Patients with ulcerative colitis (UC) have chronic symptoms, an impaired quality of life, an increased risk for complications, including colorectal cancer, and require lifelong medication or surgery 1. In the USA alone, ∼1.4 million individuals suffer from inflammatory bowel diseases (IBDs) including UC and Crohn’s disease (http://www.cdc.gov/ibd/#epidIBD). The economic costs are high, with an estimated global market for the drugs used to treat IBD at over $5 billion in the year 2009 2.
The etiology of UC is largely unknown, but there are plausible biological mechanisms for how, first, arachidonic acid (AA), a dietary n-6 polyunsaturated fatty acid (n-6 PUFA), may increase the risk for UC and, second, oleic acid (OA), an n-9 monounsaturated fatty acid (MUFA), may be protective against UC. AA is present in the phospholipid bilayer of cell membranes and can be metabolized to proinflammatory cytokines, including leukotriene B4 and prostaglandin E2. These cytokines are found in increased concentrations in the colonic mucosa of patients with UC 3–5; their levels are correlated with the degree of histological inflammation 6, and 5-amino salicylic acid drugs inhibit their formation 7. Dietary AA is chiefly derived from the n-6 PUFA linoleic acid and food sources such as red meat, and sunflower and corn oils. OA is present in high amounts in olive oil. OA competitively inhibits the enzyme [INCREMENT]5 desaturase, which in turn results in a potentially lower formation rate of AA-mediated proinflammatory metabolites, which further could lead to less colonic mucosal inflammation and potentially prevent UC. To confirm the experimental mechanisms for the role of these fatty acids, epidemiological studies are required, demonstrating that, first, high dietary intakes of AA, and, second, low dietary intakes of OA increase the risk of developing incident UC. The ideal study methodology would be prospective cohort work, which reduces both selection and recall biases associated with retrospective case–control work. To date, several epidemiological studies 8–13, only two of them being cohort studies 12,13, have reported that a diet rich in the essential n-6 PUFA, linoleic acid, which is metabolized to AA, increases the risk for UC. To the best of our knowledge, only one epidemiological study has investigated the association between OA and the etiology of UC, which reported no association 14. However, a limitation of these previous investigations includes a lower accuracy of the dietary monitoring instrument utilized – namely, food frequency questionnaires – which when used to measure habitual dietary intake in the populations studied perhaps measures this to a lower extent. A more precise method for dietary assessment is through 7-day food diaries, in which participants record all foods eaten, including quantities, brands, and cooking method 15. The accuracy of this method has been demonstrated, for example, when comparing the intake of energy from dietary fat, using 16-day-weighed records as the gold standard; the correlation coefficient was 0.77 for food diaries compared with 0.64 for food frequency questionnaires.
The aim of this study was therefore to conduct the first prospective cohort investigation assessing whether there are associations between the development of incident UC and the dietary intake of both AA and OA, using nutritional information derived from 7-day food diaries, which are to date the most accurate method of measurement of diet in large-scale epidemiological studies. Showing such associations between OA and AA and IBD would support, first, the measurement of their intake in further etiological studies and, second, the assessment of such a dietary modification in patients with established UC.
The cohort consisted of 25 639 healthy men and women, initially without UC, aged between 40 and 79 years, resident in the county of Norfolk, UK, who participated in the European Prospective Investigation into Cancer (EPIC)-Norfolk study. EPIC-Europe is a multicentered prospective cohort study of ∼520 000 participants from 10 European countries, including the cohort in Norfolk, which was designed to investigate the relationship between diet and the incidence of cancer and chronic diseases 16. Participants in the EPIC-Norfolk study were recruited from general practices in rural, suburban, and inner city areas, between the years 1993 and 1997, after obtaining their informed consent. At enrollment, they completed questionnaires on demography, medical history, medications, and physical activity. Information on cigarette smoking 17 and aspirin use 18, which can influence the etiology of IBD, was also recorded. At the baseline health check, a nurse explained how to complete the 7-day food diary, the first day of which was recorded with the nurse as a 24-hour recall of the participant’s previous day’s dietary intake. Entries for the remaining 6 days were then made at home by the participants themselves, who reported all foods eaten, including food types, portion sizes, brands, cooking methods, and recipes, in eight separate meal times each day. The names of commercially prepared foods from products consumed were included in the diary to allow more accurate nutritional assessments. Portion sizes were estimated by comparing the amount of food eaten with either a series of supplied photographs of varying household measures or recognizable standard units. Intake was reported in a series of meal slots, with space for information on additional snacks and other information. Information on consumption of supplements containing fish oils, as a binary variable (yes or no), was recorded. The 7-day food diaries were returned to the study headquarters, where they were coded by one of several trained nutritionists, who entered the data into a specially designed computer program called Data In to Nutrients for Epidemiological Research (DINER). Each food or drink entry in the diary was matched to one of 11 000 food items and 55 000 portion sizes within DINER, by selecting the one that best described it. DINER therefore facilitated the translation of information of food from participant-reported free text to structured data, which could then be electronically converted into nutrient values or food groups. A further series of programs and databases (DINERMO) performed detailed checks and translated the structured data into weights of food and nutrient data for analysis. The underlying nutrient database is based on the UK food composition database. An example of the detail of this method is that 337 specific types and brands of breakfast cereals are included in DINER. Each 7-day food diary took ∼2.5 h to code and interpret, with an average of 220 individual food and drink items reported by each participant in his/her diary. The computer program checked for potential errors in the coded diaries such as unexpectedly large portion sizes or duplication, and anomalies were checked by the nutritionists.
After recruitment, EPIC participants were followed up until June 2004 to identify those individuals who were subsequently diagnosed with new onset UC. Incident cases of UC were identified using information from self-reports on new illnesses and relevant medications recorded on follow-up questionnaires and from hospital inpatient and pathology databases. For all these potential cases, a consultant gastroenterologist reviewed the clinical notes to confirm the diagnosis, on the basis of standard radiological, endoscopic, and histological criteria. Clinical information was collected on both the confirmatory investigations and the extent of colonic inflammation. All participants with UC at recruitment and those who were diagnosed with UC less than 6 months after recruitment into the EPIC-Norfolk study were excluded. The latter helped ensure that the dietary data truly reflected the participants’ intake before the occurrence of symptoms and subsequent diagnosis.
The protocol of the EPIC study was approved by the Norwich District Ethics Committee, and all participants gave written, informed consent at recruitment for their clinical notes to be reviewed (NDEC Ref 98/122).
In this analysis, a cohort-nested case–control study, each case was matched with four controls for sex, date of birth (±6 months), date of recruitment into EPIC (±3 months), and similar follow-up times. Baseline characteristics including demographic characteristics and the median intakes of dietary fatty acids were compared between cases and controls using the Mann–Whitney test for nonparametrically distributed variables. The dietary intakes of AA and OA were divided into tertiles, from the distribution across both cases and controls, and conditional logistic regression was used to calculate the odds ratios (ORs) and 95% confidence intervals (95% CIs) for developing incident UC, adjusted for total energy intake and cigarette smoking. In a further multivariate analysis the above covariates were included, plus the essential n-6 PUFA, linoleic acid (which is metabolized to AA), and the three n-3 PUFAs (eicosapentanoic acid, docosahexanoic acid, α-linolenic acid), and finally OA, an n-9 MUFA. The rationale for the latter adjustment is that there is evidence for n-3 PUFAs, including eicosapentanoic and docosahexanoic acid, having anti-inflammatory properties, and they may therefore play a role in the prevention of UC 7. Finally, the trend ORs were adjusted for the addition of three separate variables, namely, the use of supplements containing fish oils, the use of aspirin, and social class. Adjustments were made for social class as this may be a surrogate marker for both education level and eating habits. All adjustments for potential confounders were defined before the analysis. The attributable fractions (AFs) for the proportion of cases associated with higher intakes of OA and AA were calculated using the equation AF=∑[((OR−1)/OR)×(% cases in that tertile)], in which the sum is of the higher two tertiles.
The data were analyzed using the IC-10 STATA statistical software package.
A total of 26 incident cases of UC (58% men, median age at diagnosis 70.4 years, range 46.8–80.8 years) were identified after a median follow-up period of 3.8 years (range 0.5–8.3 years). All patients showed histological evidence of UC, and 73% had undergone assessment of their disease extent, which was maximally up to the splenic flexure in 53% (Table 1). The proportion of smokers and the median energy intake were similar between cases and controls, although, as reported in previous work 19, cases consumed less docosahexanoic acid and eicosapentanoic acid.
The median intake of AA, without adjustment for any other variables, was similar among cases and controls (0.053 vs. 0.058 g/day, P=0.78). There was a lower intake of OA among cases, but without further multivariate analysis this was not statistically significant (17.75 vs. 18.51 g/day, P=0.60). In the first multivariate analysis, which included cigarette smoking and total energy intake only, there were no effects of increasing intake of AA or OA. However, in the analysis including adjustment for the intake of other dietary fatty acid variables, the two higher tertiles of AA intake were both associated with a statistically significant increased risk for incident UC (highest vs. lowest tertile, OR 6.09, 95% CI 1.05–35.23), with a trend across tertiles (trend OR 2.43, 95% CI 1.06–5.61). The attributable fraction for the two higher tertiles of AA intake was 57%. In the same analysis for OA, the two highest tertiles of intake were both inversely associated with the development of UC, with the highest tertile reaching statistical significance (OR 0.03, 95% CI 0.002–0.56) and an inverse trend across tertiles (OR 0.30, 95% CI 0.10–0.90, P=0.03).
The attributable fraction for the highest two tertiles of OA intake was 41%, that is the proportion of cases that could be prevented if the association was causal (Table 2). For both AA and OA, including the use of fish oil supplements, the use of aspirin, or social class did not affect the effect sizes for the trends across tertiles. In a sensitivity analysis of nonsmokers, this demonstrated an AA trend OR of 2.91 (95% CI 1.11–7.66, P=0.03) and an OA trend OR of 0.27 (95% CI 0.07–1.10, P=0.07). There were inadequate numbers of smokers among the cases to adjust for trend.
The main findings of this study were a six-fold increase in the odds of developing UC, associated with the highest tertile of intake of dietary AA, and an even larger decrease in the odds of developing UC among participants consuming higher amounts of OA. The strong effect sizes and dose–responses for both fatty acids provide support for causal associations between diet and incident UC. The attributable fractions suggest that these nutrients may be relevant in approximately half of all patients with UC. The lack of associations when all dietary fatty acids were included in the model suggests that there are complex interactions between these nutrients that can affect the development of UC.
AA is an n-6 PUFA (C20:4 n-6) obtained chiefly from diet and also through endogenous synthesis from the essential fatty acid, linoleic acid. The main sources of AA are linoleic acid, red meat and meat with a high fat content (e.g. pork, dark turkey, and chicken offal), polyunsaturated margarines, and cooking oils derived from soy, sunflower, rapeseed, and corn. OA, an n-9 MUFA, is the main fatty acid in olive oil and is also found in lesser quantities in pecan and peanut oils. Additional food sources that contain OA include grain-based desserts, chicken dishes, sausages, hot dogs, bacon, ribs, and dishes containing nuts and seeds 20. Our results support the a-priori hypothesis of a role for the AA-mediated proinflammatory pathway in the etiology of UC. AA is incorporated into colonocyte phospholipid membranes and its enzymatic conversion leads to the release of leukotrienes, including leukotriene B4, through the lipoxygenase pathway, and prostaglandins, such as prostaglandin E2, through the cyclo-oxygenase pathway. These AA-derived proinflammatory mediators are elevated in the serum and mucosa of patients with UC 3; there is a correlation between their mucosal levels and the histological activity of the disease 4–6, and 5-aminosalycilic acid drugs inhibit the formation of these eicosanoids 7. The potential etiological protective mechanism of OA is through the competitive inhibition of the enzyme [INCREMENT]5 desaturase, which reduces linoleic acid to AA. Consistent experimental and epidemiological data suggest that both these fatty acids should be measured in future dietary etiological and therapeutic studies on UC.
A major methodological advantage of this prospective study was the use of the 7-day food diaries to measure the habitual dietary intake of participants, a feature unique to the EPIC-Norfolk cohort. We are not aware of any other cohort studies investigating UC that have used such diaries in large-scale epidemiological studies. Using 16-day records as the gold standard, diaries are more accurate than either food frequency questionnaires or 24-h dietary recalls 21. This accuracy is because of the comprehensive recording of all individual foods, drinks, brands, and portion sizes, with each participant generating ∼220 lines of text on foods eaten (Figs 1–3). After completion, each diary was coded by a trained nutritionist using a specifically designed in-house computer program that converted participant-recorded free text into food and portion codes. Data were then processed to calculate nutrient composition using a food table from a database containing information on 55 000 food items. Further methodological advantages of this prospective study were that there were minimal selection and recall biases, as participants who served as cases and controls were drawn from the same baseline population, and information on diet was collected before the development of symptoms. There appeared to be minimal follow-up bias as the number of actual cases diagnosed was similar to that expected 22. Our work has limitations in that the number of cases was relatively small and the patients were from a defined geographical area; however, the incidence of ∼13/100 000/year was consistent with published data from the Europe-wide EC-IBD study 22. The expected incidence and the stability of the Norfolk population imply that follow-up bias should be minimal 21.
A major limitation of our study is that this cohort mainly included middle-aged to older individuals and traditionally UC is reported as a disease affecting younger individuals. However, this problem may only be relevant to women as the EC-IBD study reported that the incidence of UC peaked at ages between 25 and 34 years among men. Above this age, no significant decline with increasing age was observed in men, although in older women the incidence was larger than that observed in men 21. The EPIC cohort is the only European cohort with detailed dietary information on both male and female participants, who are monitored for IBD, and we are unaware of cohorts with similar detailed dietary data from younger populations. A further limitation was that the population studied was exclusively White, which limits the generalizability. There is evidence to suggest that ethnic origin can affect the bioavailability of some dietary components such as isoflavones, and further work investigating diet in ethnically diverse populations would be instructive 23.
In other aspects our cohort can be generalized in that both sexes were studied, the clinical distribution of colonic inflammation was similar to that expected, and the intake of the fatty acids was comparable with that in the general population 24. In addition, importantly, the average daily intake of n-6 PUFAs and MUFAs was similar across the age range of 19–64 years 24. Although our work demonstrated strong associations between OA and AA and UC, it is plausible that there are other dietary and environmental factors associated with, and that are involved in, the intake of AA and OA. This possible residual confounding will be addressed by continuing to investigate other nutrients, which will be included in future statistical models to see whether the associations with AA and OA are maintained. One such potential factor is dietary sulfur, which is present in higher quantities in red meat and is potentially harmful to the colonic mucosa 25.
Our findings support the emerging epidemiological evidence that AA may be involved in the etiology of UC, including work in which fatty acid intake was estimated from both food frequency questionnaires and adipose tissue biopsies 12,13. A prospective cohort study of 203 193 mainly middle-aged men and women, from five European countries (the IBD in EPIC study), including the Norfolk cohort, reported that higher dietary intakes of linoleic acid, which is converted to AA, as measured by food frequency questionnaires, were associated with a more than two times risk for developing UC (OR 2.49; 95% CI 1.23–5.07), with a significant trend across quartiles (trend OR 1.32, 95% CI 1.04–1.66) 12. Further, in the prospective EPIC-Denmark cohort, in which the dietary intake of AA was assessed using a biomarker, namely fatty acid concentrations in adipose tissue biopsies, there was an increased risk for incident UC in participants in the highest quartile (OR 4.16, 95% CI 1.56–11.04) 13. The large US cohort study of 170 805 women, which subsequently identified 338 incident cases of UC, reported no associations between AA and risk, although less accurate food frequency questionnaires were used to measure diet. Finally, two case–control studies reported statistically significant odds ratios of 5.1 (95% CI 1.0–26.7) 8 and 3.3 (95% CI 0.96–11.5) 9 for the higher intakes of total PUFAs and the chances of UC. Therefore, generally most, but not all, epidemiological work on associations with AA, as measured using three different dietary methods, dose–response effects, and the biological plausibility, are supportive evidence for a causal association. In a previous study, incident IBD in middle-aged women was found to be associated with increased protein intake, which may confound the association as AA is mainly derived from red meat 26. We assessed the relationship between animal protein and disease within our cohort and found no link (trend OR 0.47, 95% CI 0.15–1.44). Hypothetically, dietary OA may prevent the development of UC by inhibiting the AA-mediated proinflammatory pathway, and we documented an inverse association. Statistically, a nonsignificant inverse association was observed between OA and UC incidence for increasing quartiles of intake in the larger European EPIC study (in which our UK subcohort participated) 12; however, no link was noted in the US cohort 14. Again, this may have been because of the use of the less accurate food frequency questionnaires in both studies. In this statistical analysis we used tertiles, rather than quartiles as used in the full cohort, to generate greater statistical stability as there were fewer cases. A further new finding from this Norfolk study is that the previous Europe-wide EPIC study reported data on ALA and total n-6 PUFA, and not AA specifically (an important source of n-6 PUFAs). To confirm a causal association with the individual fatty acids, confirmatory evidence is required from further epidemiological studies on other populations and, ideally, randomized controlled trials in patients, investigating whether low intakes of AA and higher intakes of OA improve clinical outcomes. The number of incident Crohn’s disease cases in this cohort was only 11, which was insufficient for a meaningful statistical analysis.
A future research area to pursue is the potential interaction between diet and the gut microbiome in the etiology of IBD. There is increasing evidence that both lipid and carbohydrate intakes may alter the composition of the gut flora, which can affect mucosal inflammatory activity 27,28. Colonic bacterial flora stimulate inflammation in mouse models of UC, with an increased beef tallow intake leading to higher levels of proinflammatory fecal Bacteroidaceae and an increased fish oil intake resulting in higher quantities of the less inflammatory fecal Bifidobacteria 29. Further, patients with IBD have enhanced serological and T-cell responses to enteric microbial antigens and respond favorably to antibiotic treatment 30,31. Research is required to investigate whether AA increases and OA reduces the levels of these micro-organisms and consequently influence the etiology of UC. The balance between dietary n-6, n-9, and n-3 fatty acids is likely to be important as this can affect the inflammatory process. For example, n-3 PUFAs affect gene expression by binding to PPAR-γ, which downregulates proinflammatory interleukin-8 and interleukin-6 32,33. A higher dietary intake of n-9 MUFAs may reduce the effect of n-6 PUFAs by competitively inhibiting the metabolism of the latter, thereby reducing colonic inflammation.
In summary, this prospective study reports large positive associations between an increasing dietary intake of AA and inverse associations between an increasing dietary intake of OA, and the development of UC. Both these findings are supported by plausible biological mechanisms, large effect sizes, a dose–response, the temporality of the associations, and consistent findings from other epidemiological studies 12,13. To clarify whether the association is causal, these dietary fatty acids need to be studied in other populations and nutritional interventional trials need to be conducted in patients with established disease. Although previous small-scale dietary intervention studies have attempted to assess the influence of increased n-3 PUFA intake on clinical outcomes with some success, as far as we are aware, no trials have assessed a combined high n-3, n-9 PUFA and low n-6 PUFA dietary modification in UC 34,35. If the associations with fatty acids are confirmed, nutritional advice to the population may lead to a decrease in the incidence of UC and dietary modifications for patients may complement existing drug treatments.
The authors acknowledge the contribution of the staff and participants of the EPIC-Norfolk study.
de Silva: study design, statistical analysis and interpreting data, drafting the manuscript, and approval of the final draft submitted. Luben: statistical analysis and interpreting data, critical review of the manuscript, and approval of the final draft submitted. Shrestha: data collection, critical review of the manuscript, and approval of the final draft submitted. Khaw: planning the study, collecting data, analysis and interpreting data, critical review of the manuscript, approval of the final draft submitted. Hart: planning the study, collecting data, analysis and interpreting data, critical review of the manuscript, approval of the final draft submitted.
The EPIC-Norfolk study is supported by grants from the Medical Research Council program (G0401527, G1000143) and Cancer Research UK program (C864/A8257).
Conflicts of interest
There are no conflicts of interest.
1. Eaden JA, Abrams KR, Mayberry JF.The risk of colorectal cancer in ulcerative colitis
: a meta-analysis.Gut2001;48:526–535.
3. Sharon P, Ligumsky M, Rachilewitz D, Zor U.Role of prostaglandins in ulcerative colitis
. Enhanced production during active disease and inhibition by sulfasalazine.Gastroenterology1978;75:638–640.
4. Sharon P, Stenson WF.Enhanced synthesis of leukotriene B4 by colonic mucosa in inflammatory bowel disease.Gastroenterology1984;86:453–460.
5. Nishida T, Miwa H, Shigematsu A, Yamamoto M, Iida M, Fujishima M.Increased arachidonic acid
composition of phospholipids in colonic mucosa from patients with active ulcerative colitis
6. Wardle TD, Hall L, Turnberg LA.Inter-relationships between inflammatory mediators released from colonic mucosa in ulcerative colitis
and their effects on colonic secretion.Gut1993;34:503–508.
7. Travis SPL, Jewell DP.Salicylates for ulcerative colitis
: their mode of action.Pharmacol Ther1994;63:135–161.
8. Geerling BJ, Dagnelie PC, Badart-Smook A, Russel MG, Stockbrügger RW, Brummer RJ.Diet as a risk factor in the development of ulcerative colitis
.Am J Gastroenterol2000;95:1008–1013.
9. Reif S, Klein I, Lubin F, Farbstein M, Hallak A, Gilat T.Pre-illness dietary factors in inflammatory bowel disease.Gut1997;40:754–760.
10. Hart AR, Luben R, Olsen A, Tjonneland A, Linseisen J, Nagel G, et al..Diet in the aetiology of ulcerative colitis
: a European prospective cohort study.Digestion2008;77:57–64.
11. Sakamoto N, Kono S, Wakai K, Fukuda Y, Satomi M, Shimoyama T, et al..Dietary risk factors for inflammatory bowel disease: a multi-center case–control study in Japan.Inflamm Bowel Dis2005;11:154–163.
12. Hart AR.Linoleic acid, a dietary n-6 polyunsaturated fatty acid, and the etiology of ulcerative colitis
– a nested case–control study within a European prospective cohort study.Gut2009;58:1606–1611.
13. de Silva PS, Olsen A, Christensen J, Schmidt EB, Overvaad K, Tjonneland A, Hart AR.An association between dietary arachidonic acid
, measured in adipose tissue, and ulcerative colitis
14. Ananthakrishnan AN, Khalili H, Konijeti GG, Higuchi LM, de Silva P, Fuchs CS, et al..Long-term intake of dietary fat and risk of ulcerative colitis
and Crohn’s disease.Gut2013. [Epub ahead of print].
15. Jakes RW, Day NE, Luben R, Welch A, Bingham S, Mitchell J, et al..Adjusting for energy intake – what measure to use in nutritional epidemiological studies?Int J Epidemiol2004;33:1382–1386.
16. Bingham S, Riboli E.Diet and cancer-the European Prospective Investigation into Cancer and Nutrition.Nat Rev Cancer2004;4:206–215.
17. Calkins BM.A meta-analysis of the role of smoking in inflammatory bowel disease.Dig Dis Sci1989;34:1841–1854.
18. Chan SS, Luben R, Bergmann MM, Boeing H, Olsen A, Tjonneland A, et al..Aspirin in the etiology of Crohn’s disease and ulcerative colitis
: a European prospective cohort study.Aliment Pharmacol Ther2011;34:649–655.
19. John S, Luben R, Shrestha SS, Welch A, Khaw KT, Hart AR.Dietary n-3 polyunsaturated fatty acids
and the aetiology of ulcerative colitis
: a UK prospective cohort study.Eur J Gastroenterol Hepatol2010;22:602–606.
21. Banim PJ, Luben RN, Bulluck H, Sharp SJ, Wareham NJ, Khaw KT, Hart AR.The aetiology of symptomatic gallstones quantification of the effects of obesity, alcohol and serum lipids on risk. Epidemiological and biomarker data from a UK prospective cohort study (EPIC-Norfolk).Eur J Gastroenterol Hepatol2011;23:733–740.
22. Shivananda S, Lennard-Jones J, Logan R, Fear N, Price A, Carpenter L, van Blankenstein M.Incidence of inflammatory bowel disease across Europe: is there a difference between north and south? Results of the European Collaborative Study on Inflammatory Bowel Disease (EC-IBD).Gut1996;39:690–697.
23. Vergne S, Sauvant P, Lamothe V, Chantre P, Asselineau J, Perez P, et al..Influence of ethnic origin (Asian vs. Caucasian) and background diet on the bioavailability of dietary isoflavones.Br J Nutr2009;102:1642–1653.
24. Henderson L, Gregory J, Irving K, Swan G.The National Diet & Nutrition Survey. Vol 2. London: TSO; 2003. Chapter 5, pp. 53–59.
25. Jowett SL, Seal CJ, Pearce MS, Phillips E, Gregory W, Barton JR, Welfare MR.Influence of dietary factors on the clinical course of ulcerative colitis
: a prospective cohort study.Gut2004;53:1479–1484.
26. Jantchou P, Morois S, Clavel-Chapelon F, Boutron-Ruault MC, Carbonnel F.Animal protein intake and risk of inflammatory bowel disease: the E3N prospective study.Am J Gastroenterol2010;105:2195–2201.
27. Turnbaugh P, Ridaura VK, Faith JJ, Rey FE, Knight R, Gordon JI.The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice.Sci Transl Med2009;1:6ra14.
28. Swidsinski A, Ladhoff A, Pernthaler A, Swidsinski S, Loening-Baucke V, Ortner M, et al..Mucosal flora in inflammatory bowel disease.Gastroenterology2002;122:44–54.
29. Hekmatdoost A, Feizabadi MM, Djazayery A, Mirshafiey A, Eshraghian MR, Yeganeh SM, et al..The effect of dietary oils on cecal microflora in experimental colitis in mice.Indian J Gastroenterol2008;27:186–189.
30. Sartor RB.Microbial influences in inflammatory bowel diseases.Gastroenterology2008;134:577–594.
31. Macpherson A, Khoo UY, Forgacs I, Philpott-Howard J, Bjarnason I.Mucosal antibodies in inflammatory bowel disease directed against intestinal bacteria.Gut1996;38:365–375.
32. Marion-Letellier R, Déchelotte P, Iacucci M, Ghosh S.Dietary modulation of peroxisome proliferator-activated receptor.Gut2009;58:586–593.
33. Marion-Letellier R, Butler M, Déchelotte P, Playford RJ, Ghosh S.Comparison of cytokine modulation by natural peroxisome proliferator-activated receptor gamma ligands with synthetic ligands in intestinal-like Caco-2 cells and human dendritic cells – potential for dietary modulation of peroxisome proliferator-activated receptor gamma in intestinal inflammation.Am J Clin Nutr2008;87:939–948.
34. Uchiyama K, Nakamura M, Odahara S, Koido S, Katahira K, Shiraishi H, et al..N
-3 polyunsaturated fatty acid diet therapy for patients with inflammatory bowel disease.Inflamm Bowel Dis2010;16:1696–1707.
35. Grimstad T, Berge RK, Bohov P, Skorve J, Gøransson L, Omdal R, et al..Salmon diet in patients with active ulcerative colitis
reduced the simple clinical colitis activity index and increased the anti-inflammatory fatty acid index-a pilot study.Scand J Clin Lab Invest2011;71:68–73.