OUP user menu

Journal of Crohn's and Colitis: 10 (10)


Laurence J. Egan, Ireland

Associate Editors

Shomron Ben-Horin, IsraelSilvio Danese, ItalyPeter Lakatos, HungaryMiles Parkes, UKJesús Rivera-Nieves, USABritta Siegmund, GermanyGijs van den Brink, NLSéverine Vermeire, Belgium


Published on behalf of

Risk factors of non-alcoholic fatty liver disease in patients with inflammatory bowel disease

Achuthan Sourianarayanane, Gaurav Garg, Thomas H. Smith, Mujtaba I. Butt, Arthur J. McCullough, Bo Shen
DOI: http://dx.doi.org/10.1016/j.crohns.2012.10.015 e279-e285 First published online: 1 September 2013


Background: Metabolic risk factors are associated with non-alcoholic fatty liver disease (NAFLD), but they are less frequent in inflammatory bowel disease (IBD).

Aim: This study evaluates the frequency of NAFLD and its risk factors among IBD patients including anti-TNF-α therapy.

Methods: IBD patients who underwent abdominal imaging from January, 2009 to December, 2010 were analyzed in this nested, case-controlled study. IBD patients with NAFLD by imaging were compared with those who had no evidence of NAFLD (control).

Results: Among 928 IBD patients, 76 (8.2%) had evidence of NAFLD by imaging, and were compared to 141 patients without NAFLD evaluated (study: control ratio = ~ 1:2). NAFLD patients were older (46.0 ± 13.3 vs. 42.0 ± 14.1 years; p = 0.018) and had a later onset of IBD compared to the control group (37.2 ± 15.3 vs. 28.7 ± 23.8 years; p = 0.002). Metabolic syndrome was present in 29.0% of NAFLD patients, with a median Adult Treatment Panel risk factor of 2 [Interquartile range 1,3]. Patients not receiving anti-TNF-α therapy had a higher occurrence of NAFLD (p = 0.048). In multivariate analysis, hypertension (OR = 3.5), obesity (OR = 2.1), small bowel surgeries (OR = 3.7), and use of steroids at the time of imaging (OR = 3.7) were independent factors associated with NAFLD.

Conclusion: NAFLD occurred in 8.2% of the IBD population. NAFLD patients were older and had a later onset of IBD disease. IBD patients develop NAFLD with fewer metabolic risk factors than non-IBD NAFLD patients. It is also less common among patients who received anti-TNF-α therapy.

  • Anti-TNF-α therapy
  • Fatty liver
  • Inflammatory bowel disease
  • Risk factors

1 Introduction

Non-alcoholic fatty liver disease (NAFLD) is a clinico-pathological syndrome with a histologic spectrum ranging from benign steatosis to non-alcoholic steatohepatitis (NASH)1 along with subsequent development of cirrhosis and its complications in a proportion of these patients. The prevalence of NAFLD is increasing in the Western population, and studies have shown a high prevalence in the US (about 30%),2,3 where it has become the most common chronic liver disease.2 The third report of the expert panel on detection, evaluation, and treatment of high blood cholesterol in adults, Adult Treatment Panel (ATP) III defined metabolic syndrome by the presence of three of five risk factors: waist girth > 102 cm (M) or > 88 cm (F), or body mass index (BMI) > 30; arterial blood pressure ≥ 130/85 mm Hg; triglycerides (TG) ≥ 150 mg/dL; high density lipoprotein (HDL) cholesterol < 40 mg/dL for males or < 50 mg/dL for females; and fasting glucose ≥ 110 mg/dL.4 The prevalence of NAFLD is increased in patients who are obese and have metabolic syndrome.5,6 In patients with NAFLD, metabolic syndrome was present in 60% of women and 30% of men.7

Patients with chronic inflammatory diseases, such as inflammatory bowel disease (IBD), which consists of Crohn's disease (CD) and ulcerative colitis (UC), are prone to under-nutrition, and some require nutritional supplementation. Gut inflammation and bowel resection surgery can result in decreased absorptive capacity. Protein energy malnutrition (PEM) is frequently seen in patients with IBD.8 Depending on the severity of IBD, weight loss has been reported in 65–76% of CD patients and in 18–62% of UC patients.9 The reported frequency of under-nutrition in active IBD patients ranged from 25.0% to 69.7%, and the frequency of severe malnutrition ranged from 1.3% to 31.6%.10 On the other hand, metabolic syndrome and obesity are expected to be less common in IBD patients than in NAFLD patients, and normal lipid levels are reported in some patients with short gut syndromes.11

Tumor necrosis factor (TNF)-α plays an important role in the pathogenesis of both IBD and NAFLD. Pentoxifylline, an anti-TNF-α agent, was shown to benefit non-alcoholic steatohepatitis (NASH) patients.1215 Infliximab, an anti-TNF-α therapy used in IBD, may also benefit NAFLD patients,16 as it normalized transaminases and improved inflammation, necrosis, and fibrosis in an animal model,17,18 although it has not yet been proven to be clinically effective. Other treatments for IBD, such as steroids, could also affect the prevalence of NAFLD.

There are a number of other concurrent liver diseases reported in IBD patients, including primary sclerosing cholangitis (PSC).19,20 Fatty liver is frequently seen (38–49%) in CD or UC patients with abnormal liver test results such as elevated transaminases.1922 Recent increasing awareness of NAFLD in IBD is evidenced by the publications of case reports and series.23 An Italian study24 found steatosis by ultrasound in nearly 40% of IBD patients, who had a mean BMI of 21; however, to our knowledge, there are no studies from the US evaluating the frequency of NAFLD in IBD patients and the metabolic profiles of these patients. We hypothesized that a subgroup of IBD patients with their “unique” clinical risk factors would develop NAFLD. The hypothesis was tested by the following aims: 1) to assess the frequency of NAFLD in patients with CD or UC, and 2) to evaluate the risk factors for NAFLD in this population, including treatment with anti-TNF-α therapy.

2 Patients and methods

Medical records of consecutive patients undergoing evaluation and treatment for IBD at our tertiary care center between January, 2009 and December, 2010 were reviewed. Detailed demographic and clinical data, including data pertaining to IBD and liver diseases, were obtained. The study was approved by our Institutional Review Board. All patients were seen by IBD/nutrition specialists and, if indicated, by colorectal surgeons, and were managed according to standard of care.

2.1 Inclusion and exclusion criteria

Inclusion criteria were patients with 1) IBD and 2) abdominal imaging performed at our institution. Patients who had other types of liver diseases or who had a prior diagnosis of a liver disease other than NAFLD were excluded.

2.2 Study groups

Various imaging modalities (ultrasound of liver, CT scans with and without contrast, CT enterography, and MRI) were used to evaluate patients for evidence of NAFLD. The study (i.e., NAFLD) group included patients who had evidence of steatosis by any modality of imaging evaluated. This study group was compared with a sample population selected from the remainder of IBD patients who did not have NAFLD on imaging. The control group (i.e., no NAFLD) consisted of patients who did not have fatty liver on imaging. For analysis in this study, the control group of 141 patients (nearly 1:2 compared to study group) was obtained by analyzing patients who had presented in a fixed 2-month period from January, 2009 to February, 2009 and did not have steatosis on imaging.

2.3 Diagnostic criteria

The diagnoses of CD and UC were made based on a combined assessment of symptomatology, endoscopy, histology, and abdominal imaging. The pattern and distribution of CD and UC were obtained from endoscopic, radiological and operative findings. A diagnosis of NAFLD was made by an expert GI radiologist based on imaging of the liver. Typical radiographic findings indicative of NAFLD include a heterogeneous appearance of the liver, echogenicity exceeding that of the renal cortex or spleen by ultrasound, greater attenuation than in the spleen and blood, the appearance of intrahepatic vessels as relatively hypo-attenuated structures by unenhanced CT scan, and signal intensity loss on opposed-phase images in comparison with in-phase images by MRI.25 Six (7.9%) NAFLD patients also had liver biopsies, the results of which correlated with the imaging results.

2.4 Study variables

The type of IBD (CD or UC), the site of involvement of the CD (terminal ileal, colitis, small bowel, or combination) or UC (right side, left side, or pan colitis), and the need for parenteral nutrition were considered. In addition, current and prior treatments were evaluated, including therapy with biological agents (different forms of anti-TNF-α), immunomodulators [6 mercaptopurine (6MP), azathioprine], corticosteroids (prednisone and budesonide), or 5 amino-salicylic (5 ASA) agents. Surgical therapies were also considered (nature of surgery, type and length of bowel loss). Variables related to NALFD disease such as metabolic risk factors, liver serological tests, and histologic features were also reviewed.

2.5 Outcome measurements

The primary outcomes were the frequency of NAFLD detected by abdominal imaging and risk factors for NAFLD in this IBD cohort.

2.6 Statistical analysis

Univariable analysis was performed to compare subjects with and without NAFLD. Student t-tests or the non-parametric Wilcoxon rank sum tests were used for continuous factors, and Pearson chi-square tests were used for categorical variables. In addition, a multivariable logistic regression analysis was performed to assess risk factors associated with NAFLD in IBD patients. An automated stepwise variable selection method performed on 1000 bootstrap samples was used to choose the final multivariable model. p < 0.05 was considered statistically significant. All analyses were performed using SAS (version 9.2 software, The SAS Institute, Cary, NC) and R (version 2.13.1, The R Foundation for Statistical Computing, Vienna, Austria).

3 Results

A total of 1545 patients with diagnoses of IBD were evaluated at our facility. Among these patients, 602 who did not have imaging performed at our facility and 14 patients who had other liver pathology or risk factors (primary sclerosing cholangitis, viral hepatitis, or excessive alcohol consumption) were excluded from the study. The remaining 928 patients were included for further analysis.

3.1 Frequency of NAFLD

Among these 928 patients, 76 had evidence of NAFLD on imaging, giving a prevalence of NAFLD in IBD of 8.2% (95% CI: 6.4%, 9.9%). Their mean age was 46 ± 13.3 years, which is higher than the age of the remaining 852 patients without NAFLD (42.7 ± 14.2 years; p = 0.034).

3.2 Univariable comparison of demographic and clinical variables

Of the 852 patients who had imaging done at our center and but who did not have evidence of NAFLD, 141 patients who were evaluated during the 2-month period between January, 2009 and February, 2009 (a study: control ratio of nearly 1:2) were used as a comparison group. Their mean age was 42.0 ± 14.1 years, which was representative of the larger cohort of 852 patients without NAFLD (mean age of 42.7 ± 14.2 years), but was again lower than the mean age of the NAFLD group (p = 0.018) [Table 1 ]. Similar to this age disparity, the age of onset of IBD was higher in patients who had NAFLD compared to those without NAFLD (37.2 ± 15.3 vs. 28.7 ± 23.8 years; p = 0.002) [Table 1].

View this table:
Table 1

Demographic and clinical data.

NAFLD (N = 76)No NAFLD (N = 141)p-Value
Male31 (41.3)61 (43)0.69
Age at study46 ± 13.342 ± 14.10.018
Age at IBD37.2 ± 15.328.7 ± 23.80.002
Crohn's disease39 (53.4)70 (49.6)
Ulcerative colitis34 (46.6)71 (50.4)
Imaging modality0.99
CT abdomen32 (42.1)59 (41.8)
CTE37 (48.7)68 (48.2)
MRI abdomen2 (2.6)4 (2.8)
Ultrasound5 (6.6)10 (7.1)
  • NAFLD: non-alcoholic fatty liver disease; IBD: inflammatory bowel disease; CT abdomen: CT abdomen with and without contrast; CTE: computed tomogram enterography.

  • Values presented as Mean ± SD with t-test; Median [P25, P75] with Wilcoxon rank sum test, or N (%) with Pearson's chi-square test.

Among those patients with NAFLD, the mean age at the time of diagnosis of NAFLD was 45.9 ± 13.5 years, with a median time of 4.6 years from the diagnosis of IBD to the diagnosis of NAFLD [Interquartile range (IQR): 0.97, 13.3]. There were no differences in the gender or diagnosis (CD vs. UC) between the two groups, as shown in Table 1. The imaging modalities performed were similar between the two groups. There were no differences between the groups in terms of the distribution of disease or the extent of bowel involvement, for either CD or UC in patients (p = 0.1; data not shown).

Patients with NAFLD had a higher BMI (30.4 vs. 27.0; p < 0.001) and a higher frequency of obesity (51.3% vs. 27.7%; p < 0.001), hypertension (27.6% vs. 9.2%; p < 0.001), and metabolic syndrome (as defined as a minimum of three of five ATP III risks) (29.0% vs. 9.9%; p = 0.003). Furthermore, the NAFLD patients had a greater number of ATP III metabolic risk factors (median of 2 vs. 1; p < 0.001) and a higher frequency of metabolic risk factors (2 or more risk factors compared to 0 or 1) than the control group (Table 2 ).

View this table:
Table 2

Metabolic profile.

NAFLD (N = 76)No NAFLD (N = 141)p-Value
Body Mass Index30.4 ± 8.527.0 ± 6.1< 0.001
Obese39 (51.3)39 (27.7)< 0.001
Diabetes9 (11.8)10 (7.1)0.23
Hypertension21 (27.6)13 (9.2)< 0.001
Metabolic syndrome22 (29.0)14 (9.9)0.003
ATP metabolic risk factors2 [1,3]1 [1,2]< 0.001
02 (2.6)6 (4.3)
126 (34.2)91 (64.5)
226 (34.2)30 (21.3)
317 (22.4)7 (5)
44 (5.3)3 (2.1)
51 (1.3)4 (2.8)
HDL cholesterol43.2 ± 15.955.0 ± 19.90.049
Triglyceride133.5 [118.0, 224.0]112.0 [62.0,175.0]0.11
LDL cholesterol102.0 ± 55.1105.2 ± 44.10.84
TSH1.5 [0.86,2.8]1.8 [1.02,3.0]0.21
Hypothyroidism4 (5.3)3 (2.1)0.21
Vitamin D level22.4 [13.9,31.8]27.8 [18.1,38.3]0.23
  • ATP: Adult Treatment Panel; NAFLD: non-alcoholic fatty liver disease; TG: triglycerides; HDL: high density lipoprotein; LDL: low density lipoprotein; TSH: thyroid stimulating hormone.

  • Values presented as Mean ± SD with t-test; Median [P25, P75] with Wilcoxon rank sum test, or N (%) with Pearson's chi-square test.

Although there were no differences in the mean serum albumin levels or the median AST or ALT levels between the two groups, the number of patients with serum albumin below 3.5 g/dL was higher in the NAFLD group compared to the control group (34.4% vs. 16.7%; p = 0.007). Similarly, a higher number of patients with NAFLD had an elevated ALT (51.7% vs. 32.8%; p = 0.014) (Table 3). There were 11 patients in the NAFLD group (15%) with evidence of cirrhosis. A diagnosis of cirrhosis in these patients was based on liver histology (n = 4), abdominal imaging (n = 10), or the presence of ascites (n = 4), esophageal varices and splenomegaly (n = 4). There were no differences in age, gender, metabolic profile, or IBD treatment between those NAFLD patients with cirrhosis and those without (data not shown).

View this table:
Table 3

Liver characteristics.

NAFLD (N = 76)No NAFLD (N = 141)p-Value
Albumin g/dL*3.8 ± 0.74.0 ± 0.70.074
Aspartate transaminase*23.0 [17.0,35.0]20.0 [16.0,29.0]0.12
Alanine transaminase*25.0 [14.0,38.0]18.0 [13.0,31.0]0.13
Albumin < 3.5*21 (34.4)20 (16.7)0.007
Aspartate transaminase > 35*15 (24.6)20 (16.4)0.18
Alanine transaminase > 30 if male; alanine transaminase > 19 if female31 (51.7)40 (32.8)0.014
Platelets277.2 ± 120.2309.6 ± 103.40.054
Creatinine0.84 [0.68,0.98]0.80 [0.69,0.98]0.33
  • Note: * Missing some data.

  • NAFLD: non-alcoholic fatty liver disease; values presented as Mean ± SD with t-test; Median [P25, P75] with Wilcoxon rank sum test, or N (%) with Pearson's chi-square test.

Among the various medical and surgical treatments for IBD, patients with NAFLD had prior small bowel surgery more frequently (19.7% vs. 9.9%; p = 0.043), received anti-TNF-α therapy less often (17.1% vs. 29.7%; p = 0.048), and were more likely to be using corticosteroids at the time of imaging (25.0% vs. 9.2%; p = 0.002) (Table 4). There was, however, no difference in the number of patients who underwent any form of surgical therapy (small bowel or colon), who had prior exposure to corticosteroids or immunomodulators such as 6MP or azathioprine, or who needed parenteral nutrition (Table 4).

View this table:
Table 4

IBD-related medical and surgical therapy and NAFLD.

NAFLD (N = 76)No NAFLD (N = 141)p-Value
Any bowel surgery for IBD42 (55.3)78 (55.3)0.99
Terminal ileum resection21 (28.4)39 (27.7)0.91
Colectomy31 (40.8)55 (39.0)0.8
Small bowel surgery15 (19.7)14 (9.9)0.043
Parental Nutrition3 (3.9)5 (3.5)0.95
Use of anti-TNF-α biologics13 (17.1)35 (29.7)0.048
History of steroids44 (57.9)93 (66.0)0.24
Current steroid use19 (25.0)13 (9.2)0.002
Use of immunomodulators14 (18.9)43 (30.5)0.068
  • IBD: inflammatory bowel disease; NAFLD: non-alcoholic fatty liver disease.

  • TNF: tumor necrosis factor.

  • Values presented as N (%) with Pearson's chi-square test.

3.3 Multivariable analysis of risk factors for NAFLD

In multivariable logistic regression analysis, small bowel surgery, hypertension, obesity, corticosteroid use at the time of imaging, anti-TNF-α biologic therapy, and therapy with immunomodulators were included in the final model. Subjects who had small bowel surgery were 3.7 times more likely to have NAFLD than those who did not require surgery, after adjusting for all other variables in the model. In addition, hypertension was associated with a 3.5 times higher likelihood of NAFLD. Compared to non-obese subjects, obese patients were twice as likely to be diagnosed with NAFLD. Finally, subjects who were on steroids at the time of evaluation were almost 4 times more likely to have NAFLD (Table 5 ).

View this table:
Table 5

Risk factors for NAFLD in IBD patients: multivariable logistic regression analysis.

Odds ratio (95% CI)p-Value
Small bowel surgery3.7 (1.5, 9.3)0.005
Hypertension3.5 (1.5, 8.1)0.004
Obesity2.1 (1.05, 4.0)0.035
Steroid use at imaging3.7 (1.5, 9.3)0.005
No use of anti-tumor-necrosis factor-α biologics1.9 (0.83, 4.3)0.13
No use of immunomodulators1.7 (0.74, 3.9)0.21
Age0.3 (0.1,1.05)0.06

4 Discussion

This study evaluated the frequency of NAFLD documented by imaging in a large cohort of IBD patients. NAFLD was diagnosed in 8.2% of IBD patients, which is lower than the frequency reported in the general population of the United States, where an estimated 33.6% have NAFLD.3 However, this rate is similar to other reported studies, in which approximately 10% IBD patients had NAFLD.20,23 In our study, IBD patients with NAFLD were older than those without NAFLD. This could reflect a similar trend seen in population-based studies, where an increase in the prevalence of metabolic risks leading to NAFLD increased with age.26,27 Patients diagnosed with NAFLD were also initially diagnosed with IBD at a higher age, with a median time of 4.6 years between the diagnosis of IBD and the diagnosis of NAFLD. In our study, IBD patients with NAFLD had a higher number of metabolic risk factors (ATP III) than those with IBD alone and a similar higher prevalence of metabolic syndrome (29.0% vs. 9.9%; p = 0.003). The prevalence of metabolic syndrome was however lower among our study patients both with NAFLD and without when compared to its prevalence among the general US population of 30–60% and ~ 30% with and without NAFLD respectively.7,28 A low occurrence of metabolic syndrome in IBD, especially among CD patients (7%), is also reported.29 This could also explain the difference in age between groups, with a lower frequency of metabolic risk (and hence a lower frequency of NAFLD) among who developed IBD at a younger age.

The late onset of IBD in those with NAFLD, along with the increased occurrence of metabolic risk in those patients, may also suggest a different phenotype of IBD or a different patient population with different risks. Patients with NAFLD had undergone small bowel surgery more frequently despite less usage of anti TNF-α therapy without any other significant difference in the nature of disease (CD vs. UC), the pattern of disease (site of involvement) or overall need for steroid and immunomodulators therapy. Although it is known that metabolic risk factors increase with age,27 significant differences in the prevalence of these factors have only been seen over decades, whereas there was only a 4-year age difference found between those with NAFLD and those without NAFLD in our study. The age difference between the patients with NAFLD and those without cannot therefore entirely account for the difference in metabolic risk factors between the two groups. Moreover, age was not a significant factor in the multivariate analysis. There has been evidence of alcohol like steatosis reported in patients who have undergone jejuno-ileal bypass surgeries30,31 or similar obesity related surgeries and also in patients who have undergone massive small bowel resections for other causes.30,32 In these patients mild to moderately severe centrilobular fatty changes as seen in NAFLD patients on liver histology were found.32 An increase in plasma free fatty acids, decreased carnitine concentrations and essential fatty acid deficiency are considered to play a role in causing liver steatosis following small bowel surgeries.30 Although, in our patients massive small bowel resection was not seen, the need for small bowel surgery may suggest increased involvement of small bowel by IBD with resultant decreased small bowel function as seen in jejuno-ileal bypass surgery patients or after massive small bowel resection.

Steroid use at the time of evaluation was a significant factor among those with NAFLD by both univariate and multivariate analyses. There are reports of new onset of liver steatosis following 40 days of aggressive systemic steroid therapy in IBD.33 The increased occurrence of steroids use at the time of evaluation could also suggest its frequent use in the management of difficult IBD, although the duration and dose of steroids were not evaluated in this retrospective study. On the other hand, exposure to anti-TNF therapy may be protective against the development of NAFLD. Although clinical studies are lacking, anti-TNF-α therapy has been found to reduce liver disease in animal studies.17,18 This, along with the finding that small bowel surgery was required more frequently in patients with NAFLD, may indicate a different phenotype of IBD with metabolic risks as compared to the IBD of patients without NAFLD.

There are many studies in the literature where NAFLD, a common chronic liver disease in the United States, was found in patients with IBD, although these patients are expected to be under-nourished9,10 rather than obese, as is frequently seen in patients with NAFLD. Many of these studies are case reports or case series,2123,3436 or studies on the presence of NAFLD in IBD patients with abnormal liver tests.33,37 This last category includes the largest study so far, which identified NAFLD by imaging in 120 of 786 IBD patients with abnormal liver tests. Although an Italian study evaluated the prevalence of NAFLD in the IBD population by ultrasound imaging, patients in this study had a surprisingly high prevalence of steatosis (35–40%) despite a low body mass index (BMI) of 21.24 Many other smaller studies had suggested the prevalence to be around 10%.20,23 Our study evaluated nearly 1000 IBD patients by imaging over a 2-year period to determine the prevalence of NAFLD in this population. We were also able to associate potential metabolic risk factors with the development of NAFLD in IBD patients and also to relate the IBD disease pattern and its treatment to the development of NAFLD, which had not been previously evaluated.

The findings of this study have clinical implications. Currently, routine liver evaluation for NAFLD is not done in all IBD patients. With the current study demonstrating that NAFLD develops in a subset of IBD patients with relatively few metabolic risk factors, assessment for NAFLD should be considered. Our study demonstrates that IBD patients with hypertension who are older and who require steroid therapy frequently are at increased risk for the development of NAFLD even if they do not have metabolic syndrome, as hypertension and diabetes were associated with NAFLD in IBD patients. Hence, liver imaging may be of value in older IBD patients who have elevated blood pressure with obesity especially when they require frequent steroid therapy for control of their IBD.

There are limitations to this study. In our study, diagnosis of NAFLD was based on imaging. Ultrasound imaging modality has a sensitivity, specificity and positive predictive value (80 to 100%) in detecting steatosis.38,39 CT and MRI imaging have been found to have a higher sensitivity and specificity compared to ultrasound imaging.38,40 The sensitivity by these imaging modalities decreases when fat content in liver is less than 30%.25 The results can be also be lower, by ultrasound imaging being an operator dependent especially in patients with morbid obesity (BMI > 40) when the sensitivity is less (40%).38 In our study, the mean BMI of study group was 30.4 ± 8.5 and the control group was 27.0 ± 6.1. Hence, most of patients evaluated had BMI less than 40. Moreover, most of our patients had CT or MRI imaging, which have better sensitivity compared to ultrasound imaging modality in the diagnosis of steatosis.38,40 Although we were diligent in obtaining records of the medications patients received, both in the past and at the time of imaging, the frequency and duration of medications such as steroids and anti-TNF-α therapy could not be addressed in this retrospective study. Because patients in the two groups were similar in terms of IBD disease phenotype and the total number of surgical interventions required, it can be presumed that the life time and frequency of steroid use were also similar between the two groups, although this cannot be confirmed. A future prospective study could address the finding that more patients with NAFLD were on steroids at the time of evaluation and assess if these patients require steroid medication more often. This would allow us to assess the effect of cumulative steroid dose in these patients, and also may explain the less frequent use of anti-TNF-α therapy in NAFLD patients, which would provide a possible explanation for the observed IBD disease phenotype variability. A prospective study could also eliminate some additional variables concerning the liver tests and other IBD treatments which were not fully analyzed in this study. However, a prospective study is less likely to address the age difference seen in our study, which was also not found to be a significant factor in the multivariate analysis.

In conclusion, patients with IBD appeared to have a lower risk for NAFLD than the general population. Those who developed NAFLD tended to be older, to develop IBD at an older age, and to require small bowel surgery more often.

Conflict of interest

None of the authors have conflicts of interest.



Authors role:

AS: Study design, data acquisition, analysis, interpretation, and manuscript drafting/revision.

GG: data acquisition, analysis, manuscript revision.

THS: data acquisition, interpretation, manuscript drafting/revision.

MIB: data acquisition, interpretation, manuscript drafting/ revision.

AJM: design, supervision, manuscript drafting/revision.

BS: study concept, interpretation, manuscript drafting/revision.

All authors read and approved the final manuscript.


  • Accepted as poster: American college of Gastroenterology, Las Vegas, 2012.

aminosalicylic acids
Adult Treatment Panel
Crohn's disease
high density lipoprotein
inflammatory bowel disease
low density lipoprotein
protein energy malnutrition
primary sclerosing cholangitis
non-alcoholic fatty liver disease
non-alcoholic steatohepatitis
tumor necrosis factor
thyroid stimulating hormone
ulcerative colitis


View Abstract