Diet and Inflammatory Bowel Disease

Review of Patient-Targeted Recommendations

Jason K. Hou; Dale Lee; James Lewisk


Clin Gastroenterol Hepatol. 2014;12(10):1592-1600. 

In This Article

Patient-Targeted Diet Recommendations

Internet Search

We performed an Internet query of dietary recommendations for IBD using 2 separate search engines, Google and Bing, for the following searches: "Crohn's disease diet" and "ulcerative colitis diet." The top 30 hits on each engine were reviewed (Supplementary Appendices 1–4 Each site was assessed for recommendations to include or exclude specific food categories. The food categories evaluated included macronutrients, food groups, cooked versus raw foods, taste classifications, fiber content, specific foods, and beverages. General food recommendations, foods thought to be possible triggers of disease, and dietary recommendations specific to periods of disease flare were aggregated ( Table 1 ). Recommendations for physician and/or dietician consultation were also assessed on each site. Search results leading to sites containing the same Web address stem and dietary recommendations were considered duplicates and were only included once. For CD, the Web search resulted in 47 unique sites and the UC query identified 55 unique sites. Each Web site's recommendations to include or exclude the food categories were noted. If a Web site gave recommendations supporting both inclusion and exclusion of a food category, this was noted as a "conflicting" recommendation. We included food categories with recommendations from 10 or more Web sites included in our search.

Recommendations on vegetables, fruits, and fiber were particularly common. The vegetable food categories included were raw vegetables, cruciferous vegetables, cooked vegetables, and any vegetables. If a recommendation on a particular vegetable was made, but did not specify raw, cooked, or cruciferous, it was included as a recommendation for "any vegetable." Comparisons of dietary recommendations between CD and UC using chi-square testing showed similar rates of recommendations except for fatty and fried foods. Fatty and fried foods were recommended to be excluded in 100% of CD sites compared with 71% of UC sites (P = .02). Given the general similarity, the food recommendations for CD and UC are presented together.

Of the 28 food categories included, avoidance (by ≥80% sites mentioning the food) was recommended for 11 categories: raw vegetables, cruciferous vegetables, citrus fruit, red meat, carbonated beverages, coffee and tea, alcohol, fatty and fried foods, spicy foods, sugars, and seeds and popcorn. Of these categories, cruciferous vegetables, alcohol, carbonated beverages, and sugars were recommended to be avoided by 100% of sites mentioning the food category. Inclusion (by ≥80% sites mentioning the food) was recommended for 5 categories: cooked vegetables, fish, poultry, lean protein, and high-protein diet. Of these categories, cooked vegetables, poultry, and lean protein were recommended to be included by 100% of sites mentioning the food. The food categories having a higher percentage of "conflicting" recommendations for both inclusion and exclusion were any vegetables (21%), any fruit (32%), nuts (17%), and whole grains (18%).

Our Web search analysis demonstrated that patient-targeted dietary recommendations are highly restrictive and frequently conflicting. These recommendations may result in patient confusion and unnecessarily restrictive diets in patients who are already at risk for nutritional deficiencies.

Defined Diets

Defined diets are dietary regimens prescribed based on an underlying "theory" of how food interacts with the body. There are several defined diets that have been touted to affect intestinal inflammation and other medical conditions.[46–50] Defined diets are promoted in the lay literature through anecdotal success stories but to date lack rigorous scientific assessment. In this review we detail 3 defined diets that are commonly advocated for patients with IBD in the lay literature: the specific carbohydrate diet (SCD); the fermentable oligosaccharides, disaccharides, and monosaccharides (FODMAP) diet; and the Paleolithic diet (Paleo). This review is not advocating the use of these diets to treat patients with IBD, but rather highlights the underlying philosophy and potential nutritional impact of these diets on patients with IBD.

Specific Carbohydrate Diet

The SCD was first described by Dr. Sidney Haas in 1924 as a means to treat celiac disease.[48] The SCD was popularized for the treatment of IBD by biochemist Elaine Gottschall through her lay book Breaking the Vicious Cycle after her daughter was reportedly cured of UC using the SCD.[46] The SCD is also promoted in the lay literature to manage other diseases, including celiac disease, constipation, hyperactivity, night terrors, and autism.[46]

The underlying theory of the SCD is that disaccharide and polysaccharide carbohydrates are poorly absorbed in the human intestinal tract, resulting in bacterial and yeast overgrowth and subsequent overproduction of mucus. These effects are hypothesized to result in small bowel injury thus perpetuating the cycle of carbohydrate malabsorption and intestinal injury.[46] Strict adherence to the diet is recommended, because any exposure to restricted carbohydrates is hypothesized to worsen bacterial overgrowth and exacerbate mucosal damage. There are significant variations in diet recommendations within the community of SCD and related diets. Although not as restrictive as the SCD, the gluten-free diet has also been advocated by some patients to treat IBD. For the purposes of this review, we focus on the specific recommendations as described in Breaking the Vicious Cycle.[46]

Dietary Restrictions and Allowances on the Specific Carbohydrate diet. The SCD restricts all but simple carbohydrates ( Table 2 ). The only carbohydrates permitted are monosaccharides: glucose, fructose, and galactose. Fresh fruits and vegetables are universally acceptable with the exception of potatoes and yams. Certain legumes (ie, lentils, split pea) are permitted; however, others (ie, chickpeas, soybeans) are not. No grains are permitted in the SCD. Saccharin and honey are permitted in addition to moderate use of sorbitol and xylitol. Canned fruits and vegetables are not permitted because of possible added sugars and starches.

Unprocessed meats are permitted in the SCD without limitation. However, processed, canned, and most smoked meats are restricted because of possible sugars and starches used in additives. Milk is not permitted in the SCD because of lactose content. However, certain lactose-free cheeses are permitted as is homemade lactose-free yogurt.


The underlying mechanistic theory of the FODMAP diet overlaps with the SCD; poorly absorbed carbohydrates result in bacterial overgrowth.[50] The FODMAP diet has been studied primarily for irritable bowel syndrome and functional gastrointestinal disorders. However, despite the similar mechanistic theories, the SCD and FODMAP diets are diametrically opposed when it comes to honey and many fruits and vegetables ( Table 2 ).[50] Although they are similar in the restrictions of cereal grains and unrestricted meat, the FODMAP diet is highly restrictive on certain fruit and vegetable intake, whereas the SCD has unrestricted fruit and vegetable intake except for potatoes and yams.

Paleolithic Diet

The Paleo diet was introduced by Dr. Walter L. Voegtlin, a gastroenterologist, who published a lay book titled Stone Age Diet: Based on In-Depth Studies on Human Ecology and the Diet of Man.[49] A scientific review of the Paleo diet was published in the New England Journal of Medicine in 1995 further describing the evolutionary rationale for the Paleo diet and contrasting it with the modern diet.[47] The underlying hypothesis behind the Paleo diet is that the human digestive tract is poorly evolved to handle the modern diet that resulted from development of modern agricultural methods. It is hypothesized that exposure of the human digestive tract to foods that were not present at the time of human evolution may result in modern diseases. Because the primary principle behind the Paleo diet is based on assumptions of evolutionary biology, there is no mechanistic theory as to the effect of diet on intestinal inflammation specifically. The lack of mechanistic theory has also led to great variations in the recommended foods and restrictions. There are multiple variations of the Paleo diet published in the lay literature, including the Caveman, Stone-Age, and Hunter-gatherer diets.

Dietary Allowances and Restrictions on the Paleo Diet. The Paleo diet emphasizes intake of lean, nondomesticated (game) meats and noncereal plant-based foods (ie, fruits, roots, legumes, and nuts).[47] The Paleo diet is not as prescriptive as the SCD regarding food types but rather focuses on the source and balance of caloric intake. Lean protein is recommended to be the source of 30%–35% of daily caloric intake. The balance of subtypes of polyunsaturated fatty acids (PUFAs), n-6 and n-3, are recommended to be as low as 2:1, in contrast to estimates of the modern diet ratio of upward of 11:1. Consumption of lean protein from nondomesticated meat is recommended to reach recommended PUFA ratios. The Paleo diet hypothesizes that domesticated livestock raised on grain-based feed has unfavorable fat composition and should be avoided. In addition to lean nondomesticated meats, the Paleo diet advocates a very-high-fiber diet from noncereal-based plant sources, up to 45–100 g/day.[47] There is debate in the Paleo diet community regarding the acceptability of potatoes and legumes in the Paleo diet.

Scientific Evidence Regarding Defined Diets for Inflammatory Bowel Disease

There are no formal published studies on the benefits of either the SCD or Paleo diets in the management or prevention of IBD. Only a few small pilot studies have evaluated the FODMAP diet in patients with IBD. Both the SCD and FODMAP diets purport that carbohydrates lead to bacterial overgrowth. Detecting bacterial overgrowth (ie, an increase in the abundance of bacteria) is difficult and standard tests are fraught with misclassification. Furthermore, dietary composition is correlated with the composition of the gut microbiome as measured in terms of relative abundance. For example, Wu and coworkers[18] demonstrated that consumption of carbohydrates was positively correlated with the relative abundance of most but not all firmicutes within human feces. Similarly, Hoffmann and coworkers[51] found that consumption of carbohydrates is positively correlated with the proportional abundance of Candida and the methanogen Archaea Methanobrevibacter. The link between diet and abundance of certain bacteria, yeast, and Archaea is complex but may represent a syntrophic relationship.[51] For example, Candida may use starch, liberating simple carbohydrates that are used by bacteria, such as Prevotella and Ruminococcus, which in turn produce substrates for fermentation that can be used by Methanobrevibacter to produce CH4 and or CO2.

How bacterial overgrowth may result in intestinal inflammation is unclear. The SCD postulates that bacterial overgrowth results in fermentation and subsequent production of short-chain organic acids that are injurious to the small intestinal mucosa. However, Breaking the Viscous Cycle references only case studies on systemic D-lactic acidosis, not mucosal concentrations of organic acids or mucosal injury.[52–54] The FODMAP authors hypothesize bacterial overgrowth may result in increased intestinal permeability, which has been associated with the pathogenesis of CD.[55,56]

There have been two small pilot studies evaluating the FODMAP diet in IBD.[57,58] The first was performed in 8 patients with UC who had undergone colectomy. Median stool frequency per day dropped from 8 to 4 (P = .02) after initiation of the low-FODMAP diet in the retrospective analyses; however, no benefit was observed in 5 patients who were studied prospectively.[58] In the second study, 72 patients with IBD were retrospectively evaluated after education regarding a low-FODMAP diet. Based on self-report, 70% of patients remained adherent on the diet after 3 months, and symptoms of pain, bloating, and diarrhea improved among those adherent to the diet (P < .02).[57] These limited retrospective studies are supportive of dietary interventions to improve IBD symptoms but may be biased because of their retrospective nature and lack of objective data regarding inflammatory changes associated with dietary intervention. Symptomatic response in patients with IBD to these dietary interventions may also suggest a component of functional gastrointestinal symptoms or nonceliac gluten intolerance.

The existing data on dietary risk factors are not clear regarding the role of carbohydrates in the development of IBD. A systematic review of dietary risk factors for IBD included 5 studies reporting the association of carbohydrate intake and risk of developing IBD showing conflicting results.[59] The two most recent and largest cohort studies showed no association between carbohydrate intake and UC risk.[12,60] None of the included studies specifically differentiated monosaccharides from other carbohydrates, which could limit its applicability to SCD recommendations. However, the existing data do not strongly support the role of carbohydrates in the development of IBD or in perpetuating intestinal inflammation.

The carbohydrate malabsorption/bacterial overgrowth theory does not incorporate observations of increased risks of IBD associated with high protein or high fats. As discussed previously, Chiba and coworkers[43] demonstrated a reduction in disease relapse for patients with CD on a semivegetarian diet compared with a control omnivorous diet. The semivegetarian diet included brown rice, soybeans, seaweed, yam, potato, onion, and corn, foods restricted by both the SCD and FODMAP diets. In a prospective cohort study, Jantchou and coworkers[60] observed a positive association of high animal protein intake with the development of IBD (hazard ratio, 3.01; 95% confidence interval, 1.45–6.34). Similarly, high-fat diets have been associated with an increased risk of development of both CD and UC.[4,13,14,61] Although therapeutic trials of omega-3 PUFA to treat IBD have not proved successful, there are signals that the balance of omega-6:omega-3 PUFA may affect IBD risk.[13,44] The concept of balance between PUFA does correspond conceptually with the Paleo diet; however, means to assess ratios of PUFA in a practical manner in the modern diet are challenging.

Both the SCD and Paleo diets advocate a high-fiber diet but restrict cereal grain–based fiber. Fermentation of dietary fiber in the colon produces short-chain fatty acids, which act as an energy source for colonocytes, modulate the local immune response (attenuate interleukin-6, interleukin-8, and tumor necrosis factor-α), and modify the intestinal microbial flora.[62] Interventions using grain-based fiber have demonstrated potential therapeutic benefits in UC. Hallert and coworkers[63] performed a pilot study of 22 patients with UC in remission, demonstrating that an increase of dietary fiber intake of 60 g of oat bran daily can increase fecal butyrate levels by 36% without an exacerbation in symptoms. Kanauchi and coworkers[64] performed an open-label control trial of 18 patients with mild-to-moderate UC, treated with 20–30 g/day of germinated barley foodstuff. An improvement of bowel-related symptoms benefit was observed (P < .05) in the germinated barley foodstuff-treated group as were increased fecal concentrations of Bifidobacterium and Eubaterium limosum. In contrast to the SCD and Paleo diets, these data suggest that inclusion of cereal-based grains may be beneficial to patients with UC in particular. Further study of the role of cereal grain and noncereal grain based fiber in IBD is required.

Both the SCD and Paleo diets have the potential to contribute to vitamin D deficiency. This is a particular concern given the association of vitamin D deficiency and increased risk of surgery and hospitalization.[65] In patients expressing interest in either of these diets, assessment of vitamin D status may be important.