SEARCH

SEARCH BY CITATION

Keywords:

  • colorectal carcinoma;
  • insulin resistance;
  • NAFLD;
  • screening colonoscopy;
  • ultrasound examination

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Stadlmayr A, Aigner E, Steger B, Scharinger L, Lederer D, Mayr A, Strasser M, Brunner E, Heuberger A, Hohla F, Steinwendner J, Patsch W, Datz C (Oberndorf Hospital, Oberndorf; Paracelsus Medical University, Salzburg, Austria). Nonalcoholic fatty liver disease: an independent risk factor for colorectal neoplasia. J Intern Med 2011; 270: 41–49.

Abstract.

Background and aims.  Nonalcoholic fatty liver disease (NAFLD) is the hepatic manifestation of insulin resistance (IR), and IR is associated with an increased risk of colorectal carcinoma (CRC). Increased echogenicity suggesting NAFLD is a frequent incidental finding on ultrasound examination. We aimed to systematically evaluate whether NAFLD is an independent risk factor for colonic neoplasia.

Patients and methods.  One thousand two hundred and eleven patients (603 males, 60.6 ± 9.6 years; 608 females, 61.1 ± 10.3 years) who underwent screening colonoscopy according to national screening recommendations for CRC were evaluated in a cross-sectional study. Colorectal adenomas were classified as tubular adenoma, advanced adenoma (villous features, size ≥1 cm or high-grade dysplasia) or carcinoma. NAFLD was diagnosed by increased echogenicity on ultrasound examination after serological exclusion of infectious, immunological, hereditary or alcoholic aetiology.

Results.  Nonalcoholic fatty liver disease was diagnosed in 367 (60.8%) males and in 265 (43.5%) females. The total rate of adenomas was increased in subjects with NAFLD (243/367 vs. 107/236 in males, P = 0.010; 94/265 vs. 78/343 in females; P = 0.014). In particular, more tubular adenomas (127/367 vs. 56/236; P = 0.006), adenomas of the rectum (40/367 vs. 8/236; P = 0.004) and more cancers (6/367 vs. 1/236; P < 0.001) were observed in males with NAFLD. In females with NAFLD, more tubular adenomas (59/265 vs. 48/343; P = 0.011) and adenomas of the proximal colon (51/265 vs. 40/343; P = 0.041) were observed. Multivariate regression analyses demonstrated an independent association of colorectal adenomas with hepatic steatosis after adjustment for age, sex, body mass index and glucose intolerance (OR 1.47; 95% CI 1.079–2.003; P = 0.015).

Conclusion.  Patients with NAFLD undergoing screening colonoscopy reveal significantly more CRC precursor lesions and early CRC compared with subjects without NAFLD. This elevated risk is independent from other manifestations of IR. These findings suggest that detecting fatty liver on ultrasound should heighten the awareness for referral to screening colonoscopy.


Abbreviations:
AIH

autoimmune hepatitis

BMI

body mass index

BP

blood pressure

CRC

colorectal carcinoma

CRP

C, reactive protein

CVD

cardiovascular disease

GGT

gamma-glutamyl transpeptidase

HDL

high-density lipoprotein

HOMA-IR

homeostasis model assessment of insulin resistance

IGF

insulin-like growth factor

IR

insulin resistance

LDL

low-density lipoprotein

MetS

metabolic syndrome

NAFLD

nonalcoholic fatty liver disease

OGTT

oral glucose tolerance test

T2DM

type 2 diabetes mellitus

US

ultrasound examination

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Colorectal cancer (CRC) is the second most common cancer worldwide and a leading cause of cancer mortality [1]. Most CRCs originate from benign adenomatous precursor lesions that progress to malignancy over approximately 10 years because of the accumulation of mutations in oncogenes and tumour suppressor genes [2, 3]. Screening for CRC by colonoscopy effectively reduces CRC-associated mortality, as precursor lesions can be endoscopically or surgically removed at an early stage [4, 5]. Most American and European guidelines therefore recommend screening of asymptomatic adults above 50 years of age with adapted recommendations for specific risk groups [6, 7]. Improved knowledge regarding the risk stratification of the target population is desirable to improve screening strategies and recommendations.

An increased lifetime risk for CRC amongst other cancers has been demonstrated for patients with type 2 diabetes mellitus (T2DM) [8–10]. Elevated serum concentrations of insulin and insulin-like growth factors (IGF) 1 and 2 in patients with insulin resistance (IR) seem to exert growth-promoting effects on CRC precursor lesions, increasing the likelihood of the development of malignancies [11, 12]. These associations have resulted in recommendations for earlier and more frequent colonoscopy screening in patients with T2DM, particularly those receiving insulin therapy [13].

Nonalcoholic fatty liver disease (NAFLD) is the hepatic manifestation of IR and the most common chronic liver disease worldwide. The histological and clinical spectrum of NAFLD ranges from benign steatosis to nonalcoholic steatohepatitis with marked inflammation and scarring, which may progress to cirrhosis, end-stage liver disease and hepatocellular carcinoma [14, 15]. The rising prevalence of NAFLD parallels the increase in obesity-related conditions such as hypertension, atherogenic dyslipidaemia and T2DM, which have been collectively termed the metabolic syndrome (MetS) [16, 17]. The MetS and its components are characterized by an increased risk of cardiovascular disease (CVD) and cancers, including CRC. With regard to cardiovascular disease, it has recently been demonstrated that NAFLD is not only a marker of IR, but also increases cardiovascular disease risk beyond established risk factors [18].

As IR represents a common risk factor for both CRC and NAFLD, and increased echogenicity of the liver on ultrasound (US) examination is a frequent finding in the routine clinical setting, we aimed to systematically evaluate whether NAFLD is an independent risk factor for colonic neoplasia.

Subjects and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Subjects

The study cohort consisted of 1382 consecutive Caucasians (702 males, 40–76 years; 680 females, 31–88 years) who underwent colonoscopy for CRC screening according to national screening recommendations at a single centre between June 2007 and December 2009. The study was approved by the local ethics committee, and informed consent was obtained from all participants.

Study concept and execution

Study participants were examined on two consecutive days, and venous blood was collected and analysed following an overnight fast. In series, a medical history was taken, and dietary practices were assessed using a validated food-frequency questionnaire. A physical examination and right upper quadrant US examination were performed. On the second day, all subjects underwent complete colonoscopy.

Laboratory assessment

Full blood counts were obtained in all subjects by standard laboratory methods. Erythrocyte sedimentation rate was measured in citrated plasma. Blood was centrifuged, and plasma was analysed for aspartate aminotranspeptidase and alanine aminotranspeptidase activities and levels of triglycerides, total cholesterol, high-density and low-density lipoprotein cholesterol, glucose, insulin and C-reactive protein (CRP). Glycated haemoglobin A1c was measured by HPLC using an Adamts H-8160 analyzer (Menarini, Florence, Italy). The homeostasis model assessment-insulin resistance (HOMA-IR; fasting insulin [μU L−1] × fasting glucose [mmol dL−1]/22.5) was used to assess IR.

US examination

Right upper quadrant US examination was performed using an ATL HDI 5000 machine (Phillips Medical Systems, Vienna, Austria). The examinations were carried out by one of the five physicians with 5–21 years of experience. The liver was considered ‘normal’ if the echogenicity was homogenous and similar to or slightly higher than the echogenicity of the renal parenchyma. The liver was considered ‘fatty liver’ if areas of significant increased echogenicity in relation to the renal parenchyma were found. The severity of sonographic steatosis was not graded [19]. The diagnosis of NAFLD was based on specific findings on right upper quadrant US examination (as noted above) and after exclusion of viral, autoimmune and hereditary (Wilson’s disease, hereditary haemochromatosis, alpha-1 antitrypsin deficiency) liver disease and excess alcohol consumption (determined with a questionnaire; cut-off >20 g day−1, a limit considered to be below the traditional level for alcohol-induced liver disease [20]).

A total of 632 patients (45.7%) had evidence of NAFLD on US examination. In patients with elevated liver enzymes and/or a positive US result suggestive of hepatic steatosis, serum was analysed for the evidence of unrecognized hepatitis C virus infection, hepatitis B virus infection, autoimmune hepatitis (AIH), Wilson’s disease or alpha-1 antitrypsin deficiency. Subjects with increased serum ferritin or transferrin saturation underwent HFE genotyping to exclude hereditary haemochromatosis. Ninety-nine (14.1%) male subjects were excluded because of incomplete colonoscopies (n = 10), a history of recent colorectal polypectomy (n = 1), asymptomatic ulcerative colitis (n = 3), other extraintestinal malignancies (n = 3), viral hepatitis (n = 10), AIH (n = 1), homozygous hereditary haemochromatosis (n = 3), a history of excess alcohol consumption (>20 g day−1, n = 34) or refusal to participate in the study (n = 34). Seventy-two (10.6%) women were excluded because of incomplete colonoscopies (n = 18), a history of recent colorectal polypectomy (n = 2), asymptomatic ulcerative colitis (n = 1), asymptomatic Crohn’s disease (n = 2), other extraintestinal malignancies (n = 2), viral hepatitis (n = 17), AIH (n = 1), a history of excess alcohol consumption (>20 g day−1, n = 10) or refusal to participate (n = 19). Thus, data from 603 males (85.9%) and 608 females (89.4%) were included in the final analysis.

Metabolic syndrome was evaluated as defined by the National Cholesterol Education Program Adult Treatment Panel [21]. Blood pressure was determined twice by a nurse after a 5-min rest in a sitting position, and the average was taken as the measurement of blood pressure. Waist circumference was measured at the highest point of the iliac crest with subjects standing in an upright position. The MetS was diagnosed when three or more of the following criteria were met: fasting blood glucose level ≥6.1 mmol L−1, waist circumference >102 cm in males and >88 cm in females, blood pressure ≥130/85 mmHg or current antihypertensive treatment, plasma triglycerides ≥1.7 mmol L−1 and plasma HDL <1.0 mmol L−1 in males and <1.3 mmol L−1 in females. Body mass index (BMI) was calculated as weight/height squared (kg m−2).

Colonoscopy

The laxative Klean–Prep® (containing macrogol 59.0 g, sodium sulphate 5.68 g, sodium bicarbonate 1.68 g, NaCl 1.46 g and potassium chloride 0.74 g; Norgine, Marburg, Germany) was used for bowel preparation before colonoscopy. Colonoscopic findings were classified as tubular adenoma, advanced adenoma, i.e. villous or tubulovillous features, size ≥1 cm or high-grade dysplasia [22, 23] or carcinoma after a combined analysis of macroscopic and histological results. Lesions were classified by location (i.e. proximal colon including caecum, ascending colon and transverse colon, distal colon ranging from the splenic flexure to the sigmoid and rectum alone) [24].

Statistical analysis

SigmaStat 3.1 or STATA 0.8 software packages (Erkrath, Germany) were used for all analyses. Data are presented as mean ± SD, unless otherwise indicated. Student’s t-test was used for comparison of continuous variables or Mann–Whitney U-test in the case of non-Gaussian distribution of variables. The chi-square or McNemar’s test was used to calculate rates and proportions as appropriate. Multivariate regression analysis was used to determine variables independently associated with the number of polyps. Logistic regression analysis was used to determine variables predicting the presence of adenoma. As this was a cross-sectional study, the dependent variable in the logistic regression equation is the log odds of having adenoma at the time of examination. Throughout, a two-tailed P-value <0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Characteristics of the study cohort

A total of 1382 subjects were enrolled after exclusion of 99 males and 72 females. Data from 603 males and 608 females were included in the final analysis (Fig. 1). NAFLD was diagnosed in 367 of 603 (60.9%) males and 265 of 608 (43.6%) females. Seventy-eight males (21.3%) and 90 females with fatty liver (34.0%) also had elevated liver transaminase activities. Clinical and biochemical features of participants with and without NAFLD are shown in Table 1.

image

Figure 1. Study participants.

Download figure to PowerPoint

Table 1. Clinical and biochemical characteristics of the study cohort grouped as subjects with and without NAFLD
 Males (n = 603)Females (n = 608)
NAFLDNo NAFLD P-valueNAFLDNo NAFLD P-value
  1. Unless indicated, data are given as mean ± SD. P-values indicate the difference between groups as calculated by chi-square test, t-test or Mann–Whitney U-test.

  2. NAFLD, nonalcoholic fatty liver; NS, not significant; BMI, body mass index; BP, blood pressure; LDL, low-density lipoprotein; HDL, high-density lipoprotein; OGTT, oral glucose tolerance test; HOMA-IR, homeostasis model assessment-insulin resistance; HbA1c, glycated haemoglobin A1c; GGT, gamma-glutamyl transpeptidase; AST, aspartate transaminase; ALT, alanine transaminase; CRP, C-reactive protein.

  3. aSubjects with T2DM (n = 83) were excluded.

Number of subjects367236 265 343  
Age (years)62.1 ± 9.361.4 ± 10.3NS62.61 ± 9.3359.91 ± 10.89<0.001
BMI (kg m−2)28.5 ± 5.225.1 ± 3.5<0.00129.48 ± 5.9324.80 ± 3.77<0.001
Waist (cm)105.5 ± 12.795.5 ± 8.8<0.001101.02 ± 13.1987.16 ± 12.12<0.001
Waist/hip ratio1.00 ± 0.060.95 ± 0.06<0.0010.92 ± 0.070.85 ± 0.11<0.001
Systolic BP (mmHg)136.5 ± 17.8130.1 ± 14.2<0.001135.80 ± 21.11129.10 ± 19.49<0.001
Diastolic BP (mmHg)83.2 ± 11.278.6 ± 9.0<0.00181.95 ± 11.8579.07 ± 11.430.003
Total cholesterol (mmol L−1)5.4 ± 1.15.4 ± 1.0NS5.7 ± 1.205.6 ± 1.00NS
LDL cholesterol (mmol L−1)3.5 ± 1.03.5 ± 0.9NS3.7 ± 1.103.5 ± 0.900.006
HDL cholesterol (mmol L−1)1.4 ± 0.41.6 ± 0.4<0.0011.6 ± 0.401.9 ± 0.50<0.001
Triglycerides (mmol L−1)1.8 ± 1.11.2 ± 0.7<0.0013.6 ± 1.902.5 ± 1.20<0.001
Uric acid (mmol L−1)0.35 ± 0.070.33 ± 0.07<0.0010.30 ± 0.070.25 ± 0.06<0.001
Fasting glucose (mmol L−1)a6.1 ± 1.45.6 ± 1.1<0.0016.16 ± 1.865.38 ± 0.76<0.001
Fasting insulin (μU mL−1)a13.8 ± 12.88.6 ± 9.80.00314.07 ± 13.898.06 ± 9.80<0.001
OGTT 1 h (mmol L−1)a10.7 ± 2.39.5 ± 2.0<0.00110.51 ± 2.739.30 ± 2.37<0.001
OGTT 2 h (mmol L−1)a7.8 ± 2.57.1 ± 2.30.0128.31 ± 2.497.36 ± 1.89<0.001
HOMA-IRa4.1 ± 4.02.2 ± 2.7<0.0013.5 ± 2.742.00 ± 2.73<0.001
HbA1c (%)a5.7 ± 0.625.5 ± 0.570.0306.04 ± 0.925.62 ± 0.36<0.001
GGT (U L−1)59.9 ± 86.442.6 ± 56.30.01143.34 ± 69.3627.74 ± 46.040.001
AST (U L−1)27.8 ± 20.023.6 ± 12.40.00624.78 ± 14.0021.74 ± 11.320.003
ALT (U L−1)33.0 ± 26.822.9 ± 12.5<0.00127.26 ± 22.0819.94 ± 18.05<0.001
Erythrocyte sedimentation rate7.5 ± 8.96.2 ± 7.5NS12.97 ± 17.078.73 ± 9.57<0.001
CRP (mg dL−1)0.47 ± 0.930.33 ± 0.72NS0.63 ± 1.720.37 ± 1.020.022
Haemoglobin (g dL−1)15.1 ± 1.314.7 ± 1.50.00614.12 ± 2.3613.79 ± 1.150.025

In total, 341 of 1211 enrolled study participants (28.1%) were found to have at least one colorectal lesion. The prevalence of colorectal lesions was 34% in the NAFLD group and 21.7% in the control group (P < 0.001). Furthermore, 40% of male patients with NAFLD had colorectal lesions compared with 28.3% without NAFLD (P = 0.010), whereas 25.6% of female patients with NAFLD had colorectal lesions compared with 17.2% without NAFLD (P = 0.014). Of note, more tubular adenomas and colorectal carcinomas were found amongst patients with NAFLD, whereas the number of advanced adenomas was not different between groups in either males or females. With regard to the location of adenomas, in males with NAFLD, a significantly higher number of lesions were detected in the rectum, whereas in females with NAFLD, significantly more lesions were detected in the proximal colon. The details of these analyses are summarized in Table 2.

Table 2. Colonoscopic findings in males and females with and without NAFLD
 MalesFemales
With NAFLD (n = 367)Without NAFLD (n = 236) P-valueWith NAFLD (n = 265)Without NAFLD (n = 343) P-value
  1. In section A, each patient was only counted once with the highest graded lesion; i.e. patients with advanced adenomas and tubular adenomas were allocated to the advanced adenoma group.

Total number of patients with lesions147 (40.1%) 67 (28.4%)0.00568 (25.7%)59 (17.2%)0.015
Total number of lesions2431070.01094780.014
(A) Histology of removed lesions (number of patients)
 Tubular adenoma (≥1)120530.00959480.011
 Advanced adenoma (≥1)21130.9449110.921
 Colorectal carcinoma (≥1)61<0.00101NS
(B) Histology of removed lesions (number of lesions)
 Tubular adenoma213900.01184660.014
 Advanced adenoma24160.92310110.578
 Colorectal carcinoma61<0.00101NS
(C) Location (total number of lesions)
 Proximal colon109580.35351400.041
 Distal colon94410.07935280.121
 Rectum4080.0048100.941
(D) Number of colorectal lesions
 1 lesion85450.27548470.170
 2 lesions35120.0671470.052
 ≥3 lesions27100.166650.665

As the total number of adenomas was found to be increased in patients with NAFLD, and NAFLD is associated with the MetS, we evaluated the prevalence of the various components of the MetS in patients with and without adenomas. We found that apart from the higher rate of NAFLD in patients with adenomas, impaired glucose tolerance and higher levels of fasting glucose, fasting insulin and hypertension were more common in males with adenomas but the rate of the MetS was not different between males with and without adenomas.

In females, adenomas were associated with the MetS, impaired fasting glycaemia and higher levels of fasting insulin. Additionally, HDL was significantly more often decreased in patients with adenomas than in those without. No differences in the prevalence of other components of the MetS were found. Amongst males with colorectal lesions, 27 (12.6%) were receiving antidiabetic medication compared with 20 (5.1%) of those without colorectal lesions (P = 0.002). No statistical difference was found in females.

Besides a higher prevalence of fatty liver, patients with adenomas presented significantly more often with some components of the MetS (including elevated levels of fasting glucose and plasma triglycerides, and decreased plasma HDL in females and hypertension in males).

Glucose intolerance, antidiabetic medication, impaired fasting glycaemia and impaired glucose tolerance were more often found in the adenoma group. Significantly fewer adenomas were observed in glucose-tolerant patients. Impaired fasting glycaemia was defined as fasting glucose ≥6.1 and <7.0 mmol L−1. Impaired glucose tolerance was defined as an oral glucose tolerance test (OGTT) at 2 h ≥7.8 and <11.1 mmol L−1. Glucose-tolerant patients were those with fasting glucose <6.1 mmol L−1, no diabetic medication and an OGTT at 2 h <7.8 mmol L−1; there were significantly fewer adenomas in this group of patients. The prevalence of smoking was not different in patients with adenoma and without adenoma (data not shown). CRP level was significantly higher in the adenoma group. These results are summarized in Tables 3, S1 and S2.

Table 3. Components of the metabolic syndrome (MetS), CRP and glucose tolerance/intolerance in all patients with and without adenomas
  All patients (n = 1211)With adenoma (n = 341)Without adenoma (n = 870) P-value
  1. NAFLD, nonalcoholic fatty liver; MetS, metabolic syndrome; HOMA-IR, homeostasis model assessment-insulin resistance; HbA1c, glycated haemoglobin A1c; OGTT, oral glucose tolerance test; CRP, C-reactive protein.

Age60.88 ± 9.9863.96 ± 9.5659.96 ± 9.88<0.001
NAFLD632215417<0.001
MetS (3/5 criteria)3201072130.024
1. Fasting glucose (≥6.1 mmol L−1)310114196<0.001
2a. Waist circumference (>102 cm) in males194/430 741200.365
2b. Waist circumference (>88 cm) in females258/402 552030.243
3. Plasma triglycerides (≥1.7 mmol L−1)3091032060.023
4a. Plasma HDL (<1.0 mmol L−1) in males 66 23 430.983
4b. Plasma HDL (<1.3 mmol L−1) in females 97 28 690.049
5. Hypertension7812435380.003
Fasting insulin (μU mL−1)10.92 ± 11.6113.18 ± 16.0110.05 ± 9.240.006
HOMA-IR 2.92 ± 3.26 3.60 ± 4.42 2.66 ± 2.650.003
Glucose-intolerant patients517185332<0.001
Glucose-tolerant patients694156538<0.001
CRP4.01 ± 7.474.06 ± 7.053.99 ± 7.900.047

Age, sex, MetS and its components, hepatic steatosis, glucose intolerance and CRP were associated with the presence and number of colonic polyps (P < 0.05). However, in logistic regression models with MetS or its individual components, no association with colorectal adenomas independent of age and sex was found for MetS, its individual components or CRP, whereas hepatic steatosis, glucose intolerance and BMI remained associated with adenoma after adjustment for age and sex. No interaction between sex and steatosis (P = 0.989) or glucose intolerance (P = 0.796) with respect to adenoma prediction was noted. Likewise, in multiple regression analyses, the number of polyps was not associated with the components of the MetS or CRP, but the association with steatosis (P = 0.037) was maintained, whereas associations with BMI (P = 0.073) and glucose intolerance (P = 0.064) were borderline significant. These results are summarized in Tables 4 and 5. Patients who received treatment for T2DM at the time of the examination were excluded from the analysis as an increased risk for colorectal adenomas has been demonstrated for patients with T2DM, and antidiabetic medication and particularly insulin may influence the development of colorectal adenomas.

Table 4. Logistic regression model for variables predicting the presence of adenoma
ParameterOdds ratio95% CI P-value
  1. 02*log(likelihood): for this model = 1337.29; χ2 = 101.79; P = 0.000.

  2. CI, confidence interval.

Sex (male = 1/female = 2)0.5450.411–0.721<0.001
Age (per year increment)1.0441.028–1.059<0.001
Body mass index (kg m−2)1.0210.987–1.0550.236
Liver steatosis (no = 0/yes = 1)1.4701.079–2.0030.015
Glucose intolerance (no = 0/yes = 1)1.3831.037–1.8460.027
Table 5. Multiple regression analysis
ParameterβSE P-value
  1. The absolute number of adenomas was used as the dependent variable.

  2. β, regression coefficient; SE, standard error.

Sex (male = 1/female = 2)−0.1570.029<0.001
Age (per year increment)0.1600.029<0.001
Body mass index (kg m−2)0.0580.0330.073
Liver steatosis (no = 0/yes = 1)0.0680.0330.037
Glucose intolerance (no = 0/yes = 1)0.0560.0300.064

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Obesity-related sequelae represent a growing health concern in Western societies. Being overweight or obese is causally associated with an increased incidence of T2DM, cardiovascular disease and several forms of malignancy. IR has been established as the common mechanism behind these different disease entities. Here, we report an increased rate of colorectal neoplasia in patients with hepatic steatosis on US examination, a surrogate marker for NAFLD.

Over the past decade, NAFLD has been firmly established as the liver manifestation of the IR syndrome, also termed the MetS. NAFLD is the most frequent cause of elevated liver enzymes in the general population [20]. The high prevalence of NAFLD in the general population is consistent with the results of our study showing that 60.9% of males and 43.5% of females had NAFLD. In our study, the diagnosis of NAFLD was based on the results of liver US examination. In general, liver biopsy is regarded as the gold standard for obtaining a diagnosis of NAFLD. However, it is important to consider that inter- and intraobserver variations also limit the diagnostic accuracy of liver histology in NAFLD. A 96.2% concordance rate between US and liver biopsy was recently reported [25]. It is likely that such a high rate may not be achieved in daily clinical practice; nonetheless, this finding supports the choice of US as a useful diagnostic tool for NAFLD in the primary care setting, in which performance of liver biopsy is not feasible in all at-risk individuals.

We evaluated colonoscopy findings in male and female Caucasian patients with regard to the presence or absence of NAFLD and found that colorectal lesions, particularly tubular adenomas and carcinomas, were significantly more prevalent in patients with NAFLD, irrespective of age, sex and other manifestations of IR. In line with previous reports, we found a lower overall rate of colorectal adenomas in female compared with male subjects [26, 27]. Of particular note, NAFLD increased the likelihood of CRCs and lesions of the rectum in males and lesions of the proximal colon in females. Similar differences in the location of colorectal adenomas have recently been reported from a study in an Asian population [28]. However, to what extent sex hormones, sex-specific lifestyle and dietary habits as well as genetic and epigenetic factors determine the location in which colorectal polyps and consequently cancers originate remains to be determined. In contrast to previously reported data, we did not observe an association between smoking habits and colorectal adenomas. Currently, we have not been able to identify an explanation for this lack of association in our cohort (data not shown).

Both the MetS and IR have previously been associated with colorectal neoplasia in epidemiological studies [29, 30]. Our results suggest that the presence of liver fat may serve as a particularly important clinical manifestation of IR predicting the presence of colorectal adenomas beyond the classical components of the MetS or IR. Support for the notion that steatosis is a powerful risk factor for diseases related to the MetS comes from studies that identified NAFLD as a stronger predictor than the MetS of first-time cardiovascular events [31]. With regard to the increased risk of cardiovascular disease in patients with NAFLD, it is generally assumed that, along with accumulated lipids, the hepatic tissue becomes an additional source of factors that further augment the proinflammatory systemic milieu [32]. Similarly, dysfunctional adipose tissue in obese individuals is associated with an increased cancer rate because of its propensity to secrete several cancerogenic and growth-promoting factors such as interleukin (IL)-6 or tumour necrosis factor (TNF)-α and a decreased expression of protective factors such as adiponectin. Our results suggest that an analogous relationship may exist between hepatic steatosis and the development of neoplastic precursor lesions and probably also colorectal cancer, as the steatotic and inflamed liver may secrete growth-promoting factors into the systemic circulation [33]. The exact nature of these soluble mediators derived from the steatotic liver which may be involved in tumour development and/or growth is unclear. Of note, adiponectin is a key mediator of adipose tissue effects in obesity, but it is minimally expressed in the liver. Similarly, IL-6 is normally secreted from Th1 lymphocytes and also from inflamed adipose tissue but the amount derived from the liver is minimal. Hence, factors such as TNF-α and IGF-1, as well as plasminogen activator inhibitor 1 (PAI-1), vascular endothelial growth factor (VEGF) and angiopoietin, will need to be investigated in future studies with regard to their expression in NAFLD compared with normal livers. By activating the PI3K/AKT pathway, IGF-1 is a well-known mediator of cell proliferation. Likewise, changes in TNF-α/NF-κB signalling, which is crucial for cell survival and apoptosis, may be linked to the increased rate of polyps in patients with NAFLD. One may even speculate that NAFLD-derived PAI-1, VEGF or angiopoietin may be involved in metastasis and neoangiogenesis of malignancies and thus cancer progression [32, 34].

Because of an increased incidence of colorectal neoplasia, screening guidelines from different gastroenterological societies have already suggested earlier and more frequent colorectal cancer screening for patients with T2DM, especially those receiving insulin therapy [6, 7]. In line with this, we also observed an independent association between increased risk of colorectal lesions and glucose intolerance in addition to the associations with liver steatosis. In routine clinical practice, the diagnosis of hepatic steatosis is frequently obtained from US examination commonly performed for other indications. Hence, our results imply that patients who receive the diagnosis of liver steatosis during US examination may particularly benefit from referral for screening colonoscopy.

In summary, we report an increased prevalence of colorectal adenomas in patients with NAFLD compared with similar patients without NAFLD. Hence, hepatic lipid accumulation may either be an indicator or, analogous to the increased cardiovascular risk in patients with NAFLD, an additional source of a systemic neoplasia-promoting milieu. These findings should increase the awareness of the necessity to perform CRC screening in patients with hepatic steatosis on US examination.

Conflict of interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

None of the authors has any potential financial conflicts of interest to disclose with regard to this investigation.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

We are grateful to Karin Schwenoha and Carmen Winkler, Oberndorf Hospital, for laboratory technical support.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and methods
  5. Results
  6. Discussion
  7. Conflict of interest
  8. Acknowledgements
  9. Funding
  10. References
  11. Supporting Information

Table S1. Components of the metabolic syndrome (MetS) and glucose tolerance/intolerance in males with and without adenomas.

Table S2. Components of the metabolic syndrome (MetS) and glucose tolerance/intolerance in females with and without adenomas.

FilenameFormatSizeDescription
JOIM_2377_sm_supp.doc66KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.