Significantly higher rates of multiple and proximally located adenomas among patients with diabetes mellitus: A cross-sectional population-based study

Abstract

Background

Diabetes mellitus (DM) is associated with a greater risk for colorectal cancer (CRC).

Objective

The objective of this article is to examine the endoscopic phenotype and histopathology of colorectal polyps in patients with vs without DM.

Methods

We conducted a cross-sectional study of patients who underwent colonoscopy at our university hospital and who completed a questionnaire. We collected endoscopy and histopathology data regarding colorectal adenomas and serrated polyps. Cox regression analyses were used to estimate adjusted prevalence ratios (PRs).

Results

We examined a total of 3654 patients (mean age (SD): 62 (12) years, 47% males). Of them, 360 (9.9%) had DM. Overall, the prevalence of colorectal adenomas (42% vs 32%, p < 0.01), multiple (≥3) adenomas (12% vs 7%, p = 0.01) and proximal adenomas (30% vs 19%, p < 0.01) was higher in patients with vs without DM. Multivariable analysis showed that the prevalence of adenomas (PR 1.17, 95% CI; 1.02–1.34), multiple (PR 1.37, 95% CI; 1.00–1.86) and proximal (PR 1.37, 95% CI; 1.16–1.62) adenomas was higher in patients with vs without DM, especially in men.

Conclusion

Patients with DM harbor more frequently multiple and proximal adenomas than those without DM. Close colonoscopic surveillance of DM patients is important to maximize the effectiveness of colonoscopic CRC prevention.

Keywords

Diabetes mellitus colorectal cancer colorectal neoplasm epidemiology colonoscopy prevention

Introduction

Patients with diabetes mellitus (DM) have a 20% to 40% greater risk of developing colorectal cancer (CRC) than non-diabetic individuals.¹ The gradual increase of the incidences both of DM and CRC over the past decade, with approximately 347 million patients with DM and 663,000 CRC cases worldwide in 2008,^2,3 raises concerns about the implications for CRC-prevention strategies. Increased life expectancy and shared risk factors (e.g. obesity, lack of physical activity and unfavorable dietary habits) may explain the rise in incidence rates of both diseases.⁴ The exact mechanisms underlying such risk remain, however, unclear. Insulin resistance in type 2 DM (T2DM) leads to chronic hyperinsulinism, possibly promoting the proliferation of colonic epithelial cells, either directly or indirectly, by increasing insulin-like growth factor (IGF), which, in turn, may lead to an accelerated progression from adenoma to cancer.⁵ This hypothesis is supported by a few studies demonstrating an increased risk of colorectal adenomas in T2DM patients, especially in those on exogenous insulin therapy.^6–8 It has been suggested that CRC patients with DM have a worse prognosis than those without DM, in particular a greater overall mortality (20%–40%) and a shorter CRC disease-free survival, perhaps due to suboptimal cancer therapy.⁹ Collectively, these data highlight the need for improvements in the prevention of CRC in diabetic patients. Although studies endorse the importance of CRC screening in diabetic patients,¹⁰ little is known about the prevalence, endoscopic phenotype (e.g. location, size, shape) and histopathology of colorectal polyps in diabetic patients. Also, information on age- and gender-specific variations in the prevalence of colorectal polyps in diabetic patients is sparse.¹¹ In view of the increasing number of CRC screening initiatives worldwide, it is important to clarify the magnitude of CRC risk in diabetic patients. Such information may provide the framework for customized screening programs, in terms of preferred test modality, age of initiation and frequency of surveillance intervals, with the goal of improving the effectiveness of CRC prevention. In a cross-sectional study, we therefore examined the prevalence and endoscopic phenotype of colorectal polyps in diabetic vs non-diabetic men and women, with special attention to the colonic subsite.

Materials and methods

Study population

In the framework of an ongoing initiative aiming to improve the quality of colonoscopy at our university hospital, we have previously trained all endoscopists on the detection, diagnosis and endoscopic resection of colorectal polyps, with emphasis on the non-polypoid colorectal lesions, as described elsewhere.¹² Special attention was paid to targets for measuring the quality of colonoscopy.¹³ From February 2008 to February 2012, we have included all consecutive patients who underwent an elective colonoscopy at the endoscopy unit of our university hospital for symptoms, screening or surveillance indications. We excluded patients aged <30 years, to preclude the inclusion of individuals with type 1 DM. Furthermore, patients with inflammatory bowel disease, hereditary CRC syndromes (i.e. known gene mutations), those fulfilling the World Health Organization (WHO) criteria for serrated polyposis syndrome, patients with a personal history of CRC, prior colonic resection or an incomplete colonoscopy were excluded. In patients who underwent more than one colonoscopy during the study period, data from the most extensive colonoscopy examination only were evaluated. Index colonoscopy data were used when a second colonoscopy was performed because of onset of new symptoms, polypectomy or post-polypectomy surveillance. In cases of incomplete first colonoscopy or poor bowel preparation requiring a second-look colonoscopy, data from the second colonoscopy were used. The study was approved by the local institutional review board (MEC 14-4-046).

Data collection

Clinical, endoscopic and histopathology data

Complete clinical, endoscopic and histopathology data were collected using a digital, standardized case report form. Bowel preparation was classified by the endoscopist as adequate or poor, as described previously.¹⁴ Colorectal polyps were divided according to location into proximal (i.e. proximal from and including the splenic flexure) and distal (i.e. distal from the splenic flexure). Size of the polyps was measured endoscopically (biopsy forceps or mini-snare) and classified into diminutive (i.e. <6 mm), small (i.e. 6–9 mm) and large (i.e. >9 mm). Multiple adenomas were defined as presence of at least three adenomas. Non-polypoid colorectal neoplasms were defined as lesions with a height less than half of their diameter.¹⁵ Histopathologic classification of all colorectal polyps was performed by two experienced gastrointestinal (GI) pathologists according to the WHO classification.¹⁶ Adenomas included tubular, tubulovillous and villous adenomas, while serrated polyps included hyperplastic polyps, sessile-serrated adenomas/polyps (SSA/Ps) with and without dysplasia and traditional serrated adenomas. Advanced adenomas were defined as adenomas sized ≥10 mm, containing high-grade dysplasia or any villous component.

Questionnaires

A self-administered questionnaire concerning demographic features, presence of DM, smoking, body mass index (BMI), medication (i.e. use of aspirin/nonsteroidal anti-inflammatory drugs (NSAIDs)), alcohol consumption, family history of CRC, and fruit and vegetable intake was obtained from all patients, shortly after the colonoscopy. Inconsistent or missing data were verified through medical records and recorded as missing in case such information could not be retrieved. Patients who provided informed consent and completed the questionnaire were included in the multivariable analyses. Patients were classified into diabetic and non-diabetic based on self-reporting. BMI was calculated using height and weight. A positive family history of CRC was defined as ≥1 first-degree relatives with CRC.¹⁷ Alcohol use was categorized as no use/sporadic use, 1–2 alcoholic consumptions per day and >2 alcoholic beverages per day. Smoking status was categorized as current, former or never and medication use was categorized as daily vs non-daily/no use. Information regarding fruit and vegetable intake was derived from the question: “Do you consume the following products (e.g. bread, fruit)?” When answered by yes, the patient was asked to assess the consumption of fruit in pieces per week and the amount of vegetables in grams per day. Indications for aspirin use in this population were primary or secondary prevention of cardiovascular events, for which standard daily doses of 100 mg were used.

Statistical analysis

Primary endpoints of this study were the prevalence of colorectal polyps, either adenomas or serrated polyps, in patients with vs without DM. Based on previous studies,^7,18 we estimated that a sample size of at least 155 diabetic women and 1350 non-diabetic women would be required to detect a 10% difference in polyp prevalence (30% vs 20%) with a power of 80% and an alpha of 0.05 in women. As polyps are more common in men than in women,¹⁸ we considered higher polyp prevalence (35% vs 25%) in the power analysis for men. We assumed that a sample size of 176 diabetic men and 1535 non-diabetic men would be sufficient to detect a 10% difference. Secondary endpoints were the characterization of the endoscopic features (e.g. location, size, shape) and histopathology of colorectal polyps in DM vs non-DM stratified by gender. Prevalence ratios (PRs) of colorectal polyps in DM vs non-DM patients were calculated using Cox-regression analyses with equal times of follow-up assigned to all individuals and robust variance as described previously by Barros and Hirakata.¹⁹ Possible confounders were selected a priori, based on clinical relevance and previous literature data. Multivariable models were adjusted for age at the time of colonoscopy, gender, BMI, family history of CRC, smoking status, alcohol intake, use of NSAIDS/aspirin, indication for colonoscopy, bowel preparation, and fruit and vegetable intake. Individuals with missing data were excluded from the analyses. In order to compare the adenoma prevalence in the distal vs proximal colon independent of the number of polyps found on colonoscopy, sensitivity analyses were performed excluding patients with more than one adenoma found on colonoscopy. Differences in continuous variables were analyzed using Mann-Whitney U and the independent samples t-test, differences in categorical values with the Chi-square test. Two-sided p values ≤ 0.05 were considered statistically significant and all risk ratios are presented with 95% confidence intervals (CIs). Statistical analyses were conducted using Stata (version 12, StataCorp, College Station, TX, USA).

Results

Figure 1 shows the study diagram. A total of 9363 patients with 11,067 colonoscopies were included. Of them, 6394 (68.3%) patients returned the questionnaire, of whom 6.7% declined to fill out the questionnaire, leaving 5771 participants further examined in this study. Participating responders were on average five years older than non-responders (mean (SD) age 60 (15) years vs 55 (19), p < 0.01) and had more often ≥1 adenoma (30% vs 23%, p < 0.01), ≥1 serrated polyp (16% vs 11%, p < 0.01) and ≥1 SSA/P (1.0% vs 0.7%, p = 0.241).

Figure 1.

Study flowchart.

Clinical characteristics

A total of 3654 (mean (SD) age 62 (12) years, 47% males) individuals were included in the final analyses, of whom 360 (9.9%) reported to have DM. Of all participants, 1203 individuals (33%) had ≥1 adenoma and 625 individuals (17%) ≥1 serrated polyp. Clinical characteristics of the study population subdivided according to presence of DM are described in Table 1. Indications for colonoscopy were symptoms (79.8% of cases), screening (9.5%) or surveillance (10.7%).

Table 1.

Clinical characteristics and risk profile of the study participants stratified according to type 2 diabetes mellitus status

Characteristic	Total	Type 2 DM	No type2 DM	p value
Characteristic	n = 3654	n = 360 (9.9%)	n = 3294 (90.1%)	DM vs non-DM
Mean (SD) age at colonoscopy (years)	61.6 (11.7)	66.0 (9.8)	61.1 (11.8)	<0.01
Gender (% male)	47.4	53.9	46.6	<0.01
Indication for colonoscopy (%)				0.01
Symptoms	79.8	81.9	79.6
Screening	9.5	5.3	9.9
Surveillance	10.7	12.8	10.5
Bowel preparation (%)				0.02
Adequate	94.2	91.4	94.5
Poor	5.8	8.6	5.5
Mean (SD) BMI (kg/m²)	26.1 (4.2)	28.9 (5.1)	25.7 (4.0)	<0.01
Family history of CRC (% yes)	24.9	19.7	25.5	0.02
Smoking status (%)				<0.01
Never	36.3	28.3	36.3
Ex-smoker	48.3	57.8	48.3
Current smoker	15.4	13.9	15.4
Alcohol use (%)				<0.01
No or sporadic use	59.9	74.2	58.3
1–2 Alcoholic consumptions a day	33.7	20.0	35.2
>2 Alcoholic consumptions a day	6.4	5.8	6.5
Use of NSAIDs/aspirin (% yes)	29.6	43.9	28.1	<0.01
Mean (SD) vegetable intake (grams/day)	187 (81)	187 (86)	187 (80)	0.56
Mean (SD) fruit intake (pieces/week)	6.6 (3.6)	6.7 (3.6)	6.6 (3.6)	0.36

BMI: body mass index; DM: diabetes mellitus; CRC: colorectal cancer; NSAIDs: nonsteroidal anti-inflammatory drugs.

Notably, patients with DM were more likely to have a poor bowel preparation than those without DM (9% vs 6%, p = 0.02). Compared with non-diabetic patients, patients with DM were older (mean age (SD) 61 (12) years vs 66 (10) years, p < 0.01), had a higher mean BMI (25.7 kg/m² vs 28.9 kg/m², p < 0.01), and used more often NSAIDS/aspirin daily (28% vs 44%, p < 0.01). Concerning lifestyle factors, diabetic patients were more often ex-smokers (58% vs 48%, p < 0.01) than the non-diabetic patients. Approximately 74% of the diabetic patients reported never/sporadic use of alcohol compared with 58% in the non-diabetic population (p < 0.01). No significant differences were found between diabetic and non-diabetic patients regarding reported fruit and vegetable intake.

In the analyzed population, the prevalence of colorectal polyps and CRC differed between men and women. Compared to women, the prevalence of colorectal adenomas (33% vs 25%, p < 0.01), advanced adenomas (17% vs 11%, p < 0.01) and CRC (4.1% vs 2.4%, p < 0.01) was higher in men. No differences were found between men and women regarding serrated polyps or SSA/P.

Prevalence ratios of colorectal polyps in patients with vs without DM in relation to subsite, number of adenomas, and histopathology

In Table 2 data regarding the prevalence and prevalence ratios of colorectal polyps in patients with vs without DM are presented. Overall, the prevalence of colorectal adenomas (42% vs 32%, p < 0.01), multiple (≥3) adenomas (12% vs 7%, p = 0.01) and proximal adenomas (30% vs 19%, p < 0.01) were higher in patients with than in those without DM. No differences were found between diabetic and non-diabetic patients regarding advanced adenomas, serrated polyps and SSA/Ps.

Table 2.

Total and gender-specific prevalence ratios of colorectal neoplasms in patients with vs patients without diabetes mellitus

Colorectal polyps	Total		Men		Women
	Non-DM (n = 3294)	DM (n = 360)	Non-DM (n = 1536)	DM (n = 194)	Non-DM (n = 1758)	DM (n = 166)
≥1 Adenoma
Cases (%)	1053 (32%)	150 (42%)	625 (41%)	95 (49%)	428 (24%)	55 (33%)
PR^d	1 (ref)	1.15 (1.01–1.31)	1 (ref)	1.11 (0.95–1.29)	1 (ref)	1.22 (0.96–1.55)
PR^e	1 (ref)	1.17 (1.02–1.34)	1 (ref)	1.12 (0.96–1.30)	1 (ref)	1.29 (1.02–1.65)
≥1 Distal adenoma^a
Cases (%)	694 (21%)	91 (25%)	398 (26%)	53 (27%)	296 (17%)	38 (23%)
PR^d	1 (ref)	1.07 (0.88–1.30)	1 (ref)	0.97 (0.76–1.25)	1 (ref)	1.23 (0.91–1.67)
PR^e	1 (ref)	1.10 (0.90–1.34)	1 (ref)	0.99 (0.76–1.27)	1 (ref)	1.34 (0.98–1.82)
≥1 Proximal adenoma^b
Cases (%)	621 (19%)	109 (30%)	397 (26%)	76 (39%)	224 (13%)	33 (20%)
PR^d	1 (ref)	1.37 (1.16–1.62)	1 (ref)	1.37 (1.13–1.66)	1 (ref)	1.36 (0.97–1.90]
PR^e	1 (ref)	1.37 (1.16–1.62)	1 (ref)	1.40 (1.15–1.69)	1 (ref)	1.35 (0.96–1.90)
≥3 Multiple adenoma
Cases (%)	229 (7%)	43 (12%)	152 (10%)	32 (16%)	77 (4%)	11 (7%)
PR^d	1 (ref)	1.40 (1.03–1.90)	1 (ref)	1.44 (1.01–2.06)	1 (ref)	1.29 (0.70–2.37)
PR^e	1 (ref)	1.37 (1.00–1.86)	1 (ref)	1.44 (1.02–2.05)	1 (ref)	1.18 (0.59–2.33)
≥1 NP adenoma
Cases (%)	192 (6%)	32 (9%)	114 (7%)	20 (10%)	78 (4%)	12 (7%)
PR^d	1 (ref)	1.27 (0.89–1.87)	1 (ref)	1.21 (0.77–1.90)	1 (ref)	1.40 (0.77–2.56)
PR^e	1 (ref)	1.31 (0.90–1.90)	1 (ref)	1.21 (0.76–1.93)	1 (ref)	1.50 (0.81–2.79)
≥1 Advanced adenoma^c
Cases (%)	441 (13%)	59 (16%)	250 (16%)	41 (21%)	191 (11%)	18 (11%)
PR^d	1 (ref)	1.05 (0.82–1.34)	1 (ref)	1.15 (0.86–1.55)	1 (ref)	0.87 (0.55–1.38)
PR^e	1 (ref)	1.04 (0.80–1.34)	1 (ref)	1.15 (0.84–1.56)	1 (ref)	0.89 (0.56–1.41)
≥1 Serrated polyp
Cases (%)	555 (17%)	70 (19%)	272 (18%)	43 (22%)	283 (16%)	27 (16%)
PR^d	1 (ref)	1.13 (0.90–1.41)	1 (ref)	1.22 (0.92–1.63)	1 (ref)	1.01 (0.70–1.45)
PR^e	1 (ref)	1.03 (0.82–1.29)	1 (ref)	1.21 (0.90–1.61)	1 (ref)	0.85 (0.59–1.24)

DM: diabetes mellitus; PR: prevalence ratio; NP: non-polypoid; HGD: high-grade dysplasia; CRC: colorectal cancer; NSAID: nonsteroidal anti-inflammatory drugs; ref: reference.

Distal from the splenic flexure. ^bProximal from the splenic flexure. ^cAdenoma with HGD, villous component, or size ≥10 mm.

PR adjusted for age. ^ePR adjusted for age, colonoscopy indication, bowel preparation, BMI, family history of CRC, smoking status, alcohol use, NSAID/aspirin use, fruit and vegetable intake.

Age-adjusted analyses showed that DM was associated with an increased prevalence ratio for having ≥1 adenoma (PR 1.15, 95% CI 1.01–1.31), proximal adenomas (PR 1.37, 95% CI 1.16–1.62) and multiple adenomas (PR 1.40, 95% CI 1.03–1.90). In the multivariable adjusted model presence of DM was significantly associated with greater prevalence ratios of adenomas (PR 1.17, 95% CI: 1.02–1.34), proximal adenomas (PR 1.37, 95% CI: 1.16–1.62) and multiple adenomas (PR 1.37, 95% CI: 1.00–1.86). Because both the prevalence rates of multiple adenomas and proximal adenomas were increased in patients with DM, a sensitivity analysis was conducted excluding patients with more than one adenoma. The prevalence of proximal adenoma remained increased (PR 1.34 (95% CI: 0.99–1.82)) in diabetic (14%) vs non-diabetic (9%) patients without multiple adenomas. Multivariable analysis showed similar results (PR 1.43 (95% CI: 1.05–1.94)).

Age- and gender-specific associations of DM with colorectal polyps

In Figure 2 age group-specific distribution of the prevalence of colorectal adenomas is depicted for 442 diabetic patients and 3511 non-diabetic individuals. For all age groups, the prevalence of ≥1 adenoma was higher in patients with vs without DM, though only statistically significant in the 65–75 years age group. Concerning patients in the pre-screening age range (45–55 years), those with DM had again a higher prevalence of adenomas than those without DM (29% vs 24%, p = 0.423).

Figure 2.

Age-specific prevalence of ≥1 colorectal adenoma, with 95% CI, in patients with and patients without diabetes mellitus. Because only missing data on age and diabetes status were excluded, the total number of individuals is higher than that in the multivariate analysis.

With regard to gender-specific associations, Table 2 shows that in men with DM, the prevalence of proximal (PR 1.40, 95% CI 1.15–1.69) and multiple adenomas (PR 1.44, 95% CI 1.02–2.05) were significantly increased in multivariable adjusted models compared to men without DM. In diabetic women, the prevalence of ≥1 adenoma (PR 1.22, 95% CI 0.96–1.55), proximal adenomas (PR 1.36, 95% CI 0.97–1.90), as well as distal adenomas (PR 1.23, 95% CI: 0.91–1.67), were again higher than in non-diabetic women, though differences did not achieve statistical significance.

Discussion

In this population-based study of patients undergoing elective colonoscopy, we found that diabetic men and women have a moderately increased prevalence of adenomas compared with non-diabetic individuals, in particular multiple and proximally located adenomas, which may partly explain the greater risk for developing CRC. Our findings underscore the importance of careful colonoscopic examination and close surveillance of diabetic patients to optimize the quality and effectiveness of CRC prevention in this higher-risk subgroup.

The findings of our study are of potential relevance for the post-polypectomy surveillance of diabetic patients, as both the multiplicity²⁰ and proximal location of colorectal adenomas²¹ are independent predictors for the development of metachronous neoplasms and are therefore incorporated in calculating the post-polypectomy surveillance intervals. Proximally located colorectal neoplasms appear to be often non-polypoid in shape,²² which partly explains why such lesions are frequently overlooked,²³ especially in patients with poor bowel preparation. Indeed, in our present study, 12% of the diabetic patients had multiple, often proximally located adenomas. Of all DM cases with detected adenomas, 21% had at least one nonpolypoid-shaped adenoma. Of note, nearly one in 10 diabetic patients in our study lacked adequate bowel preparation during the colonoscopic examination. Taken together, these data emphasize the importance of high-quality standards in screening and colonoscopy surveillance of diabetic patients to ensure protection against cancer.

Epidemiologic studies^24,25 previously described an increased risk for CRC in diabetic patients and a higher risk for proximally compared to distally located CRC, although data are conflicting. In the quality colonoscopy era, however, few studies examined the endoscopic phenotype (site, multiplicity, shape) of colorectal neoplasms in diabetic patients. An association was reported between the presence of proximal adenomas in diabetic patients and the duration of insulin use and fasting insulin levels: The study by Wong et al.⁸ showed an increased risk of proximal adenomas (odds ratio (OR) 1.8, 95% CI 1.1–2.9) in diabetic patients after at least two years of exposure to exogenous insulin compared to diabetic patients who were not exposed to exogenous insulin. Yoshida et al.²⁶ observed an increased risk for proximally located polyps with each 5 µU/ml increment in fasting serum insulin levels at the time of colonoscopy (OR 1.8, 95% CI 1.2–2.5). With respect to the multiplicity of adenomas detected, Cha et al.²⁷ found a higher prevalence of multiple adenomas (44% vs 28%) in pre-diabetic vs non-diabetic patients with adenomas. In a study by Suh et al.,²⁸ DM appeared to be associated with multiple adenomas (OR 2.8, 95% CI 1.8–4.4), particularly in men of older age (>65 years).

Data regarding gender-specific adenoma and CRC relative risk in diabetic patients are scarce. In a study of 600 postmenopausal women, Elwing et al.⁷ found that diabetic women were more likely to have colorectal adenomas (OR 1.75, 95% CI 1.05–2.87) at screening colonoscopy than those without DM, in keeping with our present study. In contrast to our data, these authors found a trend toward a greater proportion of proximally located advanced adenomas in postmenopausal diabetic women. Difference in the overall proportions of advanced adenomas between Elwing’s study (6% of the non-diabetic women) and ours (11% of the non-diabetic women) may partly explain such discrepancies. Furthermore, difference in indications for colonoscopy (screening in Elwing’s study vs predominantly diagnostic colonoscopy in ours) could play a role. A recent cross-sectional study by Krämer et al.⁶ investigating the association of type 2 DM with colorectal neoplasia (CRN) in 1554 participants aged 50 to 74 years found an increased prevalence of CRN in diabetic women (PR 1.61, 95% CI 1.03–2.53), but not in men (PR 1.02, 95% CI 0.73–1.43). In our present study, we found an association of DM and the presence of ≥1 adenoma at colonoscopy both in men and women, albeit this was more pronounced in men.

The biological mechanisms underlying the increased risk for colorectal neoplasms in diabetic patients and the shift toward more proximally located lesions are unclear. It has been hypothesized that the “insulin resistance-hyperinsulinemia-IGF-axis” plays a major role. Previous nested, case-control studies confirmed that increased serum levels of IGF-I and insulin are associated with the risk of adenomas, advanced adenomas and CRC in men and women.^29,30

Some features of our study need to be acknowledged. Contributing to the strength of this study, data were derived from a large population-based cohort of patients undergoing elective colonoscopy at our institution. Before the outset of this population-based cohort, specialist registrars and consultant endoscopists were trained on the recognition, classification and endoscopic resection of colorectal neoplasms, in particular the subtle-appearing nonpolypoid neoplasms. A standardized endoscopy reporting system and comprehensive questionnaires were employed to capture clinical data and risk profiles. These data enabled us to carefully control for major confounders (e.g. BMI), albeit residual confounding due to non-measured variables cannot be excluded. As a limitation, the DM status in our study was ascertained based on self-reported questionnaire data, and, hence, the prevalence of DM could be underestimated (e.g. patients who were unaware that they had DM, such as those with glucose intolerance at the time of colonoscopy). Previous epidemiological studies pointed out, however, that self-reporting of DM is accurate.³¹ Data regarding the exact subtype of DM and the age at diagnosis of DM were not available. To preclude inclusion of patients with type 1 DM in our study, we excluded all patients aged <30 years. In our study, responders were older than non-responders and more likely to have colorectal polyps. A previous survey study from the Netherlands pointed out that non-response in cross-sectional population-based studies does not seem to cause bias in examined associations.³² Nevertheless, if the prevalence of colorectal polyps was overestimated because of the older age of the responders, this was the case both in diabetic and non-diabetic patients and, hence, is unlikely to have changed the main findings of this study.

Our current data reinforce the importance of stratifying for higher-risk subgroups, such as patients with DM, to increase the effectiveness of screening initiatives and colonoscopy surveillance. It remains to be shown whether such higher-risk patients would benefit from personalized cancer-prevention strategies in terms of screening test modality used, age of initiation and frequency of colonoscopy surveillance.

In conclusion, in this large population-based study, we found an increased prevalence of proximal and multiple adenomas in patients with DM, which was most pronounced in men. Close colonoscopic surveillance in patients with DM with particular attention to the proximal colon is warranted to optimize colonoscopic prevention of CRC in such a higher-risk subgroup.

Footnotes

Declaration of conflicting interests

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Larsson

Orsini

Wolk

. Diabetes mellitus and risk of colorectal cancer: A meta-analysis. J Natl Cancer Inst 2005; 97: 1679–1687.

International Agency for Research on Cancer, World Health Organization. GLOBOCAN 2012: Estimated cancer incidence, mortality and prevalence worldwide in 2012: Colorectal cancer, http://globocan.iarc.fr/factsheet.asp (accessed 27 March 2013).

Danaei

Finucane

. National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: Systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2·7 million participants. Lancet 2011; 378: 31–40.

American Institute for Cancer Research. Food, nutrition, physical activity, and the prevention of cancer: A global perspective, Washington, DC: American Institute for Cancer Research, 2007.

McKeown-Eyssen

. Epidemiology of colorectal cancer revisited: Are serum triglycerides and/or plasma glucose associated with risk? Cancer Epidemiol Biomarkers Prev 1994; 3: 687–695.

Krämer

Müller

Stegmaier

. Type 2 diabetes mellitus and gender-specific risk for colorectal neoplasia. Eur J Epidemiol 2012; 27: 341–347.

Elwing

Gao

Davidson

. Type 2 diabetes mellitus: The impact on colorectal adenoma risk in women. Am J Gastroenterol 2006; 101: 1866–1871.

Wong

Weiner

Hwang

. Insulin therapy and colorectal adenomas in patients with type 2 diabetes mellitus. Cancer Epidemiol Biomarkers Prev 2012; 21: 1833–1840.

Stein

Snyder

Barone

. Colorectal cancer outcomes, recurrence, and complications in persons with and without diabetes mellitus: A systematic review and meta-analysis. Dig Dis Sci 2010; 55: 1839–1851.

10.

Onitilo

Engel

Glurich

. Diabetes and cancer I: Risk, survival, and implications for screening. Cancer Causes Control 2012; 23: 967–981.

11.

DeCosse

Ngoi

Jacobson

. Gender and colorectal cancer. Eur J Cancer Prev 1993; 2: 105–115.

12.

Sanduleanu

Rondagh

Masclee

. Development of expertise in the detection and classification of non-polypoid colorectal neoplasia: Experience-based data at an academic GI unit. Gastrointest Endosc Clin N Am 2010; 20: 449–460.

13.

Rex

Petrini

Baron

. Quality indicators for colonoscopy. Gastrointest Endosc 2006; 63(4 Suppl): S16–S28.

14.

Kaminski

Regula

Kraszewska

. Quality indicators for colonoscopy and the risk of interval cancer. N Engl J Med 2010; 362: 1795–1803.

15.

Soetikno

Kaltenbach

Rouse

. Prevalence of nonpolypoid (flat and depressed) colorectal neoplasms in asymptomatic and symptomatic adults. JAMA 2008; 299: 1027–1035.

16.

Bosman FT, Carneiro F, Hruban RH, et al. WHO classification of tumours of the digestive system, http://www.cabdirect.org/abstracts/20113051318.html;jsessionid=F8F796937E0EB30E9DC8AA9EED286469 (2010, accessed 12 April 2012).

17.

Lieberman

Prindiville

Weiss

. Risk factors for advanced colonic neoplasia and hyperplastic polyps in asymptomatic individuals. JAMA 2003; 290: 2959–2967.

18.

Ferlitsch

Reinhart

Pramhas

. Sex-specific prevalence of adenomas, advanced adenomas, and colorectal cancer in individuals undergoing screening colonoscopy. JAMA 2011; 306: 1352–1358.

19.

Barros

Hirakata

. Alternatives for logistic regression in cross-sectional studies: An empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003; 3: 21–21.

20.

Martínez

Baron

Lieberman

. A pooled analysis of advanced colorectal neoplasia diagnoses after colonoscopic polypectomy. Gastroenterology 2009; 136: 832–841.

21.

van Heijningen

Lansdorp-Vogelaar

Kuipers

. Features of adenoma and colonoscopy associated with recurrent colorectal neoplasia based on a large community-based study. Gastroenterology 2013; 144: 1410–1418.

22.

Rondagh

Bouwens

Riedl

. Endoscopic appearance of proximal colorectal neoplasms and potential implications for colonoscopy in cancer prevention. Gastrointest Endosc 2012; 75: 1218–1225.

23.

le Clercq

Bouwens

Rondagh

. Postcolonoscopy colorectal cancers are preventable: A population-based study. Gut 2014; 63: 957–963.

24.

Limburg

Anderson

Johnson

. Diabetes mellitus and subsite-specific colorectal cancer risks in the Iowa Women’s Health Study. Cancer Epidemiol Biomarkers Prev 2005; 14: 133–137.

25.

Stram

Kolonel

. The association of diabetes with colorectal cancer risk: The Multiethnic Cohort. Br J Cancer 2010; 103: 120–126.

26.

Yoshida

Suzuki

Vallée

. Serum insulin levels and the prevalence of adenomatous and hyperplastic polyps in the proximal colon. Clin Gastroenterol Hepatol 2006; 4: 1225–1231.

27.

Cha

Lee

Joo

. Prediabetes is associated with a high-risk colorectal adenoma. Dig Dis Sci 2013; 58: 2061–2067.

28.

Suh

Kang

Kim

. Korean type 2 diabetes patients have multiple adenomatous polyps compared to non-diabetic controls. J Korean Med Sci 2011; 26: 1196–1200.

29.

Schoen

Weissfeld

Kuller

. Insulin-like growth factor-I and insulin are associated with the presence and advancement of adenomatous polyps. Gastroenterology 2005; 129: 464–475.

30.

Kaaks

Toniolo

Akhmedkhanov

. Serum C-peptide, insulin-like growth factor (IGF)-I, IGF-binding proteins, and colorectal cancer risk in women. J Natl Cancer Inst 2000; 92: 1592–1600.

31.

Jackson

DeFor

Crain

. Self-reported diabetes is a valid outcome in pragmatic clinical trials and observational studies. J Clin Epidemiol 2013; 66: 349–350.

32.

Van Loon

Tijhuis

Picavet

. Survey non-response in the Netherlands: Effects on prevalence estimates and associations. Ann Epidemiol 2003; 13: 105–110.