Composition of plant-based diets and the incidence and prognosis of inflammatory bowel disease: a multinational retrospective cohort study

Abstract

Background: Many currently proposed diets for inflammatory bowel disease (IBD) focus on increasing plant-based foods, although a vegetarian diet can still contain products such as emulsifiers and refined grains that are believed to negatively impact IBD incidence and progression. To better inform dietary management in IBD, we investigated the association between plant-based diets and the incidence and complications of IBD.

Methods: We leveraged data from the UK Biobank (UKB, 2009-2022) including 187,888 participants free of IBD at baseline and the European Prospective Investigation into Cancer and Nutrition (EPIC, 1991-2010) cohort including 341,539 individuals free of IBD across centres among Denmark, France, Germany, Greece, Italy, the Netherlands, Sweden and UK. Healthy and unhealthy diets were characterised using plant-based diet indexes (PDIs); in individual participants, these were based on the 24-h dietary recalls for UKB and food frequency questionnaires for EPIC. The primary outcome was the incidence of IBD; secondary outcomes evaluated endpoints of disease prognosis (IBD-related surgery, diabetes, cardiovascular diease, and all-cause mortality). Cox regression was applied to estimate hazard ratios (HRs).

Findings: In the UKB (925 incident IBD, median follow-up 11.6 years, IQR 1.3 years), higher adherence to healthy PDI was associated with a lower IBD risk (HR 0.75, 95% CI 0.60-0.94), while higher alignment to an unhealthy PDI associated with an increased risk (HR 1.48, 95% CI 1.21-1.82) when comparing extreme quintiles of PDIs. Among individuals with established IBD, healthy PDI was inversely associated (HR 0.50, 95% CI 0.30-0.83) and unhealthy PDI was positively associated (HR 2.12, 95% CI 1.30-3.44) with need for IBD-related surgery. We did not observe significant associations between PDIs and risk of cardiovascular disease, diabetes mellitus or mortality. In the EPIC study (548 incident IBD, median follow-up 14.5 years, IQR 7.0 years), the HR of incident IBD for healthy PDI was 0.71 (95% CI 0.59-0.85) and for unhealthy PDI was 1.54 (95% CI 1.30-1.84).

Interpretation: We provide evidence that the composition of a plant-based diet may be an important determinant of the risk of developing IBD, and of disease course after diagnosis. Further research is needed to explore the mechanistic pathways linking plant-based diets and IBD incidence and prognosis.

Funding: National Natural Science Foundation of China, Natural Science Fund for Distinguished Young Scholars of Zhejiang Province, National Undergraduate Training Program for Innovation and Entrepreneurship, CRUK Career Development Fellowship, The "Co-PI" project, Natural Science Fund for Excellent Young Scholars of Hunan Province.

Keywords: Genetic susceptibility; Incidence; Inflammatory bowel disease; Mediation analysis; Plant-based diet; Prognosis.

Fig. 1

Flowchart of the study in the UK Biobank (discovery cohort) and the European Prospective Investigation into Cancer and Nutrition study (replication cohort).

Fig. 2

Distributions of planted-based diet indexes among (a) participants free of IBD in the UK Biobank; (b) diagnosed as IBD at baseline in the UK Biobank; and (c) participants free of IBD in the European Prospective Investigation into Cancer and Nutrition study. PDI, plant-based diet index; HR, hazard ratio; CI, confidence interval.

Fig. 3

Associations between the plant-based diet indexes and incident inflammatory bowel disease in UK Biobank (n = 187,888), EPIC cohort (n = 341,539), and the meta-analysis of the results from the two cohorts (n = 529,427). HRs were adjusted for age, sex, Townsend deprivation index, education, ethnicity, body mass index, smoking status, alcohol consumption, exercise time, total energy intake and total sugar intake in UK Biobank. HRs were adjusted for age, sex, centre, education level, BMI, smoking status, alcohol consumption, total sugar, total energy, and physical activity measured in the EPIC study. The 20%, 40%, 60% and 80% percentile of PDIs in the UK Biobank and EPIC study were used as cut-off values as follow: PDI: (UKB) 46.8, 49.5, 52.0, 55.0; (EPIC) 44.0, 48.0, 52.0, 56.0; healthy PDI: (UKB) 52.0, 55.7, 58.5, 62.0; (EPIC) 46.0, 50.0, 53.0, 57.0; unhealthy PDI: (UKB) 52.0, 55.0, 58.0, 61.5; (EPIC) 48.0, 52.0, 55.0, 59.0. Numbers in bold indicates significant associations. P for heterogeneity were tested using Cochran Q test. PDI, plant-based diet index; HR, hazard ratio; CI, confidence interval; EPIC, European Prospective Investigation into Cancer, and Nutrition; PY, person-years.

Fig. 4

Associations between the plant-based diet indexes and risk of related surgery among individuals) with IBD (n = 2133) in UK Biobank. Model 1 was adjusted for age and sex, while Model 2 was additionally adjusted for Townsend deprivation index, education, ethnicity, body mass index (BMI), smoking status, alcohol consumption, exercise time, total energy intake, and total sugar intake. In this sample, the 20%, 40%, 60% and 80% percentile of PDIs were used as cutoff values as follow: PDI 46.6, 49.3, 52.0, 55.0; healthy PDI 50.7, 54.0, 57.0, 61.0; unhealthy PDI: 52.5, 56.0, 59.5, 63.0. Numbers in bold indicates significant associations. PDI, plant-based diet index; HR, hazard ratio; CI, confidence interval; EPIC, European Prospective Investigation into Cancer, and Nutrition.