Systematic Review With Meta-analysis

Dietary Intake in Adults With Inflammatory Bowel Disease

Kelly Lambert; Daniel Pappas; Chiara Miglioretto; Arefeh Javadpour; Hannah Reveley; Laura Frank; Michael C. Grimm; Dorit Samocha-Bonet; Georgina L. Hold


Aliment Pharmacol Ther. 2021;54(6):742-754. 

In This Article


A total of 8056 records were identified in the search. After exclusion of duplicates, 4835 records were screened, and 129 full-text articles were reviewed for eligibility (Figure S1). Twenty-four records were excluded as they were not contemporaneous and published prior to 2000 and 65 studies did not report useable nutrient or dietary data. A total of 40 studies thus were included in the final qualitative synthesis and 19 studies in the meta-analysis.

Characteristics of the included studies are shown in Table S2. Studies on adults with IBD were conducted in 18 countries with the majority in Japan (n = 6 studies,[27,28,35–38] 550 participants); followed by the UK and Ireland (five studies,[22,33,39–41] 486 participants) and Italy (four studies,[29,42–44] 538 participants). Of the 40 included studies most were case–control studies (19 studies, 48%) followed by cross-sectional studies (18 studies, 45%), and three cohort studies (8%).

Overall, there were 4070 participants with IBD (ulcerative colitis: 1867; Crohn's disease: 2093; 110 with IBD (unspecified type) and 11 504 controls (Table S2). The average age of participants with IBD ranged from 30[45] to 47[29] years. Studies reported an age range of participants with IBD from 18 to 81 years. The distribution of gender in study participants with IBD ranged from 0% males[46] to 100% male.[47] Of the studies with control participants, most (n = 12/19 studies) were matched for age and gender. Control participants were either healthy volunteers free from IBD in cohort studies,[23,43,48–50] or were recruited from outpatient clinics,[27,29,32] hospitals,[25,29,51] hospital staff or relatives,[29,52] or were volunteers.[22,35] The mean BMI of participants with IBD ranged from 17.4[51] to 27.4[53] kg/m2. Nine studies did not report any anthropometric data. Smoking status was reported in 16 studies, and about one-third of participants with Crohn's disease were smokers[40,45,48,54–57] compared to less than one-fifth of adults with ulcerative colitis.[26,45,48,53,58] Few studies reported the educational level of participants. Of the seven studies that reported the education level, this ranged from 39.6% to 80.8% of participants possessing a college level of education.[26,59]

The methods used to obtain information on usual dietary intake were highly variable (Table S2). The most frequently used tool was a 3-day food record (eight studies[42,47,52,53,55,60–62]), followed by a 7-day food record (five studies[32,40,44,50,54]). Various food frequency questionnaires were used which captured food frequency data over 24 hours (six studies[30,33,41,51,56,61]), 1 week (2 studies[27,28]), 1 month (three studies[23,36,37]), 3 months (one study[58]) and 1 year (two studies[26,63]) periods and ranged from 86,[57] 99,[43] 110,[23,48] 150,[38] 18[59] or 630 food items.[22] Three studies used multiple methods of a food recall and food frequency questionnaire.[32,51,62] One study[57] included information on dietary vitamin and mineral supplement use, but excluded intake from these supplements from the analysis reported.

Disease activity was also defined in a wide variety of ways. For adults with Crohn's disease the most common method for reporting disease activity was the CDAI (12 studies[37,38,42,45,51,54,55,58,59,61,63]) or the Truelove Witts score for adults with Ulcerative Colitis (four studies[28,36,45,52]). Other methods for reporting disease activity were the Harvey Bradshaw index (seven studies[22,33,48,52,56,57,60]); the Mayo score (four studies ulcerative colitis patients[44,58,59,63]); Simple Clinical Colitis Activity Index (two studies ulcerative colitis patients[50,62]). Other measures such as C Reactive Protein,[33,45] faecal calprotectin,[33,48] Ulcerative Colitis Disease Activity Index,[28] the Montreal classification[29] or self-report[39–41] were used. Disease activity was not reported in seven studies.

Information on the extent of disease and surgical history was frequently omitted. Of the 40 studies, 15 studies provided no information on disease extent and 23 studies gave no information on the surgical history. Of those where information was available, 20 studies included specific detail regarding the disease activity and participants were included with widespread involvement in the small intestine, large intestine or both. Information on other surgical procedures was challenging to synthesise due to heterogenous reporting of information. Six studies excluded participants with previous surgery leaving them with short bowel syndrome.[23,31,37,42,56,62] One study excluded participants with a previous terminal ileal resection,[39] one excluded those with any history of a small bowel resection,[49] and two excluded participants with a history of any gastrointestinal surgery.[44,59]

Usual Dietary Intake

Macronutrients. Table 2 contains a summary of the usual intake of macronutrients for adults with IBD with dietary reference values included for comparison. Overall, the daily energy intake of adults with IBD did not meet minimum requirements when reported collectively as "IBD" or as specific subgroups, with adults with active ulcerative colitis consuming the lowest energy 1454 ± 511 kcal/day. Protein intake met or exceeded recommended amounts with males and those with Crohn's disease in remission consuming the largest amounts of protein per day (82 ± 16, and 100 ± 42 g/day, respectively). Those with active Crohn's disease consumed less protein compared to those with Crohn's disease in remission. Total carbohydrate intake ranged from 210 ± 31 g/day in those with active disease to 303 ± 34 g/day in males. Daily fat intake was lowest in those with active ulcerative colitis (55 ± 11 g/day) and females (58 ± 13 g/day). Fibre intake was inadequate in all groups and was far below the Adequate Intake amount of 25–30 g/day. The lowest fibre intake was in adults with ulcerative colitis (12 ± 5 g/day) and highest in males (20 ± 5 g/day). Those in remission consumed slightly more fibre (16 ± 6 g/day) than those with active disease (14 ± 1 g/day). The percentage of energy intake from carbohydrate and fat were consistent with recommended amounts in all groups. The contribution of protein to overall daily energy intake was variable, and below the recommended amounts for the pooled IBD, ulcerative colitis, Crohn's disease, remission and active groups.

Vitamin Intake

Fat-soluble Vitamins. Data for the usual intake of vitamins is shown in Table 3. Data on vitamin A were missing for females with IBD. For the remaining subgroups, intake was inadequate for the pooled IBD group, ulcerative colitis, Crohn's disease, Crohn's disease active disease and Crohn's disease remission subgroups. Usual vitamin D intake was inadequate or borderline for most subgroups and did not meet the recommended requirement in the pooled active IBD group, pooled males and pooled females with IBD groups. Mean intake of vitamin E varied widely and was inadequate for the pooled Crohn's disease remission, Crohn's disease active disease and pooled active IBD groups. Data for usual vitamin K intake was sparse and mostly obtained from one study. Mean intake was inadequate for vitamin K for Crohn's disease, both genders and active disease groups.

Water-soluble Vitamins. The usual reported vitamin C intake was sufficient and exceeded the nutrient reference value of 45 mg/day for all subgroups. Adults with Crohn's disease did not meet the nutrient reference values for Vitamin B1, B2, B3, B6 or folate. For adults with active Crohn's disease, intake was inadequate for all water-soluble vitamins. Folate intake was particularly problematic for those with Crohn's disease, with intake of 120, 144 and 153 μg/day in the pooled, remission and active Crohn's disease groups, respectively. This represents less than 50% of the recommended dietary intake of 400 μg/day. In contrast, those with active ulcerative colitis, consumed adequate amounts of water-soluble vitamins, except folate and B12. No data were available for B5 or biotin for any subgroup other than those with ulcerative colitis, where intake was adequate. Intake of vitamin B12 was adequate for all groups except those with active ulcerative colitis and Crohn's disease. Pooled female data indicated inadequate B1 and B3 consumption and adequate B2, B6 and B12.

Mineral Intake

The usual dietary intake of minerals is shown in Table 4. Overall, there is scarce data on dietary mineral intake and detail on usual intake is missing for several groups or based on only one study. Calcium intake was inadequate for all groups and dietary sodium intake was excessive for all groups. When compared to the upper limit of adequate intake for sodium (920 mg/day), intake was as much as five times greater (the overall pooled IBD group having the highest intake of 4268 mg/day). Sodium intake ranged from 1154 to 4268 mg/day. Intake of magnesium was inadequate for all groups except the ulcerative colitis remission group. Dietary iron intake was adequate for all groups or where gender was not reported met the lower end of the female reference value. The usual intake for females was inadequate for all minerals except iron and sodium. For those with Crohn's disease in remission, data were scant. Similarly, data on intake of copper, selenium, iodine, manganese and molybdenum were frequently unavailable for many subgroups.

Food Group Intake

Quantitative data on the usual intake of food groups was obtained from five studies[48,58,59,61,64] but for some food groups from a single study. Details are shown in Table 5. None of the six subgroups of adults with IBD achieved adequate intake of legumes, fruit, vegetables or dairy. Intake of meat, fish and seafood generally met or exceeded recommend amounts. Added sugar intake was variable and did not exceed recommended amounts. There was a notable absence of food group data for females, those with active Crohn's disease, and those with Crohn's disease in remission.

Meta-analysis of Usual Intake in IBD Groups Compared to Healthy Controls

Table 6 summarises the results of the qualitative synthesis of usual intake data for adults with IBD, ulcerative colitis, Crohn's disease and healthy controls. Fibre intake was significantly lower in adults with IBD compared with healthy controls (SMD, −0.59 [95% CI: −0.73, −0.46]). This equates to a mean difference in fibre intake of −3.95 g/day (95% CI: −4.8, −3.1 g/day, data not shown) in individuals with IBD compared to healthy controls. There were no significant differences between adults with IBD and healthy controls for energy, protein, fat, carbohydrate, calcium or iron intake. The Forest plots for these analyses are shown in Figure S2.

Meta-analysis of the dietary data of adults with ulcerative colitis compared to healthy controls are also shown in Table 6. One outlier study was identified that contributed substantially to the heterogeneity in each analysis for vitamin C,[32] fat[44] and fibre.[32] Analyses without the outlier study reduced heterogeneity but did not change the outcomes, therefore analyses are reported with all data available. Dietary copper intake was significantly lower in adults with UC compared to controls (SMD, 0.39 [95% CI: 0.15, 0.62]). Fat intake was also significantly lower in ulcerative colitis compared to controls (SMD, 0.51 [95% CI: 0.16, 0.85]). This is equivalent to a daily intake 12.8 g/day lower (95% CI: −5.0, −20.6 g/day, data not shown) in UC compared with controls. The contribution of carbohydrate to total energy was significantly lower in those with UC versus healthy controls (SMD, 1.05 [95% CI: 0.24, 1.86]), There were no significant differences between adults with ulcerative colitis and controls for protein, carbohydrate, the proportion of energy from fat or protein, vitamins A, D, E, B1, B2, B3, B12, folate, magnesium, iron, zinc and manganese.

Analysis of dietary intake data for individuals with Crohn's disease compared to healthy controls indicated several important differences. These included significantly lower intake of protein and iron in individuals with CD compared to healthy controls (Table 6). Fibre intake was also significantly lower in individuals with CD compared to healthy controls (SMD, −1.38 [95% CI: −2.43, −0.33]). This is equivalent to a daily intake −4.0 g/day (95% CI: −4.8, −3.3) lower compared with healthy controls. Examination of the results for the SMD indicated there were significant differences between CD and healthy controls for carbohydrate, percentage of energy from carbohydrate and Vitamin C. However, values for the mean difference for these nutrients were no longer significant as the confidence interval included zero. Intake of energy, other macronutrients, the proportion of energy from fat and protein and calcium did not differ from controls.

Quality of Studies

Assessment of the quality of included studies is shown in Table S3. Inter-rater agreement was considered good (Cohen's kappa 0.63). Evaluation of the risk of bias was rated as positive in 31 studies (78%), neutral in eight studies (20%), and one as negative. Of the neutral studies, information on selection of study participants, handling of study withdrawals and descriptions of the study settings were the most frequently reported areas rated as unclear or absent and contributed to the overall neutral scores.