, race, and colon location in a large population: implications for screening and quality programs. Clin Gastroenterol Hepatol 2013;11:172-80. Koo JE, Kim KJ, Park HW, et al. Prevalence and risk factors of advanced colorectal neoplasms in asymptomatic Korean people between 40 and 49 years of age. J Gastroenterol Hepatol 2017;32:98-105.
FIGURE LEGENDS Figure 1. Flow diagram of study participants
SC
SUPPLEMENTAL MATERIAL
RI PT
30.
AC C
EP
TE D
M AN U
Supplementary Figure 1. Distribution of adenoma detection rate by endoscopists
ACCEPTED MANUSCRIPT Table 1. Baseline characteristics of individuals with none or low-risk adenomas No adenoma Low-risk adenomas All P value (n= 7,171) (n = 4,100) (n = 3,071) 52.7 ± 6.1
51.9 ± 5.7
53.9 ± 6.5
61.4 38.6 23.9 ± 2.8
57.0 43.0 23.7 ± 2.7
67.2 32.8 24.2 ± 2.9
59.1 40.9
62.7 37.3
<0.001 <0.001
RI PT
Age (years) Sex (%) Male Female BMI (kg/m2) Smoking status (%) <20 pack years ≥20 pack years Modest alcohol intake (%) Mild (≤10 g/day) Modest (>10 g/day) Regular aspirin use (%)
<0.001 <0.001
54.4 45.6
SC
<0.001
AC C
EP
TE D
M AN U
82.4 84.2 17.6 15.8 10.2 10.9 Family history of CRC (%) 6.9 7.1 Values are expressed as means ± standard deviation or percentages. BMI, body mass index; CRC, colorectal cancer.
80.1 19.9 9.6 6.5
0.055 0.391
ACCEPTED MANUSCRIPT Table 2. The risk of metachronous any colorectal adenomas according to baseline patient, adenoma characteristics, and quality indicator of index colonoscopy Univariate analysis Multivariate analysis OR (95% CI)
P value
1.04 (1.03-1.05) 1.97 (1.77-2.19) 1.07 (1.05-1.08)
<0.001 <0.001 <0.001
1.04 (1.03-1.05) 1.65 (1.38-1.98)
<0.001 <0.001
Smoking status 1.00 (reference)
1.00 (reference)
≥ 20 pack years
1.95 (1.70-2.23)
<0.001
Modest alcohol intake
1.46 (1.28-1.66)
<0.001
Regular aspirin use
1.01 (0.91-1.13)
0.798
Family history of CRC
1.38 (1.13-1.69)
0.002
Follow-up interval
1.01 (1.00-1.01)
0.001
Bowel preparation
1.00 (reference)
Good
0.98 (0.89-1.09)
0.731
Excellent
1.01 (0.83-1.24)
0.904
Number of adenoma Location of adenoma Distal location
1.73 (1.61-1.86)
<0.001
0.016
1.36 (1.09-1.69)
0.007
1.01 (1.00-1.01)
0.002
1.68 (1.54-1.83)
<0.001
1.00 (reference)
1.07 (0.92-1.25)
0.393
1.77 (1.37-2.28)
<0.001
EP
Proximal location Both location
TE D
Fair
Adenoma characteristics
1.23 (1.04-1.45)
M AN U
< 20 pack years
RI PT
P value
SC
Patient characteristics Age Male sex BMI
OR (95% CI)
<5
AC C
Size of adenoma (mm) 5 to <10
1.00 (reference) 1.30 (1.10-1.53)
0.002
Withdrawal time
0.97 (0.96-0.99)
0.013
0.97 (0.95-0.99)
0.016
ADR of the endoscopist
0.98 (0.97-0.99)
<0.001
0.97 (0.96-0.99)
<0.001
Quality indicator
Potential confounder ADR of the endoscopist who performed the 1.04 (1.03-1.05) <0.001 1.03 (1.02-1.05) <0.001 follow-up colonoscopy OR, odds ratio; CI, confidence intervals; BMI, body mass index; CRC, colorectal cancer; ADR, adenoma detection rate.
ACCEPTED MANUSCRIPT Table 3. The risk of metachronous advanced neoplasia according to baseline patient, adenoma characteristics, and quality indicator of index colonoscopy Univariate analysis Multivariate analysis OR (95% CI)
P value
1.06 (1.04-1.08) 2.19 (1.52-3.15) 1.04 (0.99-1.10)
<0.001 <0.001 0.126
1.05 (1.02-1.07) 1.57 (1.13-2.20)
<0.001 0.008
Smoking status 1.00 (reference)
≥ 20 pack years
1.61 (1.19-2.18)
0.002
Modest alcohol intake
1.52 (1.07-2.16)
0.018
Regular aspirin use
0.82 (0.59-1.13)
0.214
Family history of CRC
1.20 (0.69-2.10)
0.512
Follow-up interval
1.00 (0.99-1.01)
0.535
M AN U
< 20 pack years
Bowel preparation
1.00 (reference)
Good
1.22 (0.89-1.66)
0.674
Excellent
1.14 (0.62-2.07)
0.218
Number of adenoma Location of adenoma Distal location
2.03 (1.68-2.46)
<0.001
1.90 (1.56-2.31)
<0.001
0.97 (0.95-0.99)
0.026
1.00 (reference)
1.51 (1.00-2.29)
0.05
2.23 (1.26-3.96)
0.006
EP
Proximal location Both location
TE D
Fair
Adenoma characteristics
RI PT
P value
SC
Patient characteristics Age Male sex BMI
OR (95% CI)
<5
AC C
Size of adenoma (mm) 5 to <10
1.00 (reference) 1.36 (0.88-2.11)
0.166
Withdrawal time
0.98 (0.96-0.99)
0.046
ADR of the endoscopist
0.98 (0.96-1.00)
0.05
Quality indicator
Potential confounder ADR of the endoscopist who performed the 1.04 (1.01-1.06) 0.007 1.03 (1.01-1.06) 0.01 follow-up colonoscopy OR, odds ratio; CI, confidence intervals; BMI, body mass index; CRC, colorectal cancer; ADR, adenoma detection rate.
ACCEPTED MANUSCRIPT Table 4. The risk of metachronous colorectal neoplasia according to adenoma detection rate Any adenomas Advanced neoplasia P value
aORa (95% CI)
P value
0.97 (0.96-0.99)
<0.001
0.97 (0.95-0.99)
0.026
1.00 (reference)
High ADR
0.66 (0.56-0.78)
RI PT
By continuous variables ADR By binary cut-off Low ADR
aORa (95% CI)
1.00 (reference) <0.001
0.53 (0.35-0.83)
By tertiles 1.00 (reference)
Middle ADR
0.69 (0.57-0.82)
<0.001
Highest ADR
0.63 (0.52-0.75)
<0.001
a
1.00 (reference)
SC
Lowest ADR
0.005
0.68 (0.47-0.99)
0.044
0.66 (0.46-0.95)
0.025
M AN U
Estimated from logistic regression models. The multivariable analysis was adjusted for age, sex, BMI, smoking status, alcohol intake, regular aspirin use, family history of CRC, follow-up interval, bowel preparation, number of adenoma, size of adenoma, withdrawal time, and ADR of the endoscopist who performed the follow-up colonoscopy.
AC C
EP
TE D
OR, odds ratio; CI, confidence intervals; adenoma detection rate.
ACCEPTED MANUSCRIPT Table 5. The risk of metachronous advanced neoplasia by adenoma detection rate in no adenoma and low-risk adenoma Subgroup of no adenoma Subgroup of low-risk adenoma P value
aORa (95% CI)
P value
0.98 (0.94-1.01)
0.211
0.97 (0.95-1.00)
0.05
1.00 (reference)
High ADR
0.45 (0.18-1.07)
RI PT
By continuous variables ADR By binary cut-off Low ADR
aORa (95% CI)
1.00 (reference) 0.089
0.56 (0.34-0.93)
By tertiles 1.00 (reference)
Highest ADR
0.63 (0.33-1.08)
a
1.00 (reference) 0.102
SC
Lowest ADR
0.025
0.65 (0.41-0.98)
0.045
M AN U
Estimated from logistic regression models. The multivariable analysis was adjusted for age, sex, BMI, smoking status, alcohol intake, regular aspirin use, family history of CRC, follow-up interval, bowel preparation, number of adenoma, size of adenoma, withdrawal time, and ADR of the endoscopist who performed the follow-up colonoscopy.
AC C
EP
TE D
OR, odds ratio; CI, confidence intervals; adenoma detection rate.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Abbreviations: BMI, body mass index; CRC, colorectal cancer; AN, advanced neoplasia; OR,
AC C
EP
TE D
M AN U
SC
RI PT
odds ratio; CI, confidence intervals.
S0016-5107(17)32328-3
DOI:
10.1016/j.gie.2017.09.028
Reference:
YMGE 10760
To appear in:
Gastrointestinal Endoscopy
Received Date: 3 April 2017 Accepted Date: 25 September 2017
Please cite this article as: Kim TJ, Kim ER, Hong SN, Kim Y-H, Baek S-Y, Ahn S, Chang DK, Adenoma detection rate influences the risk of metachronous advanced colorectal neoplasia in low-risk patients, Gastrointestinal Endoscopy (2017), doi: 10.1016/j.gie.2017.09.028. 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.
ACCEPTED MANUSCRIPT Title: Adenoma detection rate influences the risk of metachronous advanced colorectal neoplasia in low-risk patients
RI PT
Running title: ADR and metachronous advanced neoplasia
Tae Jun Kim, MD1, Eun Ran Kim, MD1, Sung Noh Hong, MD, PhD1, Young-Ho Kim, MD,
1
SC
PhD1, Sun-Young Baek, MS2, Soohyun Ahn, PhD2, and Dong Kyung Chang, MD, PhD1
Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of
2
M AN U
Medicine, Seoul, Republic of Korea
Statistics and Data Center, Samsung Medical Center, Sungkyunkwan University School of
Medicine, Seoul, Republic of Korea
TE D
Corresponding author Dong Kyung Chang, MD, PhD
Division of Gastroenterology, Department of Medicine, Samsung Medical Center,
Republic of Korea
EP
Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul, 06351,
AC C
Tel: +822-3410-3409, Fax: +822-3410-6983, E-mail: [email protected]
Author contributions:
Study concept and design: Tae Jun Kim and Dong Kyung Chang. Acquisition, analysis, or interpretation of data: Tae Jun Kim. Writing and Drafting of the manuscript: Tae Jun Kim. Critical revision of the manuscript for important intellectual content: Eun Ran Kim, Sung Noh Hong, Young-Ho Kim, and Dong Kyung Chang.
ACCEPTED MANUSCRIPT Statistical analysis: Sun-Young Baek and Soo Hyun Ahn. Study supervision: Dong Kyung Chang.
RI PT
All authors approved the final submission.
Abbreviations used in this paper: BMI, body mass index; CRC, colorectal cancer; AN,
SC
advanced neoplasia; OR, odds ratio; CI, confidence intervals.
AC C
EP
TE D
M AN U
DISCLOSURE: The authors have no conflicts of interest to declare.
ACCEPTED MANUSCRIPT ABSTRACT Background and Aims: In individuals with either no or 1 to 2 nonadvanced adenomas, future risks of advanced colorectal neoplasia (AN) vary according to the clinical risk factors.
RI PT
However, little is known about the association between the adenoma detection rate (ADR) and the risk for metachronous AN in patients with low-risk adenomas.
Methods: We identified 7,171 participants with no or 1-2 non-advanced adenomas at first-
SC
time screening colonoscopy. The risk of metachronous AN was investigated at surveillance
M AN U
colonoscopy, according to clinical characteristics and the ADR.
Results: In multivariate analysis, the risk for metachronous AN was strongly associated with increasing age, male sex, increasing number of adenoma and the ADR of the endoscopist. With the ADR modeled as a continuous variable, each 1.0% increase in the rate of ADR predicted a 3.0% decrease in the risk of metachronous AN (adjusted odds ratio [aOR], 0.97;
TE D
95% confidence CI [CI], 0.95-0.99). With the ADR modeled using a binary cut-off (32%), the risk of metachronous AN was reduced in patients of endoscopists with an ADR ≥32% (aOR, 0.53; 95% CI, 0.35-0.83). Moreover, the risk of metachronous AN was reduced (aOR,
EP
0.66; 95% CI, 0.46-0.95) in patients of endoscopist with an ADR in the highest tertile,
AC C
compared to patients of endoscopists with ADR in the lowest tertile. The impact of ADR on metachronous AN was significant for patients with low-risk adenomas rather than patients with no adenoma.
Conclusions: In patients with low-risk adenomas, the ADR of the endoscopist was inversely associated with the risk of metachronous AN. Key words: Colonoscopy; Surveillance; Adenoma detection rate; Metachronous advanced neoplasia.
ACCEPTED MANUSCRIPT INTRODUCTION Because the majority of colorectal cancers (CRCs) develop from benign adenomas, colonoscopic screening and surveillance have been very effective in preventing CRC.1-14
RI PT
Despite the benefits of screening and surveillance, uncertainty remains with respect to surveillance policies for patients with low-risk adenomas, defined as one or two small (<1 cm) adenomas. Current guidelines recommend surveillance colonoscopy after 10 years or surveillance in 5 to 10 years for individuals with a history of no or 1 to 2 non-advanced
SC
adenomas.15-18 Korean guidelines recommend surveillance after 5 years in low-risk patients.17
M AN U
Recent studies have focused on the time interval between screening and surveillance colonoscopy, with the interval for low-risk patients having been lengthened.7,9,19,20 There is increasing evidence that most low-risk patients are also at low risk for subsequent CRC incidence and mortality, as well as being at low risk for advanced colorectal neoplasia (AN) at follow-up.9,12,20,21 In practice, however, endoscopists do not routinely follow guidelines and
TE D
many recommend shorter colonoscopy surveillance intervals for low-risk individuals.22 The principal reason to recommend a shorter interval might be the concern of interval CRC.
EP
Interval CRC is defined as CRC diagnosed after a previous colonoscopy and before the time of the next recommended examination.23 The majority of interval CRCs are thought to
AC C
result primarily from missed lesions.24-26 Therefore, high-quality colonoscopy is important for preventing interval colorectal neoplasia. Current guidelines for screening and surveillance colonoscopy also mention that high-quality colonoscopy should be the standard.16 Several studies have reported that future risk for AN can vary according to patient- and adenomarelated factors among low-risk patients.21,22,27 However, these studies did not consider colonoscopy quality-related factors, such as the ADR, bowel preparation and withdrawal time.21,27 To the best of our knowledge, there has been no prior large-scale study to evaluate the
ACCEPTED MANUSCRIPT risk for occurrence of metachronous AN among low-risk patients using patient- and adenoma-related characteristics and the colonoscopy quality indicators. The aims of this study were to determine the risk for metachronous AN among individuals with no or low-risk
RI PT
adenomas, and to evaluate whether the ADR determined by endoscopists influences the risk of metachronous AN.
SC
METHODS Study population
M AN U
This retrospective cohort study included healthy adults, aged >45 years old, who participated in a comprehensive health-screening exam at the Center for Health Promotion of the Samsung Medical Center, South Korea, from January 2005 to December 2011 (Figure 1). Because our objective was to evaluate the risk of metachronous colorectal neoplasia
TE D
according to baseline patient characteristics, adenoma characteristics and index colonoscopy quality, we included participants who underwent two colonoscopies, namely a first-time colonoscopy for screening and a second colonoscopy for surveillance (n = 9,547). We
EP
excluded 2,376 participants who met any of the following criteria: poor bowel preparation (n = 509), incomplete colonoscopy (n = 45), history of malignancy (n = 150), history of
AC C
colorectal surgery (n = 9), inflammatory bowel disease (n = 5), and other findings, except low-risk findings, such as no or 1-2 small (<1 cm) low-risk adenomas at index colonoscopy (n = 1,719). Finally, 7,171 participants with no or low-risk colorectal adenomas were included in this study.
This study was approved by the Institutional Review Board of the Samsung Medical Center, and the requirement for informed consent was waived because only de-identified data routinely collected during health screening visits were used.
ACCEPTED MANUSCRIPT Data collection The comprehensive health-screening program included demographic characteristics, anthropometric measurements, detailed physical examination, and a self-administered health
RI PT
questionnaire on smoking, alcohol intake, current medication use, family history of CRC and personal medical history including malignancy, previous colorectal polyps, and surgical history.28 Smoking status was categorized as light smoker (<20 pack years) and heavy
SC
smoker (≥20 pack years). Alcohol consumption was divided into mild (≤10 g/day) and modest (>10 g/day). Height and weight were measured to the nearest 0.1 cm and 0.1 kg,
M AN U
respectively. Each participant’s weight and height was measured with light clothing and in bare feet. Body mass index (BMI) was calculated as the weight (kg) / height (m2).
Colonoscopies
TE D
Thirty-one experienced, board-certified, gastroenterologists performed the colonoscopies after bowel preparation with 4 L of polyethylene glycol. Median year of graduation from medical school of 31 gastroenterologists is 1999 (range, 1989-2001). In this 7,171 study
EP
sample, median number of exams per gastroenterologist was 312 (range, 111-779). The location of the colorectal neoplasm was assessed by the endoscopists, and the size of each
AC C
lesion was estimated using an open biopsy forceps. The following information was recorded in the electronic medical record after the colonoscopy: number of polyps, location and size of polyps, time and results of the last colonoscopy, family history of CRC, bowel preparation (excellent/good/fair/poor), and completeness of the colonoscopy.28
Variables The primary outcome was incidence of metachronous advanced neoplasia (AN) at surveillance
colonoscopy
according
to
baseline
patient
characteristics,
adenoma
ACCEPTED MANUSCRIPT characteristics, and quality of index colonoscopy. AN was defined as any adenoma ≥1 cm, any adenoma with a villous component, high-grade dysplasia, or invasive cancer.18 Low-risk adenoma was defined as 1 or 2 adenomas, of <1 cm and without a villous component or high-
RI PT
grade dysplasia.18 Patients-related information used in the analysis included age, sex, BMI, smoking status, alcohol intake, regular use of aspirin, family history of CRC, bowel preparation, and follow-up interval. Adenoma-related characteristics included in the analysis
SC
were as follows; adenoma number, size, and location. Withdrawal time and ADR was used as quality indicators of the index colonoscopy. The ADR, determined by each endoscopist, was
M AN U
calculated from data of individuals ≥50 years old who underwent a first-time screening colonoscopy, except cases with poor bowel preparation.
Statistical analysis
TE D
Baseline characteristics of patients were summarized using conventional statistics. Continuous variables were reported as means ± standard deviation, whereas categorical variables were presented as percentages. Continuous variables were compared between
EP
groups using the Student t-test, whereas categorical variables were compared using the chisquared test. The main outcome was metachronous colorectal neoplasia at surveillance
AC C
colonoscopy. Baseline factors with a potential prognostic effect on metachronous colorectal neoplasia were initially analyzed by univariate and multivariate logistic regression analyzes. We conducted logistic regression models to estimate the adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for metachronous colorectal neoplasia associated with baseline patient, adenoma characteristics and the adenoma detection rate. Multivariate logistic regression, with step wise variable selection, was conducted by including all variables that were significant in univariate analysis. The candidate factors included age, sex, BMI, smoking status, alcohol consumption, regular aspirin use, family history of CRC, follow-up
ACCEPTED MANUSCRIPT interval, bowel preparation, number and size of baseline adenomas, withdrawal time, and the ADR of the endoscopist. We further included the ADR of the endoscopists, who performed the follow-up colonoscopy, as a potential confounding factor. We modeled the ADR as a
RI PT
continuous variable, as well as considering binary and tertile cutoffs. The Youden index, sensitivity, specificity, and positive and negative predictive values were calculated. From the highest Youden index, we selected an optimal cutoff point of ADR for predicting
SC
metachronous colorectal neoplasia.
Finally, we built a classification tree to predict the ‘hidden’ risk of metachronous AN
M AN U
among the low-risk patients. We split the root node, or each of the child nodes of the root node, by a single variable selected by recursive partitioning, a process that minimizes the sum of the node impurities in the two child nodes. In this model, age, sex, number of adenomas, and the ADR were considered inputs. The P value at each node measured the association
TE D
between the selected input variable and the response (incidence of AN). The terminal nodes were grouped according to the response (risk of metachronous AN), and the results were presented as the final set of prognostic factors. A P value < 0.05 was considered statistically
EP
significant. Statistical analyzes were performed using R version 3.3.1 (The R Foundation for
AC C
Statistical Computing, Vienna, Austria) and SAS version 9.4 (SAS Institute, Cary, NC).
RESULTS
Baseline characteristics of the study participants The baseline characteristics of the 7,171 study participants with no or low-risk adenomas are shown in Table 1. The mean (SD) age of participants was 52.7 (6.1) years, with 61.4% being male. Participants with low-risk adenomas were more likely to be older, male, have a higher BMI, and to be heavy smoker and modest alcohol consumers, compared to those without adenomas. However, no between-group difference in aspirin use and family history of CRC
ACCEPTED MANUSCRIPT was observed.
The risk of any metachronous colorectal adenomas
RI PT
Among the 7,171 participants with no or low-risk colorectal adenomas, any colorectal adenomas developed in 2,291 participants at 3 to 5 years after screening colonoscopy, with an incidence rate of 31.9%. In univariate logistic regression analysis, increasing age, male sex,
SC
increasing BMI, heavy smoking, modest alcohol intake, family history of CRC, longer duration of follow-up interval, increasing number of adenoma, larger size of adenoma,
M AN U
withdrawal time, the ADR of the endoscopists and the ADR of the endoscopist performing the follow-up endoscopy were significant predictors for any metachronous colorectal adenoma at surveillance colonoscopy. In the multivariate analysis, age (OR, 1.04; 95% CI, 1.03-1.05), male sex (OR, 1.65; 95% CI, 1.38-1.98), heavy smoking (OR, 1.23; 95% CI,
TE D
1.04-1.45), family history of CRC (OR, 1.36; 95% CI, 1.09-1.69), longer duration of followup interval (OR, 1.01; 95% CI, 1.00-1.01), increasing number of adenoma (OR, 1.68; 95% CI, 1.54-1.83), withdrawal time (OR, 0.97; 95% CI, 0.95-0.99), ADR of the endoscopist (OR,
EP
0.97; 95% CI, 0.96-0.99), and ADR of the endoscopist who performed the follow-up colonoscopy (OR, 1.03; 95% CI, 1.02-1.05) were significant predictors of any metachronous
AC C
colorectal adenoma (Table 2).
The risk of metachronous advanced neoplasia Among the 7,171 participants, AN developed in 154 participants, an incidence rate of 2.1%. In the univariate analysis, increasing age, male sex, heavy smoking, modest alcohol intake, increasing number of adenoma, proximal location of adenoma, withdrawal time, ADR of the endoscopist, and ADR of the endoscopist who performed the follow-up colonoscopy were significantly associated with the risk of metachronous AN. In multivariate analysis,
ACCEPTED MANUSCRIPT increasing age (OR, 1.05; 95% CI, 1.02-1.07), male sex (OR, 1.57; 95% CI, 1.13-2.20), increasing number of adenoma (OR, 1.90; 95% CI, 1.56-2.31), ADR of the endoscopist (OR, 0.97; 95% CI, 0.95-0.99), and ADR of the endoscopist who performed the follow-up
RI PT
colonoscopy (OR, 1.03; 95% CI, 1.01-1.06) were retained as independent predictors of metachronous AN (Table 3).
Recursive partitioning analysis of prognostic factors demonstrated that the risk of
SC
metachronous AN at follow-up was stratified based on number of adenoma, age, and the ADR of the endoscopist. These variables were also retained as significant factors for
M AN U
metachronous AN in the multivariate logistic regression model. Among the terminal nodes with different risk of metachronous AN, the highest-risk node had an 8.1% risk of incident AN, with all other nodes have a risk of AN of <5%. The highest-risk node was classified according to number of adenoma and the ADR of the endoscopist, and included patients with
TE D
2 adenomas who were examined by endoscopists with a low ADR (<32%). Aggregating the remaining nodes, the risk of metachronous AN at follow-up was 1.7%. The difference in risk of metachronous AN between the 2 groups was significant (OR, 3.98; 95% CI, 2.76-5.74; P <
EP
0.001).
AC C
Adenoma detection rate and risk of metachronous colorectal neoplasia We calculated the ADR for each of the 31 endoscopists in this study. The ADRs of the endoscopists were distributed from 18.6% to 42.0%, with a median ADR of 30.6% (Supplementary Figure 1). Mean withdrawal time of the normal screening colonoscopy was significantly related to the ADR (coefficient = 0.597, standard error = 0.027, P < 0.001). With the ADR modeled as a continuous variable, each 1.0% increase in the rate predicted a 3.0% decrease in the risk of metachronous AN (adjusted odds ratio [aOR], 0.97; 95% confidence CI [CI], 0.95-0.99) (Table 4). We selected an optimal cutoff point of ADR for predicting
ACCEPTED MANUSCRIPT metachronous colorectal neoplasia, according to the Youden index. The cutoff point of ADR was 29% for any metachronous colorectal neoplasia and 32% for metachronous AN, respectively. We then dichotomized the ADR into a low and high ADR from the optimal
RI PT
cutoff point for predicting metachronous colorectal neoplasia. With the ADR modeled by binary cut-off (32%), the risk of metachronous AN was reduced in patients of endoscopists with an ADR ≥32% (aOR, 0.53; 95% CI, 0.35-0.83). When the endoscopist ADR was
SC
considered in tertiles, the risk for metachronous AN was lower for patients of endoscopists with an ADR in the upper tertile than in patients of endoscopists with an ADR in the lowest
M AN U
tertile (aOR, 0.66; 95% CI, 0.46-0.95). Similarly, the risk was lower for patients of endoscopists with an ADR in the middle compared to lowest tertile (aOR, 0.68; 95% CI, 0.47-0.99). In subgroup analyzes, the ADR was not a significant predictor for metachronous AN among patients with no adenoma; however, the ADR was a significant predictor among
TE D
patients with low-risk adenoma (Table 5). The subgroup analysis, however, did not reveal a significant interaction between baseline adenomas (no adenoma versus low-risk adenoma; P
DISCUSSION
EP
for interaction = 0.998).
AC C
To the best of our knowledge, this is the first large-scale study to investigate prognostic factors for occurrence of metachronous AN in low-risk patients using a number of readily available clinical characteristics and ADR as a colonoscopy quality indicator. Current guidelines for surveillance colonoscopy of low-risk individuals at screening colonoscopy are heterogeneous. For individuals with 1 to 2 low-risk adenomas, recommendations for surveillance range from routine screening after 10 years,15,16 to performance of surveillance at 5 to 10 years,18 to a 5-year surveillance.17 Heterogeneity among the guidelines may be due to uncertainty of risk for metachronous colorectal neoplasia among the large and heterogeneous
ACCEPTED MANUSCRIPT group of low-risk patients. Indeed, we found that the risk for metachronous colorectal neoplasia varied according to patient characteristics, adenoma characteristics, and quality of index colonoscopy. Results from multivariate analysis demonstrated that increasing age, male
RI PT
sex, increasing number of adenoma, and the ADR of the endoscopist were independent predictors of metachronous any adenomas or AN at follow- up colonoscopy. Considering the ADR as a continuous, binary or ternary variable, a low ADR was a significant risk factor for
SC
metachronous AN. From the results of recursive partitioning analysis, patients with 2 small adenomas who were screened by endoscopists with a low ADR were identified as a hidden-
M AN U
risk group. Our results reinforced the importance of the quality of the screening colonoscopy for preventing the incidence of metachronous AN.
Several previous studies also investigated risk stratification for surveillance in low-risk patients, according to clinical characteristics.21,27 The studies revealed that several patient or
TE D
adenoma characteristics were significantly associated with incidence of metachronous AN. However, the studies did not consider quality factors of the index colonoscopy, such as bowel preparation, withdrawal time, and ADR of the endoscopists who performed the screening
EP
colonoscopy. The recommended surveillance guideline applies to individuals who underwent high-quality colonoscopy, such as complete colonoscopy with a meticulous inspection of
AC C
adequately cleaned colorectal mucosa.16 In the multivariate analysis, we included potential confounding factors. Additionally, we considered the ADR of the endoscopists, which is the most important quality indicator of interval colorectal neoplasia. Indeed, a low ADR was an independent predictor for both any metachronous colorectal neoplasia and metachronous AN. Even among low-risk patients, a low ADR was significantly associated with risk of metachronous AN. These results underline how variation in the performance of colonoscopy
ACCEPTED MANUSCRIPT may influence its effectiveness with respect to detection and prevention of colorectal neoplasia. Our results have potential clinical implications. The risk stratification may be due to
RI PT
variations in risk for metachronous colorectal neoplasia according to patient-related and adenoma-related characteristics, and quality of the index colonoscopy, even in the low-risk group. The results of our recursive partitioning analysis demonstrated that presence of 2
SC
adenomas and performance by an endoscopist with a low ADR increased the risk for metachronous AN. We defined the group with a higher risk for metachronous AN as the
M AN U
hidden-risk group among low-risk patients. Our results underlined the importance of the ADR as a prognostic factor for incident AN, to the same extent as the number, size and histology of adenomas at baseline.
The impact of ADR on metachronous advanced colorectal neoplasia was significant for
TE D
patients with low-risk adenomas rather than patients with no adenoma. This is probably due to the very low risk of metachronous advanced neoplasia in individuals without any adenomas. Indeed, in this study, the incidence of advanced colorectal neoplasia was less than
EP
1.7% in individuals with no adenoma at baseline. Therefore, the risk of metachronous
AC C
advanced neoplasia was absolutely low that this group would not have been affected by ADR. The ADR benchmark of 32% in this study was higher than the recommended ADR benchmark of 25% for all patients (30% for men and 20% for women).29 This discrepancy may be due to differences in the prevalence of colorectal adenoma according to geographic area. Recent data from United States reported a prevalence rate of colorectal adenomas of 20% to 23% among individuals 50- to 59-years-old.30 A recent Korean study reported an overall prevalence rate of 37.8% among individuals 50- to 59-years-old.31 In areas with high
ACCEPTED MANUSCRIPT prevalence rates of colorectal adenomas, the ADR benchmark may also increase accordingly. Future studies will be required to investigate these findings. Several limitations should be considered in interpreting our findings. First,
RI PT
misclassification might have occurred during the measurement of polyp size and histological diagnosis. However, as endoscopists and pathologists were unaware of the study objectives, these measurement errors would be non-differential and could possibly result in an
SC
underestimation of the association between clinical characteristics and metachronous colorectal neoplasia. Second, our study was performed at a single center and focused on
M AN U
asymptomatic Korean men and women; therefore, it may be difficult to generalize our findings to other populations. Third, several co-variates, such as smoking habit, alcohol intake and medical history, might be inaccurate based on self-report in questionnaire survey or medical records. Finally, the main outcome was metachronous lesions at 3 to 5 years. In
TE D
the current guidelines, surveillance colonoscopy was recommended at 5 to 10 years among low-risk individuals. Therefore, it is unclear whether there is a substantial difference in the risk of metachronous AN related to differences in surveillance intervals.
EP
This study also has considerable strengths, including its cohort design, large sample size,
AC C
inclusion of first-time colonoscopies, and consideration of colonoscopic quality indicators, such as bowel preparation, completeness of colonoscopy, indications for colonoscopy, and ADR of the endoscopists. Comprehensive risk variables, including patient-related data, adenoma-related data, and the quality indicators of colonoscopy, resulted in well-powered evaluations on the rare incidence of metachronous AN and more detailed analysis through multivariate adjustment. Additional strengths include the use of high quality, standardized clinical methods and a standardized self-administered health questionnaire completed by all participants.
ACCEPTED MANUSCRIPT In conclusion, there is substantial variation in the risk of metachronous AN according to readily available clinical characteristics and the ADR of the endoscopists among low-risk patients. An increasing number of adenomas, increasing patient age, male sex, and
RI PT
performing by an endoscopist with a low ADR were independent predictors for metachronous AN. The ADR was inversely associated with the risk of metachronous colorectal neoplasia, including any adenomas and advanced neoplasia in patients with low-risk adenomas. This
SC
association was consistently observed, even when the ADR was modeled either as a continuous, binary or ternary variable. Our findings support the validity of the ADR as a
M AN U
quality indicator of the performance of a screening colonoscopy even in low-risk patients. For low-risk patients who have undergone screening colonoscopy by an endoscopist with an appropriate ADR, long surveillance interval is likely to be sufficient. REFERENCES
4. 5.
6.
7. 8.
9. 10. 11.
TE D
3.
EP
2.
Brenner H, Chang-Claude J, Seiler CM, et al. Protection from colorectal cancer after colonoscopy: a population-based, case-control study. Ann Intern Med 2011;154:22-30. Kahi CJ, Imperiale TF, Juliar BE, et al. Effect of screening colonoscopy on colorectal cancer incidence and mortality. Clin Gastroenterol Hepatol 2009;7:770-5; quiz 11. Nishihara R, Wu K, Lochhead P, et al. Long-term colorectal-cancer incidence and mortality after lower endoscopy. N Engl J Med 2013;369:1095-105. Shaukat A, Mongin SJ, Geisser MS, et al. Long-term mortality after screening for colorectal cancer. N Engl J Med 2013;369:1106-14. Edwards BK, Ward E, Kohler BA, et al. Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;116:544-73. Atkin WS, Edwards R, Kralj-Hans I, et al. Once-only flexible sigmoidoscopy screening in prevention of colorectal cancer: a multicentre randomised controlled trial. Lancet 2010;375:1624-33. Brenner H, Chang-Claude J, Seiler CM, et al. Long-term risk of colorectal cancer after negative colonoscopy. J Clin Oncol 2011;29:3761-7. Holme O, Loberg M, Kalager M, et al. Effect of flexible sigmoidoscopy screening on colorectal cancer incidence and mortality: a randomized clinical trial. JAMA 2014;312:606-15. Imperiale TF, Glowinski EA, Lin-Cooper C, et al. Five-year risk of colorectal neoplasia after negative screening colonoscopy. N Engl J Med 2008;359:1218-24. Schoen RE, Pinsky PF, Weissfeld JL, et al. Colorectal-cancer incidence and mortality with screening flexible sigmoidoscopy. N Engl J Med 2012;366:2345-57. Segnan N, Armaroli P, Bonelli L, et al. Once-only sigmoidoscopy in colorectal cancer screening: follow-up findings of the Italian Randomized Controlled Trial--SCORE. J
AC C
1.
ACCEPTED MANUSCRIPT
17. 18.
19.
20.
21.
22.
23. 24. 25.
26.
27. 28. 29.
RI PT
SC
16.
M AN U
15.
TE D
14.
EP
13.
AC C
12.
Natl Cancer Inst 2011;103:1310-22. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale-Update based on new evidence. Gastroenterology 2003;124:544-60. Winawer SJ, Zauber AG, O'Brien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. The National Polyp Study Workgroup. N Engl J Med 1993;328:901-6. Zauber AG, Lansdorp-Vogelaar I, Knudsen AB, et al. Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force. Ann Intern Med 2008;149:659-69. 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 Suppl 3:SE151-63. Hassan C, Quintero E, Dumonceau JM, et al. Post-polypectomy colonoscopy surveillance: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2013;45:842-51. Hong SN, Yang DH, Kim YH, et al. [Korean guidelines for post-polypectomy colonoscopic surveillance]. Korean J Gastroenterol 2012;59:99-117. Lieberman DA, Rex DK, Winawer SJ, et al. Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2012;143:844-57. Noshirwani KC, van Stolk RU, Rybicki LA, et al. Adenoma size and number are predictive of adenoma recurrence: implications for surveillance colonoscopy. Gastrointest Endosc 2000;51:433-7. Ponugoti PL, Rex DK. Yield of a second screening colonoscopy 10 years after an initial negative examination in average-risk individuals. Gastrointest Endosc 2017;85:221-4. Chung SJ, Kim YS, Yang SY, et al. Five-year risk for advanced colorectal neoplasia after initial colonoscopy according to the baseline risk stratification: a prospective study in 2452 asymptomatic Koreans. Gut 2011;60:1537-43. Anderson JC, Baron JA, Ahnen DJ, et al. Factors Associated With Shorter Colonoscopy Surveillance Intervals for Patients With Low-Risk Colorectal Adenomas and Effects on Outcome. Gastroenterology 2017;152:1933-43 e5. Brenner H, Chang-Claude J, Seiler CM, et al. Interval cancers after negative colonoscopy: population-based case-control study. Gut 2012;61:1576-82. le Clercq CM, Bouwens MW, Rondagh EJ, et al. Postcolonoscopy colorectal cancers are preventable: a population-based study. Gut 2014;63:957-63. Bressler B, Paszat LF, Chen Z, et al. Rates of new or missed colorectal cancers after colonoscopy and their risk factors: a population-based analysis. Gastroenterology 2007;132:96-102. Singh H, Nugent Z, Demers AA, et al. Rate and predictors of early/missed colorectal cancers after colonoscopy in Manitoba: a population-based study. Am J Gastroenterol 2010;105:2588-96. Gupta S, Jacobs ET, Baron JA, et al. Risk stratification of individuals with low-risk colorectal adenomas using clinical characteristics: a pooled analysis. Gut 2015. Rampal S, Yang MH, Sung J, et al. Association between markers of glucose metabolism and risk of colorectal adenoma. Gastroenterology 2014;147:78-87 e3. Rex DK, Schoenfeld PS, Cohen J, et al. Quality indicators for colonoscopy. Gastrointest Endosc 2015;81:31-53.
ACCEPTED MANUSCRIPT
31.
Corley DA, Jensen CD, Marks AR, et al. Variation of adenoma prevalence by age, sex, race, and colon location in a large population: implications for screening and quality programs. Clin Gastroenterol Hepatol 2013;11:172-80. Koo JE, Kim KJ, Park HW, et al. Prevalence and risk factors of advanced colorectal neoplasms in asymptomatic Korean people between 40 and 49 years of age. J Gastroenterol Hepatol 2017;32:98-105.
FIGURE LEGENDS Figure 1. Flow diagram of study participants
SC
SUPPLEMENTAL MATERIAL
RI PT
30.
AC C
EP
TE D
M AN U
Supplementary Figure 1. Distribution of adenoma detection rate by endoscopists
ACCEPTED MANUSCRIPT Table 1. Baseline characteristics of individuals with none or low-risk adenomas No adenoma Low-risk adenomas All P value (n= 7,171) (n = 4,100) (n = 3,071) 52.7 ± 6.1
51.9 ± 5.7
53.9 ± 6.5
61.4 38.6 23.9 ± 2.8
57.0 43.0 23.7 ± 2.7
67.2 32.8 24.2 ± 2.9
59.1 40.9
62.7 37.3
<0.001 <0.001
RI PT
Age (years) Sex (%) Male Female BMI (kg/m2) Smoking status (%) <20 pack years ≥20 pack years Modest alcohol intake (%) Mild (≤10 g/day) Modest (>10 g/day) Regular aspirin use (%)
<0.001 <0.001
54.4 45.6
SC
<0.001
AC C
EP
TE D
M AN U
82.4 84.2 17.6 15.8 10.2 10.9 Family history of CRC (%) 6.9 7.1 Values are expressed as means ± standard deviation or percentages. BMI, body mass index; CRC, colorectal cancer.
80.1 19.9 9.6 6.5
0.055 0.391
ACCEPTED MANUSCRIPT Table 2. The risk of metachronous any colorectal adenomas according to baseline patient, adenoma characteristics, and quality indicator of index colonoscopy Univariate analysis Multivariate analysis OR (95% CI)
P value
1.04 (1.03-1.05) 1.97 (1.77-2.19) 1.07 (1.05-1.08)
<0.001 <0.001 <0.001
1.04 (1.03-1.05) 1.65 (1.38-1.98)
<0.001 <0.001
Smoking status 1.00 (reference)
1.00 (reference)
≥ 20 pack years
1.95 (1.70-2.23)
<0.001
Modest alcohol intake
1.46 (1.28-1.66)
<0.001
Regular aspirin use
1.01 (0.91-1.13)
0.798
Family history of CRC
1.38 (1.13-1.69)
0.002
Follow-up interval
1.01 (1.00-1.01)
0.001
Bowel preparation
1.00 (reference)
Good
0.98 (0.89-1.09)
0.731
Excellent
1.01 (0.83-1.24)
0.904
Number of adenoma Location of adenoma Distal location
1.73 (1.61-1.86)
<0.001
0.016
1.36 (1.09-1.69)
0.007
1.01 (1.00-1.01)
0.002
1.68 (1.54-1.83)
<0.001
1.00 (reference)
1.07 (0.92-1.25)
0.393
1.77 (1.37-2.28)
<0.001
EP
Proximal location Both location
TE D
Fair
Adenoma characteristics
1.23 (1.04-1.45)
M AN U
< 20 pack years
RI PT
P value
SC
Patient characteristics Age Male sex BMI
OR (95% CI)
<5
AC C
Size of adenoma (mm) 5 to <10
1.00 (reference) 1.30 (1.10-1.53)
0.002
Withdrawal time
0.97 (0.96-0.99)
0.013
0.97 (0.95-0.99)
0.016
ADR of the endoscopist
0.98 (0.97-0.99)
<0.001
0.97 (0.96-0.99)
<0.001
Quality indicator
Potential confounder ADR of the endoscopist who performed the 1.04 (1.03-1.05) <0.001 1.03 (1.02-1.05) <0.001 follow-up colonoscopy OR, odds ratio; CI, confidence intervals; BMI, body mass index; CRC, colorectal cancer; ADR, adenoma detection rate.
ACCEPTED MANUSCRIPT Table 3. The risk of metachronous advanced neoplasia according to baseline patient, adenoma characteristics, and quality indicator of index colonoscopy Univariate analysis Multivariate analysis OR (95% CI)
P value
1.06 (1.04-1.08) 2.19 (1.52-3.15) 1.04 (0.99-1.10)
<0.001 <0.001 0.126
1.05 (1.02-1.07) 1.57 (1.13-2.20)
<0.001 0.008
Smoking status 1.00 (reference)
≥ 20 pack years
1.61 (1.19-2.18)
0.002
Modest alcohol intake
1.52 (1.07-2.16)
0.018
Regular aspirin use
0.82 (0.59-1.13)
0.214
Family history of CRC
1.20 (0.69-2.10)
0.512
Follow-up interval
1.00 (0.99-1.01)
0.535
M AN U
< 20 pack years
Bowel preparation
1.00 (reference)
Good
1.22 (0.89-1.66)
0.674
Excellent
1.14 (0.62-2.07)
0.218
Number of adenoma Location of adenoma Distal location
2.03 (1.68-2.46)
<0.001
1.90 (1.56-2.31)
<0.001
0.97 (0.95-0.99)
0.026
1.00 (reference)
1.51 (1.00-2.29)
0.05
2.23 (1.26-3.96)
0.006
EP
Proximal location Both location
TE D
Fair
Adenoma characteristics
RI PT
P value
SC
Patient characteristics Age Male sex BMI
OR (95% CI)
<5
AC C
Size of adenoma (mm) 5 to <10
1.00 (reference) 1.36 (0.88-2.11)
0.166
Withdrawal time
0.98 (0.96-0.99)
0.046
ADR of the endoscopist
0.98 (0.96-1.00)
0.05
Quality indicator
Potential confounder ADR of the endoscopist who performed the 1.04 (1.01-1.06) 0.007 1.03 (1.01-1.06) 0.01 follow-up colonoscopy OR, odds ratio; CI, confidence intervals; BMI, body mass index; CRC, colorectal cancer; ADR, adenoma detection rate.
ACCEPTED MANUSCRIPT Table 4. The risk of metachronous colorectal neoplasia according to adenoma detection rate Any adenomas Advanced neoplasia P value
aORa (95% CI)
P value
0.97 (0.96-0.99)
<0.001
0.97 (0.95-0.99)
0.026
1.00 (reference)
High ADR
0.66 (0.56-0.78)
RI PT
By continuous variables ADR By binary cut-off Low ADR
aORa (95% CI)
1.00 (reference) <0.001
0.53 (0.35-0.83)
By tertiles 1.00 (reference)
Middle ADR
0.69 (0.57-0.82)
<0.001
Highest ADR
0.63 (0.52-0.75)
<0.001
a
1.00 (reference)
SC
Lowest ADR
0.005
0.68 (0.47-0.99)
0.044
0.66 (0.46-0.95)
0.025
M AN U
Estimated from logistic regression models. The multivariable analysis was adjusted for age, sex, BMI, smoking status, alcohol intake, regular aspirin use, family history of CRC, follow-up interval, bowel preparation, number of adenoma, size of adenoma, withdrawal time, and ADR of the endoscopist who performed the follow-up colonoscopy.
AC C
EP
TE D
OR, odds ratio; CI, confidence intervals; adenoma detection rate.
ACCEPTED MANUSCRIPT Table 5. The risk of metachronous advanced neoplasia by adenoma detection rate in no adenoma and low-risk adenoma Subgroup of no adenoma Subgroup of low-risk adenoma P value
aORa (95% CI)
P value
0.98 (0.94-1.01)
0.211
0.97 (0.95-1.00)
0.05
1.00 (reference)
High ADR
0.45 (0.18-1.07)
RI PT
By continuous variables ADR By binary cut-off Low ADR
aORa (95% CI)
1.00 (reference) 0.089
0.56 (0.34-0.93)
By tertiles 1.00 (reference)
Highest ADR
0.63 (0.33-1.08)
a
1.00 (reference) 0.102
SC
Lowest ADR
0.025
0.65 (0.41-0.98)
0.045
M AN U
Estimated from logistic regression models. The multivariable analysis was adjusted for age, sex, BMI, smoking status, alcohol intake, regular aspirin use, family history of CRC, follow-up interval, bowel preparation, number of adenoma, size of adenoma, withdrawal time, and ADR of the endoscopist who performed the follow-up colonoscopy.
AC C
EP
TE D
OR, odds ratio; CI, confidence intervals; adenoma detection rate.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Abbreviations: BMI, body mass index; CRC, colorectal cancer; AN, advanced neoplasia; OR,
AC C
EP
TE D
M AN U
SC
RI PT
odds ratio; CI, confidence intervals.