Next Article in Journal
Changes in the Sodium Content of Australian Processed Foods between 1980 and 2013 Using Analytical Data
Previous Article in Journal
A Proposal for a Study on Treatment Selection and Lifestyle Recommendations in Chronic Inflammatory Diseases: A Danish Multidisciplinary Collaboration on Prognostic Factors and Personalised Medicine

Nutrients 2017, 9(5), 500; https://doi.org/10.3390/nu9050500

Article
Macronutrient Intake and Risk of Crohn’s Disease: Systematic Review and Dose–Response Meta-Analysis of Epidemiological Studies
Department of Gastroenterology, The Central Hospital of Enshi Autonomous Prefecture, Enshi 445000, China
*
Correspondences: [email protected] (L.Z.); [email protected] (W.W.); Tel.: +86-0718-822-4327 (L.Z.)
L.Z. and S.H. contributed equally to this study.
Received: 22 April 2017 / Accepted: 11 May 2017 / Published: 15 May 2017

Abstract

:
Dietary intake is potentially associated with the onset of Crohn’s disease (CD), but evidence from epidemiological studies has remained unclear. This study aimed to evaluate the role of macronutrient intake in the development of CD. A systematic search was conducted in PubMed and Web of Science to identify all relevant studies, and the role of macronutrients in the development of CD was quantitatively assessed by dose–response meta-analysis. Four case-control studies (a total of 311 CD cases and 660 controls) and five prospective cohort studies (238,887 participants and 482 cases) were identified. The pooled relative risks (RR) for per 10 g increment/day were 0.991 (95% confidence interval (CI): 0.978–1.004) for total carbohydrate intake, 1.018 (95% CI: 0.969–1.069) for total fat intake, and 1.029 (95% CI: 0.955–1.109) for total protein intake. Fiber intake was inversely associated with CD risk (RR for per 10 g increment/day: 0.853, 95% CI: 0.762–0.955), but the association was influenced by study design and smoking adjustment. In subtypes, sucrose intake was positively related with CD risk (RR for per 10 g increment/day: 1.088, 95% CI: 1.020–1.160). Non-linear dose–response association was also found between fiber and sucrose intake and CD risk. In conclusion, this meta-analysis suggested a lack of association between total carbohydrate, fat or protein intake and the risk of CD, while high fiber intake might decrease the risk. In subtypes, high sucrose intake might increase the risk of CD.
Keywords:
macronutrient intake; Crohn’s disease; disease risk; dose–response; meta-analysis

1. Introduction

Crohn’s disease (CD) is an intestinal inflammatory disorder of unknown etiology, and its incidence has been steadily on the rise [1,2]. The disease is clinically characterized by abdominal pain, diarrhea and high rates of surgery caused by intestinal stenosis and perforation. The main therapeutic strategy is to alleviate active inflammation, maintain remission and reduce the rates of relapse and surgery. Non-steroidal anti-inflammatory drugs (e.g., 5-aminosalicylic acid (5-ASA)) are effective for mild to moderate disease, while corticosteroids, immunosuppressors (e.g., azathioprine), or biological agents (e.g., infliximab) are required for moderate to severe disease or when 5-ASA proves ineffective [3,4]. However, drug dependence, side effects, medical costs and poor life quality impose a heavy burden on CD patients. Thus, it is necessary to identify the etiology or risk factors, and prevent the disease from the source. Diet is usually considered to a potential pathogenic factor, especially for westernized dietary habits [5,6]. Secondly, the treatment with elemental diet and food exclusion has a similar effect to the drug therapy [7,8]. As a result, it was believed that certain food or nutrients might play an important role in the pathogenesis of CD. For dietary food, high consumption of fruit was reported in inverse association with the risk of CD (odds ratio (OR): 0.57, 95% confidence interval (CI): 0.44–0.74), but not in vegetables (OR: 0.66, 95% CI: 0.40–1.09) [9]. For macronutrients, fat, carbohydrate, fiber or protein intake was not significantly associated with another type of inflammatory bowel disease, ulcerative colitis [10,11]. However, the role of macronutrient intake is controversial in the pathogenesis of CD, and no meta-analyses have concentrated on this. Therefore, we conducted a systematic review and dose–response meta-analysis to quantitatively assess the role of macronutrient intake in the development of CD.

2. Materials and Methods

2.1. Search Strategy

The databases of PubMed and Web of Science were searched for relevant studies published up to 16 March 2017, using the key words including “diet*”, “carbohydrate”, “sugar”, “fiber”, “fat*” and “protein” in combination with “inflammatory bowel disease” or “Crohn’s disease”. Studies in languages other than English or Chinese were excluded. Moreover, we also reviewed the references of related studies and reviews for undetected studies.

2.2. Study Selection and Exclusion

Two authors (L.Z. and P.C.) reviewed the studies independently. The inclusion criteria were as follows: (i) case-control or cohort-based study design; (ii) contained at least three quantitative categories of macronutrient intake; (iii) evaluated the association between macronutrient intake and CD risk; (iv) presented relative risk (RR), OR, or hazard ratio (HR) estimates with 95% CI. The exclusion criteria were as follows: abstracts without full texts, reviews, case reports and pediatric studies.

2.3. Data Extraction and Quality Assessment

Two authors (L.Z. and P.C.) extracted the data by a standardized collection form. All differences were resolved by discussion with a third author (W.W.). In each study, the following information was extracted: first author, publication year, area, study design, follow-up year in prospective-designed studies, diagnostic criteria, number of cases/controls, age distribution, time at diagnosis and retrospective period in case-control studies, exposure categories, effect sizes (RR, HR, OR) with 95% CI and adjusted factors. The Newcastle–Ottawa Scale (NOS) was used to assess the methodological quality of included studies [12].

2.4. Statistical Analysis

For the low incidence of CD, OR and HR were roughly regarded as the RR in this study [10,11,13]. The assigned dose in each category was defined as the mean intake. If the mean intake per category was unavailable, we chose the midpoint of the upper and lower boundaries in each category as the assigned dose. For open-ended lower and upper categories, we defined the lowest boundary as zero and the open-ended interval length as the same with the adjacent category respectively. Groups were regarded in equal size or follow-up when cohort size or person-year per category was unavailable, and the case number per category was obtained by the method of Bekkering et al. [14]. In the Reif et al. study [15], the missing 95% CIs were obtained according to the method of Orsini [16]. In the Jantchou et al. study [17], we converted the unit of grams per kilogram of body weight to grams per day by multiplying an average weight of 60 kilograms. In the Chan et al. study [18], the person-years were calculated assuming that the interval between recruitment and diagnosis was the follow-up period.
As the cut-off points for categories varied among the studies, the RR for per 10 increment of exposure in each study was estimated before pooling the risk estimates between studies using a random-effect model by the method of Greenland and Longnecker [19] and Orsini et al. [20]. Moreover, a potential non-linear dose–response association between nutrient intake and CD risk was modeled by using restricted cubic splines with three knots at percentiles 10%, 50% and 90% [21]. The Wald test was chosen to evaluate linear or non-linear trends [22].
Subgroup analyses were conducted on study design, cohort and smoking adjustment to evaluate the stability of main results The Egger’s test was used to detect publication bias [23]. The heterogeneity among studies was estimated by Q test and I2 statistic [24], and I2 > 50% represented substantial heterogeneity. All statistical analyses were performed with Stata SE12.0 software (StataCorp LP, College Station, TX, USA).

3. Results

3.1. Study Characteristics

The search strategy resulted in 12,155 records: 10,980 from PubMed, 1119 from Web of Science and 16 through other sources. After excluding duplicated and irrelevant records, eight records (nine studies) were included in this meta-analysis [15,17,18,25,26,27,28,29] (Table 1). Persson et al. [25] classified the results by sex, which were divided into two separate reports. The two studies by Ananthakrishnan et al. [27,28] were based on the same cohort, and focused on fat and fiber respectively. Among these eight studies, four were case-control designed with a total of 311 CD cases and 660 controls, while the other five were prospective cohort designed with a total of 238,887 participants and 482 cases. Validated semi-quantitative food frequency questionnaires (FFQ) contained various food items and consumption frequency, and were used to measure macronutrient consumption in all studies. The results in most studies were statistically adjusted for certain factors, such as age and energy intake. In case-control studies, dietary habits before diagnosis were obtained to guarantee the pre-illness dietary intake. In quality assessment, the included studies had an average score of 7.78.

3.2. Carbohydrate Intake and CD Risk

Four studies investigated the association between carbohydrate intake and CD risk (Figure 1). No studies showed a significant association, and no evidence of a non-linear relationship between them was detected (p for non-linearity was 0.376). The pooled RR was 0.991 (95% CI: 0.978–1.004, I2 = 0.0%, pheterogeneity = 0.439) for per 10 g increment/day in carbohydrate intake.

3.3. Fiber Intake and CD Risk

Five studies reported the association between fiber intake and CD risk (Figure 1). The pooled RR was 0.853 (95% CI: 0.762–0.955, I2 = 0.0%, pheterogeneity = 0.730) for per 10 g increment/day in fiber intake, suggesting a protective role in the development of CD. We also found a non-linear relationship between them (p for non-linearity was 0.019) (Figure 2).

3.4. Fat Intake and CD Risk

Five studies investigated the association between fat intake and CD risk (Figure 1). Only the study of Sakamoto et al. showed a significant association (RR: 1.134, 95% CI: 1.030–1.249), and no evidence of a non-linear relationship between them was detected (p for non-linearity was 0.281). The pooled RR was 1.018 (95% CI: 0.969–1.069, I2 = 44.6%, pheterogeneity = 0.125) for per 10 g increment/day in fat intake.

3.5. Protein Intake and CD Risk

Five studies reported the association between protein intake and CD risk (Figure 1). No studies showed a significant association, and no evidence of a non-linear relationship between them was detected (p for non-linearity was 0.163). The pooled RR was 1.029 (95% CI: 0.955–1.109, I2 = 54.7%, pheterogeneity = 0.085) for per 10 g increment/day in protein intake.

3.6. Intake of the Nutrients’ Subtypes and CD Risk

Four studies reported an association between the intake of sugar subtypes and CD risk, including monosaccharide, disaccharide, starch and related subtypes (Table 2). Only sucrose was found to be significantly related with CD risk, and the pooled RR was 1.088 (95% CI: 1.020–1.160, I2 = 0.0%, pheterogeneity = 0.395) for per 10 g increment/day. A non-linear relationship was also found with a p value of 0.023 for non-linearity (Figure 2). Three studies reported an association between intake of fat subtypes and CD risk, including saturated fatty acid (SFA), monounsaturated fatty acid (MUFA), polyunsaturated fatty acid (PUFA) and their subtypes. No fat subtypes showed a significant association with CD risk. In the Jantchou et al. study [17], animal-or vegetable-derived protein showed no significant association with CD risk (the highest vs. the lowest category:RR: 2.700, 95% CI: 0.690–10.520; RR: 1.040, 95% CI: 0.280–3.800).

3.7. Subgroup Analysis and Publication Bias

Subgroup analyses were conducted on study design, cohort and smoking adjustment. No substantial changes of primary results were found between groups except for fiber intake (Table 3). The protective role of fiber intake in the development of CD was weakened by the risk factor of smoking (RR for per 10 g increment/day: 0.890, 95% CI: 0.776–1.020). The Egger’s test detected no obvious publication bias in fiber (p = 0.708), while there were not enough studies to conduct the analyses for carbohydrate, fat and protein.

4. Discussion

To the best of our knowledge, this is the first dose–response meta-analysis to study the role of macronutrient intake in the development of CD. Our results suggested a lack of association between dietary carbohydrate, fat or protein intake and CD risk, which was inconsistent with the conventional views. The Western lifestyle is characterized by high consumption of carbohydrates, fats and protein, and the prevalence of westernized dietary habits in Asian cohorts coincides with an increasing incidence of CD in those regions. Thus, these macronutrients were usually thought to be risk factors in the development of CD. Cola drinks and chocolate are rich in carbohydrates and fats, and in the study of Russel et al., both were found in positive association with the development of CD (OR (95% CI): 2.2 (1.5–3.1); 2.5 (1.8–3.5)) [30]. However, in the Racine et al. study, the dietary pattern of “high sugar and soft drinks” showed no significant association with CD risk (the highest vs. the lowest category: RR (95% CI): 1.48 (0.60–3.61)), as well as the pattern “animal fats, seafood, potatoes and alcohol” (RR (95% CI): 0.71 (0.29–1.73)) [31]. It was controversial on the role of these macronutrients in the development of CD, and environmental and genetic factors might contribute to the inconsistency in findings between studies [32,33].
For carbohydrate intake, we detected no significant association with CD risk. However, as another major subtype of carbohydrate, high fiber intake could decrease CD risk. The result was consistent with the meta-analysis of Liu et al., which also found dietary fiber intake could reduce the risk of CD (the highest vs. the lowest category: RR (95% CI): 0.44 (0.29–0.69)) [34]. In subgroup analysis, when not adjusted for smoking, a lower RR was observed with higher fiber intake, but the same finding was not present with an adjustment for smoking. This indicated that a low fiber might be simply associated with smoking (which is a risk factor for Crohn’s disease), rather than being an independent risk factor for CD [35]. Furthermore, the limited number of studies might also contribute to the inconsistency. Third, the protective role of fiber might be weakened by the risk factor of smoking.
As another major subtype of carbohydrate, sugar, was not associated with CD risk in this study, which was consistent with the study of Racine et al. However, in respect to sugar subtypes, high sucrose intake was positively associated with CD risk. In animal models, high consumption of dietary sucrose could induce the inflammation of multiple tissues [36,37,38]. Secondly, high intake of dietary sucrose could trigger endoplasmic reticulum (ER) stress which was also associated with the pathogenesis of Crohn’s disease [39,40]. Wang et al. also reported a positive association between sucrose intake and the risk of ulcerative colitis (RR for per 10 g increment/day: 1.098, 95% CI: 1.024–1.177). Thus, we thought that high intake of dietary sucrose could increase the risk of CD [11].
This meta-analysis study has several strengths. First, to our knowledge, this is the first dose–response meta-analysis to identify the role of macronutrients in the development of CD. Second, only studies with at least three quantitative categories of exposure were included which demonstrated a higher quality than those with two categories, and somewhat covered the limits in the number of studies [41]. Third, subgroup and sensitivity analyses were used to test the stability and reliability of primary results, and the results were consistent in general. There were also a few limitations in this study. First, the inclusion of case-control studies might introduce certain bias, such as recall bias, which might potentially lead to differential misclassification of various types of exposure, and exaggerate or weaken the effect estimates. Second, not all potential confounders were adjusted in each study.

5. Conclusions

This meta-analysis suggested a lack of association between total carbohydrate, fat or protein intake and the risk of CD. High fiber intake might decrease the risk of CD, but the association was influenced by study design and smoking adjustment. In subtypes, high sucrose intake might increase the risk of CD. Large-scale prospective designed studies are needed to confirm our findings.

Author Contributions

L.Z. and P.C. conceived and designed the experiments; L.Z. and P.C. performed the experiments; S.H. and P.C. analyzed the data; W.W. and Y.T. contributed reagents/materials/analysis tools; L.Z. and P.C. wrote the paper.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Molodecky, N.A.; Soon, I.S.; Rabi, D.M.; Ghali, W.A.; Ferris, M.; Chernoff, G.; Benchimol, E.I.; Panaccione, R.; Ghosh, S.; Barkema, H.W.; et al. Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review. Gastroenterology 2012, 142, 46–54. [Google Scholar] [CrossRef] [PubMed]
  2. Ng, S.C.; Tang, W.; Ching, J.Y.; Wong, M.; Chow, C.M.; Hui, A.J.; Wong, T.C.; Leung, V.K.; Tsang, S.W.; Yu, H.H.; et al. Incidence and phenotype of inflammatory bowel disease based on results from the Asian-pacific Crohn’s and colitis epidemiology study. Gastroenterology 2013, 145, 158–165. [Google Scholar] [CrossRef] [PubMed]
  3. Lichtenstein, G.R.; Abreu, M.T.; Cohen, R.; Tremaine, W. American Gastroenterological Association Institute medical position statement on corticosteroids, immunomodulators, and infliximab in inflammatory bowel disease. Gastroenterology 2006, 130, 935–939. [Google Scholar] [CrossRef] [PubMed]
  4. Wang, F.; Lin, X.; Zhao, Q.; Li, J. Adverse symptoms with anti-TNF-alpha therapy in inflammatory bowel disease: Systematic review and duration-response meta-analysis. Eur. J. Clin. Pharmacol. 2015, 71, 911–919. [Google Scholar] [CrossRef] [PubMed]
  5. Anderson, V.; Olsen, A.; Carbonnel, F.; Tjønneland, A.; Vogel, U. Diet and risk of inflammatory bowel disease. Dig. Liver Dis. 2012, 44, 185–194. [Google Scholar] [CrossRef] [PubMed]
  6. Ng, S.C.; Bernstein, C.N.; Vatn, M.H.; Lakatos, P.L.; Loftus, E.V.; Tysk, C.; O’Morain, C.; Moum, B.; Colombel, J.-F. Geographical variability and environmental risk factors in inflammatory bowel disease. Gut 2013, 62, 630–649. [Google Scholar] [CrossRef] [PubMed]
  7. Gunasekeera, V.; Mendall, M.A.; Chan, D.; Kumar, D. Treatment of Crohn’s disease with IgG4-gudided exclusion diet: A randomized controlled trial. Dig. Dis. Sci. 2016, 61, 1148–1157. [Google Scholar] [CrossRef] [PubMed]
  8. Nguyen, D.L.; Limketkai, B.; Medici, V.; Saire, M.M.; Palmer, L.; Bechtold, M. Nutritional strategies in the management of adult patients with inflammatory bowel disease: Dietary considerations from active disease to disease remission. Curr. Gastroenterol. Rep. 2016, 18, 55. [Google Scholar] [CrossRef] [PubMed]
  9. Li, F.; Liu, X.Q.; Wang, W.J.; Zhang, D. Consumption of vegetables and fruit and the risk of inflammatory bowel disease: A meta-analysis. Eur. J. Gastroenterol. Hepatol. 2015, 27, 623–630. [Google Scholar] [CrossRef] [PubMed]
  10. Wang, F.; Lin, X.; Zhao, Q.; Li, J. Fat intake and risk of ulcerative colitis: Systematic review and dose-response meta-analysis of epidemiological studies. J. Gastroenterol. Hepatol. 2017, 32, 19–27. [Google Scholar] [CrossRef] [PubMed]
  11. Wang, F.; Feng, J.R.; Gao, Q.; Ma, M.X.; Lin, X.; Liu, J.; Li, J.; Zhao, Q. Carbohydrate and protein intake and risk of ulcerative colitis: Systematic review and dose-response meta-analysis of epidemiological studies. Clin. Nutr. 2016. [Google Scholar] [CrossRef] [PubMed]
  12. Wells, G.A.; Shea, B.; O’Connell, D.; Tugwell, P. The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-Analyses. 2011. Available online: http://www.ohri.ca (accessed on 22 April 2017).
  13. Xu, C.; Han, F.F.; Zeng, X.T.; Liu, T.Z.; Li, S.; Gao, Z.Y. Fat intake is not linked to prostate cancer: A systematic review and dose-response meta-analysis. PLoS ONE 2015, 10, e0131747. [Google Scholar] [CrossRef] [PubMed]
  14. Bekkering, G.E.; Harris, R.J.; Thomas, S.; Mayer, A.M.; Beynon, R.; Ness, A.R.; Harbord, R.M.; Bain, C.; Smith, G.D.; Sterne, J.A.C. How much of the data published in observational studies of the association between diet and prostate or bladder cancer is usable for meta-analysis? Am. J. Epidemiol. 2008, 167, 1017–1026. [Google Scholar] [CrossRef] [PubMed]
  15. Reif, S.; Klein, I.; Lubin, F.; Farbstein, M.; Hallak, A.; Gilat, T. Pre-illness dietary factors in inflammatory bowel disease. Gut 1997, 40, 754–760. [Google Scholar] [CrossRef] [PubMed]
  16. Orsini, N. From floated to conventional confidence intervals for the relative risks based on published dose-response data. Comput. Methods Progr. Biomed. 2010, 98, 90–93. [Google Scholar] [CrossRef] [PubMed]
  17. Jantchou, P.; Morois, S.; Clavel-Chapelon, F.; Boutron-Ruault, M.C.; Carbonnel, F. Animal protein intake and risk of inflammatory bowel disease: The E3N prospective study. Am. J. Gastroenterol. 2010, 105, 2195–2201. [Google Scholar] [CrossRef] [PubMed]
  18. Chan, S.S.; Luben, R.; van Schaik, F.; Oldenburg, B.; Bueno-de-Mesquita, H.B.; Hallmans, G.; Karling, P.; Lindgren, S.; Grip, O.; Key, T.D.; et al. Carbohydrate intake in the etiology of Crohn’s disease and ulcerative colitis. Inflamm. Bowel Dis. 2014, 20, 2013–2021. [Google Scholar] [CrossRef] [PubMed]
  19. Greenland, S.; Longnecker, M.P. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am. J. Epidemiol. 1992, 135, 1301–1309. [Google Scholar] [CrossRef] [PubMed]
  20. Orsini, N.; Bellocco, R.; Greenland, S. Generalized least squares for trend estimation of summarized dose-response data. Stata J. 2006, 6, 40–57. [Google Scholar]
  21. Harrell, F.E.; Lee, K.L.; Pollock, B.G. Regression models in clinical studies: Determining relationships between predictors and response. J. Natl. Cancer Inst. 1988, 80, 1198–1202. [Google Scholar] [CrossRef] [PubMed]
  22. Orsini, N. Multivariate Dose-Response Meta-Analysis: An Update on glst. In Proceedings of the Nordic and Baltic Users Group Meeting, Stockholm, Sweden, 27 September 2013. [Google Scholar]
  23. Egger, M.; Davey, S.G.; Schneider, M.; Minder, C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997, 315, 629–634. [Google Scholar] [CrossRef] [PubMed]
  24. Higgins, J.P.; Thompson, S.G.; Deeks, J.J.; Altman, D.G. Measuring inconsistency in meta-analysis. BMJ 2003, 327, 557–560. [Google Scholar] [CrossRef] [PubMed]
  25. Persson, P.G.; Ahlbom, A.; Hellers, G. Diet and inflammatory bowel disease: A case-control study. Epidemiology 1992, 3, 47–52. [Google Scholar] [CrossRef] [PubMed]
  26. Sakamoto, N.; Kono, S.; Wakai, K.; Fukuda, Y.; Satomi, M.; Shimoyama, T.; Inaba, Y.; Miyake, Y.; Sasaki, S.; Okamoto, K.; et al. Dietary risk factors for inflammatory bowel disease a multicenter case-control study in Japan. Inflamm. Bowel Dis. 2005, 11, 154–163. [Google Scholar] [CrossRef] [PubMed]
  27. Ananthakrishnan, A.N.; Khalili, H.; Konijeti, G.G.; Higuchi, L.M.; de Silva, P.; Korzenik, J.R.; Fuchs, C.S.; Willett, W.C. A prospective study of long-term intake of dietary fiber and risk Crohn’s disease and ulcerative colitis. Gastroenterology 2013, 145, 970–977. [Google Scholar] [CrossRef] [PubMed]
  28. Ananthakrishnan, A.N.; Khalili, H.; Konijeti, G.G.; Higuchi, L.M.; de Silva, P.; Fuchs, C.S.; Willett, W.C.; Richter, J.M.; Chan, A.T. Long-term intake of dietary fat and risk of ulcerative colitis and Crohn’s disease. Gut 2014, 63, 776–784. [Google Scholar] [CrossRef] [PubMed]
  29. Chan, S.S.; Luben, R.; Olsen, A.; Tjonneland, A.; Kaaks, R.; Lindgren, S.; Grip, O.; Bergmann, M.M.; Boeing, H.; Hallmans, G.; et al. Association between high dietary intake of the n-3 polyunsaturated fatty acid docosahexaenoic acid and reduced risk of Crohn’s disease. Aliment. Pharmacol. Ther. 2014, 39, 834–842. [Google Scholar] [CrossRef] [PubMed]
  30. Russel, M.G.; Engels, L.G.; Muris, J.W.; Limonard, C.B.; Volovics, A.; Brummer, R.-J.M.; Stockbrugger, R.W. ‘Modern life’ in the epidemiology of inflammatory bowel disease: A case-control study with special emphasis on nutritional factors. Eur. J. Gastroenterol. Hepatol. 1998, 10, 243–249. [Google Scholar] [CrossRef] [PubMed]
  31. Racine, A.; Carbonnel, F.; Chan, S.S.; Hart, A.R.; Bueno-de-Mesquita, H.E.; Oldenburg, B.; van Schaik, F.D.M.; Tjønneland, A.; Olsen, A.; Dahm, C.C.; et al. Dietary patterns and risk of inflammatory bowel diseases in Europe: Results from the EPIC study. Inflamm. Bowel Dis. 2016, 22, 345–354. [Google Scholar] [CrossRef] [PubMed]
  32. Burisch, J.; Pedersen, N.; Cukovic-Cavka, S.; Turk, N.; Kaimakliotis, I.; Duricova, D.; Bortlik, M.; Shonová, O.; Vind, I.; Avnstrøm, S.; et al. Environmental factors in a population-based inception cohort of inflammatory bowel disease patients in Europe—An ECCO-EpiCom study. J. Crohn’s Colitis 2014, 8, 607–616. [Google Scholar] [CrossRef] [PubMed]
  33. Neuman, M.G.; Nanau, R.M. Single-nucleotide polymorphisms in inflammatory bowel disease. Transl. Res. 2012, 160, 45–64. [Google Scholar] [CrossRef] [PubMed]
  34. Liu, X.; Wu, Y.; Li, F.; Zhang, D. Dietary fiber intake reduces risk of inflammatory bowel disease: Result from a meta-analysis. Nutr. Res. 2015, 35, 753–758. [Google Scholar] [CrossRef] [PubMed]
  35. To, N.; Gracie, D.J.; Ford, A.C. Systematic review with meta-analysis: The adverse effects of tobacco smoking on the natural history of Crohn’s disease. Aliment. Pharmacol. Ther. 2016, 43, 549–561. [Google Scholar] [CrossRef] [PubMed]
  36. Collins, K.H.; Hart, D.A.; Reimer, R.A.; Seerattan, R.A.; Banker, C.W.; Sibole, S.C.; Herzog, W. High-fat high sucrose diet leads to dynamic structural and inflammatory alteration in the rat vastus lateralis muscle. J. Orthop. Res. 2016, 34, 2069–2078. [Google Scholar] [CrossRef] [PubMed]
  37. Kang, I.; Espin, J.C.; Carr, T.P.; Tomas-Barberan, F.A.; Chung, S. Raspberry seed flour attenuates high-sucrose diet-mediated hepatic stress and adipose tissue inflammation. J. Nutr. Biochem. 2016, 32, 64–72. [Google Scholar] [CrossRef] [PubMed]
  38. Beilharz, J.E.; Maniam, J.; Morris, M.J. Short-term exposure to a diet high in fat and sugar, or liquid sugar, selectively impairs hippocampal-dependent memory, with differential impacts on inflammation. Behav. Brain Res. 2016, 306, 1–7. [Google Scholar] [CrossRef] [PubMed]
  39. Hosomi, S.; Kaser, A.; Blumberg, R.S. Role of endoplasmic reticulum stress and autophagy as interlinking pathways in the pathogenesis of inflammatory bowel disease. Curr. Opin. Gastroenterol. 2015, 31, 81–88. [Google Scholar] [CrossRef] [PubMed]
  40. Gjymishka, A.; Coman, R.M.; Brusko, T.M.; Glover, S.C. Influence of host immunoregulatory genes, ER stress and gut microbiota on the shared pathogenesis of inflammatory bowel disease and type 1 diabetes. Immunotherapy 2013, 5, 1357–1366. [Google Scholar] [CrossRef] [PubMed]
  41. Orsini, N.; Li, R.; Wolk, A.; Khudyakov, P.; Spiegelman, D. Meta-analysis for linear and nonlinear dose-response relations: Examples and evaluation of approximations and software. Am. J. Epidemiol. 2012, 175, 66–73. [Google Scholar] [CrossRef] [PubMed]
Figure 1. Forest plots (random-effect model) of meta-analyses on the association between carbohydrate, fiber, fat and protein intake (per 10 g increment/day) and the risk of Crohn’s disease.
Figure 1. Forest plots (random-effect model) of meta-analyses on the association between carbohydrate, fiber, fat and protein intake (per 10 g increment/day) and the risk of Crohn’s disease.
Nutrients 09 00500 g001
Figure 2. Non-linear dose–response analysis of fiber and sucrose intake and the risk of Crohn’s disease.
Figure 2. Non-linear dose–response analysis of fiber and sucrose intake and the risk of Crohn’s disease.
Nutrients 09 00500 g002
Table 1. Characteristics of included studies.
Table 1. Characteristics of included studies.
First Author, Year, AreaStudy DesignDiagnostic CriteriaCases/Controls (Age)Time at Diagnosis (Retrospective Period #)Exposure Categories (Dietary Assessment)Risk Estimates (95% CI)Adjusted Factors
Persson, 1992, Sweden (for men) [25]Population-based case-controlThe scoring table suggested by Lennard-Jones63/147 (15–79 years)Within 4 years (5 years ago)T3 vs. T1Relative riskAge, energy intake
Protein2.2 (0.7–6.9)
Carbohydrate2.1 (0.5–8.1)
Fat1.3 (0.4–4.4)
Fiber1.2 (0.5–2.6)
(Validated FFQ)-
Persson, 1992, Sweden (for women) [25]Population-based case-controlThe scoring table suggested by Lennard-Jones89/158 (15–79 years)Within 4 years (5 years ago)T3 vs. T1Relative riskAge, energy intake
Protein0.4 (0.2–1.3)
Carbohydrate1.0 (0.2–4.3)
Fat0.7 (0.2–2.9)
Fiber0.4 (0.2–1.0)
(Validated FFQ)-
Reif, 1997, Israel [15]Population/clinic-based case-control-33/144 (mean, 29.12/29.45 years)Within 1 year from onset of symptoms (before the illness and symptoms began)T3 vs. T1Odds ratioAge, sex, country of origin, residential neighborhood, energy intake
Fiber0.40 (0.10–1.65)
(Validated FFQ)-
Sakamoto, 2005, Japan [26]Hospital-based case-controlThe criteria of the Research Committee on Inflammatory bowel disease in Japan126/211 (15–34 years)Within the past 3 years (5 years before the time of the study)Q4 vs. Q1Odds ratioAge, sex, study area, education, smoking habits
Protein2.06 (0.99–4.28)
Fat2.86 (1.39–5.90)
Carbohydrate0.53 (0.27–1.03)
Fiber0.90 (0.43–1.86)
(Validated FFQ)-
Jantchou, 2010, France (for women) [17]Prospective cohort studyClinical, radiological, endoscopic and histological criteria30/67, 504 (mean, 50.9/52.8 years)Within a median of 54.5 months (a mean follow up of 10.4 years)T3 vs. T1Hazard ratioBody weight, energy intake
Protein3.34 (0.90–12.4)
Carbohydrate1.31 (0.42–4.14)
Fat0.98 (0.25–3.88)
(Validated FFQ)-
Ananthakrishnan, 2013, USA (for female registered nurses) [27]Prospective cohort studyTypical symptoms ≥ 4 weeks; endoscopy; histology; radiography269/170, 169 (NHS I: 30–55 years; NHS II: 25–42 years)With a median age of 54 years at diagnosis (NHS I from 1984 to 2006; NHS II from 1991 to 2007)Q5 vs. Q1Hazard ratioAge, cohort, smoking, BMI, oral contraceptive use, use of post menopausal hormone therapy, regular use of NSAIDs, regular use of aspirin, energy intake
Fiber0.59 (0.39–0.90)
(Validated FFQ)-
Ananthakrishnan, 2014, USA (for female registered nurses) [28]Prospective cohort studyTypical symptoms ≥ 4 weeks; endoscopy; histology; radiography269/170, 169 (NHS I: 30–55 years; NHS II: 25–42 years)With a median age of 54 years at diagnosis (NHS I from 1884 to 2006; NHS II from 1991 to 2007)Q5 vs. Q1Hazard ratioAge, cohort, smoking, BMI, oral contraceptive use, use of post menopausal hormone therapy, regular use of NSAIDs, regular use of aspirin, energy intake
Fat0.98 (0.66–1.45)
(Validated FFQ)-
Chan, 2014, Europe [18]Prospective cohort studyRadiology; endoscopy; histology110/440 (50.1 years/50.1 years)More than 18 months after recruitment (from 1991–1998 to 2004–2010)Q5 vs. Q1Odds ratioAge, sex, center, recruitment date, follow-up period, energy intake, BMI, metabolic rate, physical activity, smoking
Carbohydrate0.87 (0.24–3.12)
(Validated FFQ)-
Chan, 2014, Europe [29]Prospective cohort studyFollow-up questionnaire, in-patient record, histology database, medical note73/292 (50.5 years/50.2 years)More than 18 months after recruitment (from 1991–1998 to 2004)Q5 vs. Q1Odds ratioAge, sex, center, recruitment date, smoking, total energy intake, BMI, dietary vitamin D and relevant fatty acids
Fat1.42 (0.26–7.67)
(Validated FFQ)-
# Retrospective period in case-control studies, and follow-up period in prospective cohort studies; T, tertile; Q, quartile; BMI, body mass index; FFQ, food frequency questionnaire.
Table 2. Intake of the nutrients’ subtypes (per 10 g increment/day) and the risk of Crohn’s disease.
Table 2. Intake of the nutrients’ subtypes (per 10 g increment/day) and the risk of Crohn’s disease.
SubtypesIncluded StudiesRR (95% CI)I2 (%)
SugarReif 1997 [15]; Chan 2014 [18]0.998 (0.969–1.027)0.0
MonosaccharidePersson 1992 (men) [25]; Persson 1992 (women) [25]0.971 (0.715–1.317)49.9
FructoseReif 1997 [15]0.843 (0.695–1.023)-
DisaccharidePersson 1992 (men) [25]; Persson 1992 (women) [25]0.988 (0.871–1.121)0.0
SucrosePersson 1992 (men) [25]; Persson 1992 (women) [25]; Reif 1997 [15]1.088 (1.020–1.160)0.0
StarchChan 2014 [18]0.994 (0.946–1.044)-
Fat
SFASakamoto 2005 [26]; Ananthakrishnan 2014 [28]0.980 (0.843–1.140)17.2
MUFASakamoto 2005 [26]; Ananthakrishnan 2014 [28]1.137 (0.842–1.536)78.8
Oleic acidAnanthakrishnan 2014 [28]; Chan 2014 [29]1.015 (0.900–1.144)0.0
PUFASakamoto 2005 [26]; Ananthakrishnan 2014 [28]1.306 (0.816–2.092)76.2
Arachidonic acidAnanthakrishnan 2014 [28]0.000 (0.000–721.226)-
Linoleic acidAnanthakrishnan 2014 [28]; Chan 2014 [29]1.097 (0.871–1.383)0.0
α–linoleic acidChan 2014 [29]0.035 (0.000–3.299)-
DHAChan 2014 [29]0.004 (0.000–1706.027) #-
EPAChan 2014 [29]799.371 (0.000–2.36 × 1011) #-
Protein
Animal proteinJantchou 2010 [17]2.700 (0.690–10.520) *-
Vegetable proteinJantchou 2010 [17]1.040 (0.280–3.800) *-
# Based on the data adjusted by smoking and total energy intake; * The highest vs. the lowest category (the mean intake per category was unavailable); SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid.
Table 3. Subgroup analyses (random effect model) of carbohydrate, fiber, fat and protein intake (per 10 g increment/day) with the risk of Crohn’s disease.
Table 3. Subgroup analyses (random effect model) of carbohydrate, fiber, fat and protein intake (per 10 g increment/day) with the risk of Crohn’s disease.
SubgroupCarbohydrateFibreFatProtein
RR (95% CI)I2 (%)RR (95% CI)I2 (%)RR (95% CI)I2 (%)RR (95% CI)I2 (%)
Study design
Case-control0.991 (0.974–1.008)19.50.815 (0.679–0.980)0.01.026 (0.930–1.132)70.11.008 (0.922–1.101)57.2
Prospective-cohort0.997 (0.969–1.026)-0.877 (0.761–1.012)-1.005 (0.965–1.048)0.01.099 (0.989–1.221)-
Cohort
Caucasian0.999 (0.981–1.018)0.00.844 (0.751–0.947)0.00.997 (0.963–1.033)0.01.015 (0.915–1.126)63.8
Asian0.983 (0.965–1.001)-1.037 (0.644–1.668)-1.134 (1.030–1.249)-1.066 (0.981–1.159)-
Adjusted for smoking
Yes0.987 (0.972–1.002)0.00.890 (0.776–1.020)0.01.045 (0.970–1.127)62.51.015 (0.915–1.126)63.8
No1.001 (0.975–1.028)14.40.782 (0.641–0.954)0.00.977 (0.916–1.043)0.01.066 (0.981–1.159)-

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Back to TopTop