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

Hypothesis: Increased consumption of emulsifiers as an explanation for the rising incidence of Crohn's disease

Carol L. Roberts, Sarah L. Rushworth, Emile Richman, Jonathan M. Rhodes
DOI: http://dx.doi.org/10.1016/j.crohns.2013.01.004 338-341 First published online: 1 May 2013


Crohn's disease (CD) incidence has increased over the past fifty years but the explanation is unclear. CD can be brought into remission by liquid enteral feeding, but the mechanism for this response is unknown. We suggest that consumption of emulsifiers in processed foods may promote CD by increasing bacterial translocation. This is supported by evidence that (i) geographical variation in CD correlates with emulsifier consumption as does the increasing incidence of CD in Japan; (ii) although CD incidence also correlates with fat consumption, the response to enteral feeding is not affected by the fat content of the feed and (iii) very small concentrations of the emulsifier polysorbate 80 enhance bacterial translocation across intestinal epithelia. Undigested emulsifiers may increase bacterial translocation, particularly in the small intestine where the mucus layer is discontinuous. The hypothesis should be testable by trials of enteral feeding with/without emulsifiers.

  • Crohn's
  • Diet
  • Incidence
  • Pathogenesis
  • Bacterial translocation
  • Polysorbate
  • Emulsifier

1 The relative importance of genetic and environmental factors in CD pathogenesis

One hundred and sixty-three gene loci associated with inflammatory bowel disease have now been identified1 but much the strongest gene association is between CD and NOD2/CARD15, with three independently associated mutations — arg702trp, gly908arg, and leu1007fsinsC. Together these confer a 15%–20% attributable population risk among cases of familial CD, yet individuals who are homozygotes for NOD2/CARD15 mutations still only have approximately 3% lifetime risk of developing the disease and only about 25% of CD patients have a family history, implying a strong role for environmental factors. Polymorphisms in the autophagy genes ATG16L1 and IRGM also associate with CD but have a much smaller attributable risk for pathogenesis. Each of these genes encode proteins that form part of the innate immune response to bacteria, indicating that a reduced ability to defend against intestinal microbiota, together with altered regulation of the immune response, are involved in CD pathogenesis.1 A consequence seems to be invasion of the mucosa by bacteria, particularly by E. coli with an adherent and invasive phenotype.2

2 Increasing incidence of CD

CD incidence in many countries has increased markedly over the past 50 years,3 particularly in Asia with 3–5-fold increases in Japan and Singapore within the 10–14 years around the millennium.4,5 The incidence in the West has also increased and meta-analyses have shown about a 4-fold increase over the past 40 years with a particularly striking rise amongst juvenile-onset CD cases.6

3 Evidence for the role of specific environmental factors

Direct evidence from interventional studies exists only for smoking as a risk factor for CD but there is indirect evidence in support of hygiene in infancy, oral contraceptive use, and various dietary risk factors. There is also growing evidence for important associations with changes in the gut microbiota, particularly an increase in mucosa-associated E. coli and reduction in faecal Faecalibacterium prauznitzii, but there is no knowledge of how other environmental factors affect these components of the microbiota.

3.1 Smoking

There is a strong association between current smoking and CD and cessation of smoking reduces risk of relapse. However the proportion of the UK population that smokes has fallen over the last three decades, a pattern which is mirrored by much of western Europe. Smoking therefore does not account for the increasing incidence of CD.

3.2 Hygiene

Increased hygiene in infancy has been suggested as the explanation for several diseases with increasing incidence including asthma and CD. Increased risk of CD has been reported in individuals who grew up in households with hot running water as compared with those who had only a cold water supply. The hygiene hypothesis seems unlikely though to explain the very marked recent increased incidence of Crohn's disease in Japan.

3.3 Oral contraceptives

Use of oral contraceptives is a modest risk factor for CD that might account for some of the increase in women, but most studies have also shown a substantial increase in CD incidence in males.

3.4 Lack of sunlight

The increased prevalence of a range of immunologically-mediated disorders, including CD, in northern Europe has been attributed to lack of sunlight and its impact on the beneficial effects of vitamin D3.6 Vitamin D3 enhances dendritic cell function, autophagy and expression of the CD associated gene NOD2 and supplementation with vitamin D3 has been shown to have promising therapeutic impact on the course of CD. This would however not explain the marked increase in incidence over time.

3.5 Diet

3.5.1 Associations between diet and Crohn's disease risk

The most consistent epidemiological association between diet and CD has been the high pre-illness intake of refined sugar.7 Exclusion trials have however been negative, whether this has been assessed as the sole intervention or as part of a high fibre, low sugar diet and it seems likely that the high sugar intake may be a correlate of some other harmful dietary factor. It has indeed proved frustratingly difficult to define an exclusion diet that will keep the majority of patients free from relapse after remission has been induced by a period of enteral feeding. Extensive rechallenge studies in a patient with CD showed that foods that are well tolerated during prolonged remission may not be tolerated in relapse, when the mucosa is presumably ulcerated.8

Fat consumption has been increasing worldwide over the past three decades, in both developed and undeveloped countries. The rapid recent rise in CD in Japan has paralleled the introduction and uptake of a western style diet and an increase in consumption of total fat, animal fat and n-6 polyunsaturated fatty acids (r for all these > 0.8) with animal fat the strongest independent factor on multivariate analysis.9,10 However meta-analysis of trials comparing high fat and low fat enteral feeds has failed to show any difference in therapeutic response and has also failed to show difference in response according to n-3 versus n-6 fat content.11

3.5.2 Induction of remission in Crohn's disease by use of liquid enteral feeding

There is good evidence supporting the benefits of enteral nutritional feeding with elemental or whole protein (polymeric) feed as either primary or adjuvant therapy for CD.11 It has been suggested that enteral feeding directly affects intestinal inflammation by regulating gene expression in the gut epithelium as well as by modulating intestinal inflammatory and immune mediators production. Equally plausible is a quantitative effect of a very low residue intake on the intestinal microbiota.

3.5.3 Perhaps it is emulsifier intake rather than fat that increases risk for Crohn's disease? Geographical variation in Crohn's disease incidence and emulsifier consumption

We have looked at available country by country data for emulsifier consumption in food and beverages, including polysorbates and sorbate esters, and lecithin.12 When this is compared with CD incidence, where known, on a matched year basis, a strong positive correlation is found (P = 0.017 by Spearman Rank correlation; Fig. 1). In Japan where data for CD incidence and emulsifier consumption were available for the same years there was again a positive correlation (Fig. 2).

Figure 1

Crohn's disease incidence by country versus emulsifier consumption in food and beverages per capita per year, taking the most recent year in which Crohn's incidence and emulsifier data were both available. [Incidence1820; emulsifier intake12].

Figure 2

CD incidence vs emulsifier intake for Japan, 1992–1998 [Incidence5; emulsifier intake12].

There is also evidence that frequent fastfood intake, likely to contain emulsifiers, confers a 3- to 4-fold greater risk for developing CD.13 Consumption of margarine, which contains both hydrogenated fats and emulsifiers, has been positively correlated with UC and CD in most, but not all, studies.14

3.5.4 Possible mechanisms to explain effects of fat and emulsifiers in CD pathogenesis

Most emulsifiers are broken down on passage through the small intestine and their detergent effects in the distal ileum and colon may arguably be small compared with the natural effects of bile acids. The commonly used emulsifier, lecithin (phosphatidyl choline), may even enhance intestinal barrier function. Other emulsifiers have however been shown to increase intestinal permeability in concentrations that could plausibly occur in the distal small intestine in man. The emulsifiers polysorbate-60 and polysorbate-80 cause lysosomal enzyme release from intestinal mucosal cells and concentration-dependent effects of non-ionic surfactants on cell permeability and cell viability have been shown.15 We have recently shown that polysorbate-80 at a concentration (0.01% vol/vol) that could be expected in the terminal ileum if just 6.7% of a recommended acceptable daily intake passed through without digestion caused a 59-fold increase in translocation of E. coli across intestinal epithelial cells.16

3.5.5 An overarching hypothesis for Crohn's disease

We have discussed the growing evidence that CD is associated with genetic and functional alterations that point towards defective clearance of bacteria. This is supported by growing evidence for intramucosal bacteria, particularly E. coli with an adherent and invasive phenotype. The earliest lesions of Crohn's disease seem to overlie Peyer's patches and there is a good correlation between the peak ages of onset of CD (in adolescents and young adults) and the greater number of Peyer's patches at that age. Peyer's patches occur mainly in the distal small intestine and are the sites where bacterial translocation is normally easiest, and increased in CD.17 Translocation occurs through the specialised microfold (M) epithelial cells that account for about 5% of the cells that make up the dome epithelium that overlies the Peyer's patches moreover the distal ileum lacks a continuous mucus layer and is particularly vulnerable to interaction with bacteria that may exist as a result of “backwash” through the ileo-caecal valve. It seems highly plausible that presence of emulsifiers in the distal ileal lumen, may add considerably to the risk of bacterial translocation (Fig. 3). Similar translocation may also occur at the lymphoid follicles in the colon which are similar, albeit smaller, to Peyer's patches.

Figure 3

Suggested model for gene–environment interactions in small intestinal Crohn's disease. The mechanism for the harmful effect of smoking on Crohn's is not well understood but there are some data suggesting impairment of phagocyte function.

4 Implications for future studies

Retrospective analysis of the effect of emulsifier content on response to enteral nutrition has not proved feasible because of unpublished variations over time in the constituents of the feeds used. Further studies are therefore needed to allow direct comparison between the therapeutic response of Crohn's disease to polymeric enteral feeds with and without permitted food emulsifiers.

Conflict of interest statement

Prof Rhodes is a member of advisory boards for Atlantic, Procter and Gamble and Falk, has received speaking honoraria from Abbott, Falk, Ferring, Glaxo Smith Kline, Procter and Gamble, Schering Plough, Shire and Wyeth, and with the University of Liverpool and Provexis UK, holds a patent for use of a soluble fibre preparation as maintenance therapy for Crohn's disease plus a patent pending for its use in antibiotic-associated diarrhoea.

Competing interests

JMR together with the University of Liverpool and Provexis UK, holds a patent for the use of a soluble fibre preparation as maintenance therapy for Crohn's disease.


We are very grateful to Dr Graham Horgan, Biomathematics and Statistics Scotland, Rowett Institute of Nutrition and Health, University of Aberdeen, Scotland for the statistical advice.

Proctor and Gamble UK kindly gave us an unrestricted grant to fund the acquisition of market research data on global emulsifier consumption.


View Abstract