Objective Non-alcoholic fatty liver disease (NAFLD) affects 20–40% of the general adult population. Due to shared risk factors, it is postulated that NAFLD patients have an increased risk of colorectal neoplasm and should be a target group for screening. The aim of this study was to examine the prevalence of colorectal neoplasm in NAFLD patients and the risk of colorectal neoplasm in relation to the severity of NAFLD histology.
Design Cross-sectional study.
Setting University hospital with case recruitment from the community and clinics.
Patients Subjects aged 40–70 years were recruited for colonoscopic screening from two study cohorts: (1) community subjects; and (2) consecutive patients with biopsy proven NAFLD. In the community cohort, hepatic fat was measured by proton-magnetic resonance spectroscopy.
Main outcome measures Prevalence of colorectal adenomas. Advanced colorectal neoplasm was defined as cancer or adenomas with villous architecture or high grade dysplasia.
Results NAFLD patients (N=199) had a higher prevalence of colorectal adenomas (34.7% vs 21.5%; p=0.043) and advanced neoplasms (18.6% vs 5.5%; p=0.002) than healthy controls (N=181). Thirteen of 29 (45%) NAFLD patients with advanced neoplasms had isolated lesions in the right sided colon. Among patients with biopsy proven NAFLD, patients with non-alcoholic steatohepatitis (N=49) had a higher prevalence of adenomas (51.0% vs 25.6%; p=0.005) and advanced neoplasms (34.7% vs 14.0%; p=0.011) than those with simple steatosis (N=86). After adjusting for demographic and metabolic factors, non-alcoholic steatohepatitis remained associated with adenomas (adjusted OR 4.89, 95% CI 2.04 to 11.70) and advanced neoplasms (OR 5.34, 95% CI 1.92 to 14.84). In contrast, the prevalence of adenomas and advanced neoplasms was similar between patients with simple steatosis and control subjects.
Conclusions Non-alcoholic steatohepatitis is associated with a high prevalence of colorectal adenomas and advanced neoplasms. The adenomas are found more commonly in the right sided colon. Colorectal cancer screening is strongly indicated in this high risk group.
- Fatty liver
- colonic polyps
- colorectal cancer
- colonic adenomas
- colonic neoplasms
- diabetes mellitus
- fatty liver
Statistics from Altmetric.com
- Fatty liver
- colonic polyps
- colorectal cancer
- colonic adenomas
- colonic neoplasms
- diabetes mellitus
- fatty liver
Significance of this study
What is already known about this subject?
Non-alcoholic fatty liver disease (NAFLD) is one of the most common chronic liver diseases, and is associated with metabolic syndrome and systemic inflammation.
Colorectal cancer is the second leading cause of cancer death worldwide, and is associated with obesity.
Screening can reduce the incidence and mortality of colorectal cancer. High risk groups are incompletely defined.
What are the new findings?
Patients with non-alcoholic steatohepatitis (NASH), the active form of NAFLD, are at highest risk of having colorectal adenomas and advanced neoplasms.
The association between NASH and colorectal neoplasm is independent of other demographic and metabolic factors.
Patients with NAFLD and NASH commonly have colorectal adenomas and advanced neoplasms in the right sided colon.
How might it impact on clinical practice in the foreseeable future?
Colorectal cancer screening is strongly indicated in NASH patients. Since proximal colonic lesions are common in these patients, colonoscopy should be considered as the preferred screening method.
Colorectal cancer is the second leading cause of cancer deaths worldwide.1 2 Most colorectal cancers develop through the adenoma–carcinoma sequence and have a long premalignant stage. Therefore, colorectal cancer screening can detect not only early cancers for curative treatment but also adenomas before they become malignant. In observational studies and randomised controlled trials, colorectal cancer screening such as faecal occult blood tests and flexible sigmoidoscopy has been shown to reduce the incidence of colorectal cancer and cancer related mortality.3–5 As a result, current guidelines support regular colorectal cancer screening in adults aged 50–70 years.6–8
However, only a minority of the population at risk undergoes screening because of perceived health, psychological, and access barriers.9 Current guidelines also do not fully address various risk factors such as gender, smoking, and race.10 Better characterisation of factors associated with advanced colorectal neoplasm may increase the screening uptake rate in higher risk individuals.
In epidemiological studies, diabetes and obesity have been associated with the development of colorectal cancer.11 12 Non-alcoholic fatty liver disease (NAFLD) is a condition closely related to diabetes, obesity, and metabolic syndrome. It has emerged as one of the most common chronic liver diseases worldwide, affecting up to 20–40% of the general adult population, and may progress to non-alcoholic steatohepatitis (NASH), cirrhosis, and liver cancer.13–15 Since NAFLD and colorectal neoplasm share many common risk factors, we postulated that NAFLD patients have increased risk of colorectal neoplasm and should be a target group for screening.
In this study, we prospectively recruited community subjects and hospital patients for screening colonoscopy. We aimed to study the prevalence of colorectal neoplasm in NAFLD patients. Our secondary aim was to evaluate the risk of colorectal neoplasm in relation to the severity of NAFLD histology.
Two cohorts were recruited from January 2008 to July 2010. The first cohort was recruited from the community. We used the census database of the Hong Kong SAR government to generate a random list of local residents, and invited them by mail and phone. Respondents underwent blood tests and proton-magnetic resonance spectroscopy (1H-MRS). Based on 1H-MRS results, subjects without NAFLD served as healthy controls and those with NAFLD constituted the community NAFLD cohort. The second cohort represented consecutive patients with biopsy proven NAFLD from the hepatology clinics of the Prince of Wales Hospital and Tseung Kwan O Hospital, Hong Kong. Subjects in both cohorts were aged 40–70 years. Subjects with other liver diseases were excluded by history taking and checking hepatitis B surface antigen, anti-hepatitis C virus antibody, and anti-nuclear antibody. Men who consumed >30 g of alcohol per day and women who consumed >20 g per day were excluded. We also excluded subjects with history of colorectal cancers or polyps, history of inflammatory bowel disease, bowel symptoms including per rectal bleeding and altered bowel habit, prior colorectal cancer screening, and contraindications to colonoscopy. All patients provided informed written consent. The study protocol was approved by the Clinical Research Ethics Committee of The Chinese University of Hong Kong.
Detailed clinical assessment was conducted within 4 weeks before liver biopsy or 1H-MRS. Bowel symptoms, drug history, alcohol intake, past medical history, and family history were recorded using a standardised questionnaire. Anthropometric measurements including body weight, body height, and waist circumference were made. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared (kg/m2). Waist circumference was measured at a level midway between the lower rib margin and iliac crest with the tape all around the body in the horizontal position. Blood tests including liver biochemistry, glucose, and lipids were performed after fasting for 8 h.
Colonoscopy was performed within 6 months of clinical assessment under conscious sedation using intravenous diazepam and pethidine by endoscopists with the experience of performing more than 500 procedures (VWSW, SWCT, TF, HLYC). Subjects were given polyethylene glycol for bowel preparation together with instructions for use. A complete examination was defined as an endoscope reaching the caecum as documented by a picture of the ileocaecal valve. All colorectal lesions found were documented for their site and size (measured by biopsy forceps that opened up to 7 mm). The site of colorectal neoplasm was defined as right sided if it was proximal to the splenic flexure. The withdrawal time of the colonoscopy procedure was at least 6 min to minimise the chance of missing lesions. Incomplete examination was excluded from analysis.
The quality of bowel preparation was graded as good (no or small volume of clear liquid, with >95% of surface seen), fair (in between good and poor), and poor (presence of semi-solid stool that could not be suctioned or washed away, and <90% of surface seen). Advanced colorectal neoplasm was defined as presence of cancer or adenomas with high grade dysplasia or villous architecture.
Proton-magnetic resonance spectroscopy
In the first cohort (community cohort), all subjects underwent hepatic triglyceride content measurement by in vivo 1H-MRS. Whole body 3.0 T scanner with a single voxel point resolved spectroscopy sequence and an echo time of 40 ms and repetition time of 5000 ms was used. A survey scan was first performed in the abdominal region to help positioning a volume measuring 20 (AP) × 15 (RL) × 40 (FH) mm within the liver. The scanner's built-in body coil was used for both signal transmission and reception. No-water-suppressed spectra was acquired using 32 signal averages and the data were exported for offline spectral analysis. Water (4.65 ppm) and lipid (1.3 ppm) peak amplitudes were measured to determine vertebral marrow fat content, which was defined as the relative fat signal amplitude in terms of a percentage of the total signal amplitude (water and fat) and calculated according to the following equation: fat content = (Ifat/(Ifat+Iwater)) × 100, where Ifat and Iwater are the peak amplitudes of fat and water, respectively. Correction for relaxation loss was not applied because of the relatively long repetition time and short echo time. A hepatic triglyceride content of 5% was used to distinguish between patients with and without fatty liver.16 17
For patients in the second cohort (hospital cohort), percutaneous liver biopsy was performed using a 16 gauge Temno needle (Cardinal Health, McGaw Park, Illinois, USA). All histological slides were read by two pathologists (AWHC and PCLC) who were blinded to the clinical data. For specimens with discrepancies in histological scoring, a consensus was reached after discussion between the two pathologists.14 NASH was diagnosed for specimens with obvious hepatocyte ballooning and intra-acinar and portal chronic inflammation.18 For analysis, NAFLD patients not fulfilling the diagnosis of NASH were labelled as having simple steatosis. Histological grading and staging of NAFLD were also scored according to the system reported by Kleiner and colleagues.16 Fibrosis was staged from 0 to 4, with stage 0=no fibrosis, 1=perisinusoidal or periportal fibrosis, 2=perisinusoidal and portal/periportal fibrosis, 3=bridging fibrosis, and 4=cirrhosis.
Statistical tests were performed using the Statistical Package for Social Sciences version 16.0. Continuous variables were expressed as mean (SD) or median (range) as appropriate. Categorical variables were compared using χ2 test or Fisher exact test as appropriate. Binary logistic regression analysis was performed to assess the interaction between NAFLD and other risk factors of colorectal neoplasm including age, gender, smoking, and family history. A two-sided p value <0.05 was taken as statistically significant.
According to our previous population screening, colorectal adenomas and advanced neoplasms were present in 24% and 13% of the study subjects, respectively.17 The prevalence of colorectal adenomas in other high risk groups (eg, obesity) is 60–90% higher than that of the general population.19 Assuming the prevalence of colorectal adenomas was 35% in NAFLD subjects and 20% in healthy controls, the recruitment of 111 NAFLD subjects and 222 controls would achieve 80% power to detect the difference at a 5% significance level.
Four hundred and forty-three subjects were assessed for eligibility. One hundred and thirty-five subjects with biopsy proven NAFLD, 64 subjects with NAFLD detected by 1H-MRS, and 181 healthy control subjects were included in the final analysis (figure 1).
The control group included slightly younger subjects and more women (table 1). NAFLD subjects in both the community cohort and hospital cohort had higher BMI, waist circumference, fasting glucose, low density lipoprotein (LDL) cholesterol, triglycerides, alanine aminotransferase, and aspartate aminotransferase values. They also had lower high density lipoprotein (HDL) cholesterol values, and were more likely to have diabetes and hypertension. In the hospital cohort, 49 (36%) patients had NASH, and 20 (15%) had bridging fibrosis or cirrhosis.
Prevalence of colorectal neoplasm in NAFLD and control subjects
Bowel preparation was good in 252 (66.3%) subjects, fair in 102 (26.8%), and poor in 26 (6.8%). The quality of bowel preparation did not differ among the study groups (p=0.14). Only one patient in the hospital cohort developed post-polypectomy bleeding which resolved after endoscopic therapy by clipping. No patient had bowel perforation.
Combining the community and hospital cohorts, NAFLD patients had higher prevalence of adenomatous polyps (34.7% vs 21.5%), polyps with villous architecture (6.0% vs 0.6%) or high grade dysplasia (18.1% vs 5.0%) than control subjects (table 2). Overall, advanced colorectal neoplasms were found in 18.6% of NAFLD subjects and 5.5% of control subjects (p=0.002). Colorectal cancers were identified in two (1.0%) NAFLD subjects and one (0.6%) control subject (p=0.65). After diagnosis, patients with colorectal cancers were interviewed again by the investigators and surgeons. None of them had bowel symptoms. All three patients had T1N0M0 cancer and recovered well after surgery. On the other hand, the prevalence of hyperplastic polyps was similar between NAFLD and control subjects.
After adjusting for demographic and metabolic factors, NAFLD remained as an independent factor associated with advanced neoplasms but not adenomas (table 3). Similarly, NAFLD subjects had higher numbers of adenomatous polyps, polyps with villous architecture and/or high grade dysplasia, and advanced colorectal neoplasms per patient (table 4). Overall, 140 colorectal adenomas were found in 199 NAFLD patients (0.7 adenoma per patient), compared to 57 adenomas in 181 controls (0.3 per subject). Among patients with adenomas, the average number of adenomas per subject was 2.0±2.2 in NAFLD patients and 1.5±0.8 in controls (p=0.050).
After stratification by age and gender, male and female NAFLD patients aged 40–50 years had a higher prevalence of colorectal adenomas and advanced neoplasms than controls, though the difference did not reach statistical significance because of the small number of subjects in each subgroup (figure 2).
NAFLD subjects were more likely to have proximal colorectal neoplasm (table 2). Among 69 NAFLD subjects with adenomatous polyps, 32 (46.4%) had lesions in the right sided colon, 22 (31.9%) had lesions in the left sided colon, and 15 (21.7%) had lesions in both areas. In contrast, among 39 control subjects with adenomatous polyps, 14 (35.9%) had right sided lesions, 23 (59.0%) had left sided lesions, and two (5.1%) had lesions in both areas. Among 29 NAFLD subjects with advanced colorectal neoplasms, 13 (44.8%) had right sided lesions, 11 (37.9%) had left sided lesions, and five (17.2%) had lesions in both areas. In comparison, among 10 control subjects with advanced neoplasms, three had right sided lesions and seven had left sided lesions.
Prevalence of colorectal neoplasm in subjects with simple steatosis and NASH
Among the 135 patients with biopsy proven NAFLD, NASH patients had a higher prevalence of adenomatous polyps (51.0% vs 25.6%) and polyps with high grade dysplasia (34.7% vs 12.8%) than subjects with simple steatosis (table 5). Overall, advanced colorectal neoplasms were detected in 34.7% of NASH patients and 14.0% of subjects with simple steatosis (p=0.011). One (2%) NASH patient and one (1%) subject with simple steatosis had colorectal cancer (p=0.74). Similarly, NASH patients had a higher prevalence of adenomas and advanced neoplasias than community controls. In contrast, the prevalence of adenomas and advanced neoplasms was similar between patients with simple steatosis and control subjects (table 5).
Using the scoring system of Kleiner and colleagues, 17 patients had NAFLD activity scores of 5 or above. Seven (41%) of them had colorectal adenomas and four (24%) had advanced neoplasms.
After adjusting for demographic and metabolic factors, NASH remained as an independent factor associated with adenomas and advanced neoplasms (table 6). Similarly, NASH subjects had higher numbers of adenomatous polyps, polyps with high grade dysplasia, and advanced colorectal neoplasms per patient (table 7). Overall, 32 adenomas were found in 86 patients with simple steatosis (0.4 adenomas per subject), compared to 63 adenomas in 49 patients with NASH (1.3 per subject). Although NAFLD patients generally had more right sided colorectal neoplasms, the distribution did not differ between subjects with simple steatosis and NASH (table 5).
In this cross-sectional study with prospective subject recruitment, NAFLD was associated with increased risk of colorectal adenomas and advanced colorectal neoplasm. Lesions in the right sided colon were more commonly found in NAFLD subjects. In addition, the increased risk was mainly observed in NASH subjects but not those with simple steatosis.
Compared to previous reports, our study included both subjects from the community and patients with biopsy proven NAFLD. The latter group provided additional data on the association between histological severity and the risk of colorectal neoplasm. Moreover, since 1H-MRS was performed in all community subjects, hepatic steatosis could be assessed accurately and objectively.20 21
BMI has been found to be associated with colorectal neoplasm in a number of retrospective studies.22–24 In another retrospective analysis of 2568 Japanese subjects who underwent two colonoscopy examinations 1 year apart, weight reduction was associated with a decreased incidence of adenomas at the second colonoscopy.25 The association translates into increased risk of colorectal cancer in obese people.12 19 Similarly, the risk of colorectal neoplasm is increased in patients with central obesity, metabolic syndrome, and type 2 diabetes.11 26–28 Recently, visceral fat as measured by imaging studies has also been found to be associated with colorectal adenomas.29 30 In a retrospective study in Korea, the prevalence of NAFLD was 41.5% in patients with colorectal adenomas and 30.2% in those without.31 Limited by its retrospective nature, the study did not describe the severity and distribution of the colorectal lesions in detail. Our study provided additional data and confirmed the strong association between NASH and advanced colorectal neoplasm.
While fatty liver can be detected by simple imaging tests such as ultrasonography, the diagnosis of NASH is based on histology. At present, this high risk group is not easy to identify in routine clinical practice. However, new biomarkers of NASH are currently under active development. For example, plasma cytokeratin-18 fragment level has been shown to have moderate accuracy in detecting NASH in cross-sectional and prospective studies.14 32 Therefore, non-invasive diagnosis of NASH in the clinic is highly probable in the future.
In the Veterans Affairs Cooperative Study and a community screening project in Hong Kong, faecal occult blood test and flexible sigmoidoscopy failed to detect a significant proportion of patients with advanced colorectal neoplasm because of the proximal location.17 33 Patients with metabolic syndrome are also more likely to develop cancer in the proximal colon.27 This has major clinical significance because such lesions may only be detected by full colonoscopy. In the current study, 46.4% of the NAFLD subjects with adenomas had lesions exclusively in the right sided colon, and 44.7% of the NAFLD subjects with advanced neoplasm had isolated right sided lesions. In other words, if only flexible sigmoidoscopy is used to screen NAFLD patients, nearly half of the significant lesions might be missed.
Current guidelines recommend colorectal cancer screening in adults aged 50–70 years.6–8 This is based on observational studies showing a low prevalence of advanced lesions in people below the age of 50.34 Nevertheless, in our study, both male and female NAFLD patients aged 40–50 years had a higher prevalence of colorectal adenomas and advanced neoplasms than controls. Although the findings need to be confirmed by other studies due to the small number of patients in each subgroup, this particularly high risk group may need earlier colorectal cancer screening.
The mechanism linking NAFLD and colorectal neoplasm is not completely understood. NAFLD represents a condition of profound insulin resistance and pro-inflammatory state.35 Insulin and insulin-like growth factor may promote the development of colorectal neoplasm through their proliferative and anti-apoptotic effects. NAFLD patients also have reduced expression of adiponectin, an adipokine with anti-inflammatory effects.35–37 In human studies, hypoadiponectinaemia is associated with increased risk of colorectal adenomas.38 Adiponectin inhibits colorectal cancer cell growth in vitro,39 and adiponectin deficient mice had increased risk of colon polyps when exposed to high fat diet.40
Our study had a few limitations. First, the demographic and metabolic profiles of NAFLD and control subjects were not matched. Since 20–30% of the general Asian population suffer from NAFLD, it was impossible to match the patients by clinical characteristics at the time of recruitment, before 1H-MRS was performed. However, NAFLD and NASH remained independently associated with colorectal neoplasm after adjusting for demographic and metabolic factors in different multivariate models. Second, although our study provided data from one of the largest liver histology cohorts, the overall number of study subjects was relatively small and could not support extensive subgroup analysis. Third, the diagnosis of fatty liver was based on histology in the hospital cohort and 1H-MRS in the community cohort. On the one hand, we need liver histology to distinguish between simple steatosis and NASH. On the other hand, liver biopsy is not ethically acceptable in the community cohort. Nevertheless, the similar prevalence of colorectal adenomas in NAFLD patients in both cohorts suggests that our observation was genuine (table 2). Finally, due to shared computer systems, endoscopists might be aware of the liver biopsy results. However, the high prevalence of colorectal adenomas in both the community and hospital NAFLD cohorts suggests that the findings were genuine because the endoscopists did not have access to the 1H-MRS results at the time of colonoscopy. The prevalence of colorectal lesions was consistent with that reported in our previous community screening project.17 Moreover, the similar prevalence of hyperplastic polyps across groups argues against ascertainment bias.
In conclusion, NASH is associated with a high prevalence of colorectal adenomas and advanced neoplasm. The adenomas are found more commonly in the right sided colon. Colorectal cancer screening is strongly indicated in this high risk group.
We would like to thank the following students and research assistants in helping with data collection: Shirley Chu, Andrew Hayward, Catherine Hayward, Mandy Law, Mia Li, and Karen Yiu.
See Commentary, p 745
Linked articles 239392.
Funding The study was supported by a grant from the Health and Health Services Research Fund sponsored by the government of Hong Kong SAR (Reference number 07080081).
Competing interests None declared.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of The Chinese University of Hong Kong.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.