Differences in dietary habits between patients with inflammatory bowel disease in clinical remission and a healthy population

Introduction

Inflammatory bowel disease (IBD) comprises a distinct group of chronic inflammatory disorders of the alimentary tract with a remitting clinical behavior. Ulcerative colitis (UC) and Crohn’s disease (CD) are the most common conditions in this spectrum [1]. The IBD pathogenesis and activity include multiple interactions among predisposing factors, such as genetics, infections, composition of the intestinal microbiota, and diet [2]. The dietary pattern has been mentioned as a possible environmental trigger for IBD. A Japanese study has demonstrated that the incidence of IBD paralleled the socio-cultural dietetic transition from the traditional to a Western diet [3]. In fact, a high-calorie, high-fat, and high-sugar diet could be a risk factor for IBD development [4]. Moreover, an unhealthy diet could alter the composition of the human microbiome. Indeed, a diet rich in milk fats may promote the growth of Bilophila wadsworthia, a sulfite-reducing organism reported to induce an enhanced inflammatory immune response of the colon in a murine model, resulting in increased levels of interleukin-10 and other proinflammatory cytokines [5,6].

On the other hand, IBD activity may affect dietary restrictions in patients with severe disease. Indeed, it has been shown that these subjects are likely to avoid consuming processed meats, alcohol and salad [7]. In a previous evaluation of UC patients in both the mild and the severe activity phase, Walton et al [8] found that these subjects had a reduced intake of fibers and an increased intake of lipids. Moreover, an appropriate diet is recommended in patients with active IBD, since it could be an adjunctive strategy in disease management. For instance, a diet poor in red meat and polyunsaturated fatty acids was associated with a high rate of maintenance of remission, according to Chiba et al [9]. Likewise, a specific carbohydrate diet, i.e., the restriction of grains and refined sugars, may improve clinical and laboratory parameters [10].

Dietary changes in IBD patients are usually suggested by the healthcare team, though self-limitations based on personal experiences and/or information drawn from the media are frequently observed. Taking into account that dietary habits and intake in patients with IBD may often be the result of a combination of information drawn from different sources, we believed that in the present study it could be of interest to investigate a singular aspect, i.e. the dietary habits of asymptomatic IBD patients in pharmacologically-induced clinical remission. These subjects do not require any dietary restriction and should eat like the healthy population.

Patients and methods

Patients and controls

We designed the present study as a cohort study according to the STROBE statement (Supplementary Table 1) [11], enrolling asymptomatic outpatients with IBD in clinical remission referred to our Unit and healthy volunteers as a control group. The study was reviewed and approved after a meeting of all authors in the Gastroenterology Unit of Bari University Hospital Policlinico (Italy). It included only routine checks in subjects seen regularly in the IBD outpatients service of our Unit. As required by Italian law, each patient gave written informed consent for each investigation performed. The control group consisted of 100 healthy volunteers, shown to be healthy on the basis of medical screening, clinical history, physical examination, routine laboratory testing and 12-lead electrocardiogram, and were taking no drugs or observing any self-prescribed dietary restrictions, selected from among the population referred to the gastroenterology general outpatients service. Written informed consent was obtained from each subject.

Table 1 Main demographic characteristics of the control group vs. the inflammatory bowel disease (IBD) group

One hundred fifty IBD patients (84 CD and 66 UC) were enrolled. The IBD diagnosis had been established by a combined assessment of symptoms, endoscopy, histology, and abdominal imaging. Additionally, for each patient we collected data about age, sex, disease clinical activity—partial Mayo and Harvey Bradshaw index (HBI) scores for UC and CD, respectively—current medications and smoking habit. Clinical remission was defined as HBI <5 for CD and partial Mayo score <2 for UC. The exclusion criteria were pregnancy or lactation, a history of alcohol or drug abuse and previous bowel surgery. A further inclusion criterion for both IBD patients and controls was a body mass index between 18 and 30 kg/m2.

Dietary survey

An experienced nutritionist guided the compilation of a food diary for 7 days according to current guidelines [12]. The dietary analysis was carried out using Winfood™ software (Medimatica Srl Unipersonale, C.da S. Giovanni, 75 - 64010 Colonnella (TE), Italy), as described elsewhere [13]. According to Italian law, “within the 4-5 years of their training, nutritionists acquire highly qualified knowledge and skills in the field of basic sciences, prevention and nutrition in general, with different competences for doctors and non-physicians”, as defined by the legislator in 2005 [14].

Results

All 150 IBD patients were in clinical remission; the mean Partial Mayo score was 1.1±0.6 (median 1, range 0-2) and mean HBI was 2.5±1.0 (median 2, range 0-4). Fifty patients were receiving immunomodulator drugs (thiopurines), while the remaining 100 subjects were given biologic treatment (50 infliximab and 50 adalimumab). All IBD patients were taking mesalazine.

In Table 1 we summarize the main demographic and clinical features of the enrolled patients and controls. The mean age of the 100 age- and sex-matched controls was 43.5±15.8 years, comparable to that of the IBD patients (42.8±12.9, P=0.71). The male/female ratio was 1:1.08 in the control group and 1:1.11 in the IBD group (P=0.97), and 7% of controls were smokers, similar to the percentage in the study group (7.3%, P=0.92).

As reported in Table 2, IBD patients showed a higher daily caloric intake than controls and they ate more total lipids and less fibers. The subanalysis according to the type of immunomodulator therapy demonstrated that the intake of macronutrients and fibers was comparable in the two groups (Table 3). In the comparison CD vs. UC, no statistically significant difference was recorded in total calories, proteins, lipids, carbohydrates or fibers (Table 4). IBD patients who were smokers consumed less fibers (9.1±4.0 vs. 12.0±4.6 g/day, P=0.046).

Table 2 Comparison of dietary intake between the controls and the inflammatory bowel disease (IBD) group, using Student’s t-test. Calories are expressed as kcal/day, macronutrients and fibers as g/day

Table 3 Comparison of dietary intake and body mass index (BMI) between patients with inflammatory bowel disease receiving thiopurines or biologics, using Student’s t-test. Calories are expressed as kcal/day, macronutrients and fibers as g/day

Table 4 Analysis of dietary intake according to inflammatory bowel disease subtype—ulcerative colitis (UC) vs. Crohn’s disease (CD)—and body mass index (BMI), using Student’s t-test. Calories are expressed as kcal/day, macronutrients and fibers as g/day

Discussion

The role of diet in both the pathogenesis and the clinical course of IBD has been amply discussed [4]. A western diet, characterized by a very high intake of fats and calories, has been implicated as an environmental factor triggering IBD [15], as demonstrated in second-generation immigrants in Sweden, who developed a comparable IBD incidence to that of the native Swedish population, possibly as a result of dietary changes [15]. In this regard, it has been proposed that a high dietary intake of omega-6 and polyunsaturated fatty acids may increase susceptibility to IBD [16]. The potentially dangerous effect of a high-calorie, fatty diet has been reported even in mouse models [17]. Moreover, an increased lipid intake may represent not only a potential risk factor for intestinal inflammation, but also an atherogenic factor in a condition featuring a well-known increased risk of cardiovascular and endothelial dysfunction [18,19]. Despite such evidence, in the present study we found a high lipid and total calorie intake in the IBD population even if the patients were asymptomatic and in clinical remission. A similar dietary pattern was observed by Walton et al [8] in the active disease phase. Finally, we could not assess the possible consequences of the different macronutrient and fiber intake in IBD patients compared to healthy subjects [20], since we selected only subjects with a satisfactory nutritional condition, and body composition was not considered helpful for our purpose. Two further limitations of the present study should be mentioned. We did not have any data about the dietary pattern before the diagnosis of IBD, since this was a prospective study, therefore we could not ascertain whether the disease onset could have induced a change in dietary habits and macronutrient intake. Additionally, the interview was aimed to estimate macronutrient intake in IBD patients and we did not ask why most patients were observing self-prescribed dietary restrictions. An excess of carbohydrates in the diet has been observed in IBD patients [21]; thus, some authors have hypothesized a possible direct correlation with the onset and activity of the disease [22]. More specifically, refined sugars have been proven to upregulate oxidative stress in mice [23]. Additionally, diets that specify a low content of industrial sugars may improve the IBD course [24]. However, large epidemiologic studies failed to confirm a conceivable link between carbohydrate intake and the risk of IBD onset [25]. In our study, we did not find any difference in carbohydrate consumption between IBD patients and controls. However, an important point for future investigations should be the analysis of carbohydrate subtypes, since short-chain types have been proven to worsen the clinical picture of IBD in several trials [26], while others, such as starch, may have a beneficial effect on mucosal inflammation [27].

A final remark about fiber assumption in IBD is justified. We observed that even IBD patients without clinical manifestations ate less fibers than controls. This finding is in complete agreement with several reports indicating that such patients are likely to avoid fibers, presumably because of worries about a disease complication, despite the fact that these dietary elements have been shown to have a beneficial effect in maintaining the remission [28-30].

In conclusion, the design of our study allows us to depict an interesting scenario of macronutrient and fiber intake in subjects with IBD, even when clinical symptoms are absent and no dietary restrictions are imposed. The most likely explanation for our findings is that their knowledge of the disease induces IBD patients to modify their dietary habits, reducing the intake of fibers and increasing lipids and overall calories. Therefore, the achievement of clinical remission did not parallel a modification of dietary habits; psychological issues [31], such as the fear of sub-occlusive complications or diarrhea, may be involved [32]. Consequently, our results suggest that a dietary imbalance should be taken into account by the healthcare team and adequately corrected, even in IBD subjects who are well. Therefore, we recommend nutritional counseling for IBD patients, aimed at correcting inappropriate dietary patterns and endorsing the importance of fibers (if not contraindicated), since this aspect has not yet been adequately highlighted and could be an interesting topic to explore in the near future.