Abstract
Background & Aims:
Diet is thought to play a role in the development of inflammatory bowel disease (IBD), though the relationship between gluten intake and risk of IBD has not been explored. The aim of this study was to determine the relationship between gluten intake and risk of incident Crohn’s disease (CD) and ulcerative colitis (UC).
Methods:
We performed a prospective cohort study of 208,280 US participants from the Nurses’ Health Study (NHS; 1986–2016), NHSII (1991–2017), and Health Professionals Follow-up Study (1986–2016) who did not have IBD at baseline or celiac disease, and who completed semi-quantitative food frequency questionnaires. We used Cox proportional hazards modeling to estimate the risk of IBD according to quintiles of cumulative average energy-adjusted dietary gluten intake over follow-up period.
Results:
We documented 337 CD cases and 447 UC cases over 5,115,265 person-years of follow-up. Dietary gluten intake was not associated with risk of IBD. Compared to participants in the lowest quintile of gluten intake, the adjusted hazard-ratios and 95% confidence intervals (CI) for participants in the highest quintile of gluten intake were 1.16 (95% CI: 0.82–1.64; Ptrend = 0.41) for CD and 1.04 (95% CI: 0.75–1.44; Ptrend = 0.64) for UC. Adjusting for primary sources of gluten intake did not materially change our estimates.
Conclusions:
In three large adult US prospective cohorts, gluten intake was not associated with risk of CD or UC. Our findings are reassuring at a time when consumption of gluten has been increasingly perceived as a trigger for chronic gastrointestinal diseases.
Keywords: inflammatory bowel disease, Crohn’s disease, ulcerative colitis, gluten
INTRODUCTION:
Inflammatory bowel diseases (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), are chronic inflammatory diseases of the gastrointestinal tract that result from a dysregulated immune response to environmental and microbial stimuli in a genetically susceptible host1,2. Though over 200 susceptibility loci have been identified, the total variance of IBD risk explained by known genetic factors is less than 14%, highlighting the significance of environmental factors in disease development3,4. Diet is thought to play a role in IBD pathogenesis, likely due to influence on gut microbiome composition, mucosal barrier function, and mucosal inflammation5.
Gluten, the dietary trigger for celiac disease, is a protein in wheat, barley, and rye. Gluten-free diets (GFD) have gained popularity, even among those without celiac disease. A recent national health and nutrition examination surveys (NHANES) survey estimated that by 2014, 2.7 million US adults without celiac disease adhered to a GFD, increasing by over 3-fold since 20096. This may be due to gastrointestinal symptoms attributed to gluten intake, as in non-celiac gluten sensitivity, or overlap of GFD with low fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAP) diet, which is known to aid symptoms of irritable bowel syndrome7. Others suggest that GFD, popularized by media and consumer-directed marketing are utilized by patients because of perceived health benefits8.
IBD is associated with an increased risk for celiac disease9, though non-celiac gluten sensitivity is also common in those with IBD10. Some patients with IBD report improvements in clinical symptoms following dietary gluten restriction11. This may be due to undiagnosed celiac disease, which constitutes a substantial portion of celiac cases 12, or an effect of gluten on IBD activity. Evidence for the role of gluten in gut inflammation, independent of celiac disease, is scant, though one study of mice with chemical-induced colitis showed an increase in gut inflammation in mice fed wheat gluten compared to those on standard diet 13.
Therefore, there is a need to examine the relationship between dietary gluten and risk of IBD in adults without celiac disease. These studies may help inform dietary strategies for prevention of IBD. Additionally, empiric gluten avoidance is likely not without consequence. We have previously shown that avoidance of gluten from whole but not refined grains may affect cardiovascular disease risk 14. Increased whole grain consumption has also been linked to decreased risk for type 2 diabetes mellitus and mortality 15,16. Finally, indiscriminate exclusion diets have been linked to malnutrition and disordered-eating 17,18. Thus, in this study, we explored the relationship between dietary gluten intake and risk of IBD in three large prospective US cohorts of men and women.
METHODS:
Study Population
We included participants in the Nurses’ Health Study (NHS), NHSII and Health Professionals Follow-Up Study (HPFS) (Supplementary Material). Briefly, NHS is a prospective cohort study of 121,700 female nurses (ages 30–55 years) enrolled in 1976, while NHSII was established in 1989 and enrolled 116,429 female nurses (ages 25–42 years). HPFS enrolled 51,529 US male physicians (ages 40–75 years) starting in 1986. Participants completed baseline and biennial questionnaires on lifestyle factors, anthropometric measurements, and medical information. Overall follow up rates for these cohorts have been reported as 90% (NHS), 85% (NHSII), and 93% (HPFS)19,20.
We included participants who completed baseline semi-quantitative food frequency questionnaires (SFFQ) in 1986 (NHS, HPFS) and 1991 (NHSII). We excluded participants with self-reported celiac disease at any time (n = 660), IBD at baseline (n = 120), missing or implausible baseline body mass index (BMI < 10 kg/m2; n = 1,487), and missing or implausible caloric intake (< 600 or > 3500 kcal/day for women, < 800 or > 4200 kcal/day for men). Baseline characteristics of those who completed SFFQ are similar to the entire cohort 21.
The study protocol was approved by the Institutional Review Boards of the Brigham and Women’s Hospital and the Harvard T.H. Chan School of Public Health, and the IRB allowed participants’ completion of questionnaires to be considered as implied consent.
Ascertainment of IBD diagnosis
Ascertainment of IBD has been previously described in detail22 (Supplementary Material). Briefly, in all cohorts, self-reported diagnosis of CD or UC was confirmed by two gastroenterologists blinded to exposure information through detailed review of medical records.
Assessment of gluten intake and covariates
In each cohort, SFFQs were administered every four years and assessed patterns of food intake during the previous year. Participants reported how often they consumed standard portions (e.g. one cup for cereals, one slice for bread, etc) of various food items, ranging from “never or less than once a month” to “6 or more times per day”. Based on food intake, nutrient intake was calculated using the Harvard Food Composition Database.
We have previously detailed our method for estimating dietary gluten intake 14,23. Briefly, the expected protein content of wheat, rye, and barley of food items was multiplied by a conservative conversion factor of 75% to account for the gluten content of protein23. Trace sources of gluten, such as that found in condiments, were not included. In two recent validation studies nested within NHSII (n = 732) and HPFS (n = 650), compared to 7-day dietary records, the energy-adjusted Spearman correlation coefficients for gluten intake derived from SFFQ were 0.55 (0.48–0.61) and 0.58 (0.52 0.64), respectively, demonstrating the validity of SFFQ in measuring dietary gluten intake23.
We used the residual method to derive energy-adjusted gluten accounting for total caloric intake, which reduces the effect of extraneous variation in nutrient reporting24. Additionally, we calculated cumulative average of energy-adjusted gluten intake, to reduce measurement errors, account for individual variation in gluten intake over time, and better represent long-term dietary patterns.
We also assessed the following covariates: BMI, physical activity, smoking status, non-steroidal anti-inflammatory drug (NSAID) use, oral contraceptive use, menopausal hormone therapy, appendectomy, family history of IBD and Alternate Healthy Eating Index (AHEI). Detailed information regarding ascertainment of these variables is included in the Supplementary Material.
Statistical Analysis
Person-time was calculated from return of baseline questionnaire to date of IBD diagnosis, date of last returned questionnaire, death, or end of follow up (2016 for NHS, HPFS; 2017 for NHSII), whichever occurred first. We modeled gluten intake as cohort-specific quintiles of cumulative average energy adjusted intake but also assessed the relationship between more recent (simple updated) and baseline energy-adjusted intake. We used Cox proportional hazards modeling to estimate hazard ratios (HR) and 95% confidence intervals (CI). The proportional hazards assumption was tested as described in the Supplementary Material. All models were stratified by age (months), time-period (in 2-year intervals), and cohort (NHS, NHSII, or HPFS). We adjusted models for BMI (< 25, 25 < 30, or ≥ 30 kg/m2), smoking status (never, past, or current), regular NSAID use (yes/no), oral contraceptive use (never, past or current), menopausal hormone therapy (never, past, current, or pre-menopausal), appendectomy (yes/no), AHEI (quintiles), physical activity (quintiles) and family history of IBD (yes/no). All variables except family history of IBD were time-varying covariates. We tested for non-linear association between gluten intake and risk of IBD using previously described methods25.
In exploratory analyses, we examined gluten intake and risk of IBD according to strata of BMI (< 25 or ≥ 25 kg/m2), baseline median age (< 45 or ≥ 45 years), smoking status (ever/never), and family history of IBD (yes/no) and evaluated for presence of effect modification using log likelihood ratio. We performed several sensitivity analyses including 4-and 8-year lag analyses, adjustment for primary sources of gluten, controlling for sugar intake, and substitution of AHEI with fruit and vegetables (Supplementary Material). All statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC). Statistical significance was defined as a p value < 0.05 using two-tailed tests.
RESULTS:
We included 208,280 participants in our study (NHS: 72,474; NHSII: 94,074; and HPFS: 41,732). Table 1 and Supplementary Table 1 show baseline characteristics of the pooled and individual cohorts, respectively, by quintiles of baseline gluten intake. Mean daily cumulative average gluten intake remained stable from baseline to follow up (5.7g (SD 2.2) and 6.0g (SD 1.6), respectively). At baseline, compared to participants in the lowest quintile of gluten intake, participants in the highest quintile had higher intake of whole and refined grains and were less likely to be current smokers. Baseline characteristics were otherwise similar.
Table 1.
Age-adjusted baseline demographics of pooled cohort by quintiles of energy-adjusted gluten intake.
| Characteristics a | Quintiles of baseline energy-adjusted gluten intake | ||||
|---|---|---|---|---|---|
| 1 (lowest) | 2 | 3 | 4 | 5 (highest) | |
| N = 41,287 | N = 41,628 | N = 41,703 | N = 41,814 | N = 41,848 | |
| Age (years) | 45.7 (10.9) | 45.2 (10.7) | 45.2 (10.6) | 45.1 (10.6) | 45.4 (10.7) |
| Sex (% female) | 81 | 80 | 80 | 80 | 80 |
| Body mass index (kg/m2) | 25.4 (5.1) | 25.2 (4.9) | 25.1 (4.8) | 25.0 (4.7) | 24.5 (4.6) |
| Alternate Healthy Eating Index score | 50.1 (11.9) | 49.9 (11.2) | 50.1 (11.0) | 50.4 (11.1) | 52.2 (11.3) |
| Physical activity, MET-hours/week | 18.7 (25.8) | 18.2 (24.7) | 18.1 (23.5) | 17.9 (23.3) | 19.2 (26.3) |
| Smoking status (%) | |||||
| Never | 51 | 54 | 55 | 56 | 57 |
| Past smoker | 30 | 30 | 31 | 31 | 32 |
| Current smoker | 20 | 16 | 14 | 13 | 11 |
| Regular NSAIDs use (%) b | 18 | 18 | 18 | 17 | 16 |
| History of appendectomy (%) | 20 | 19 | 19 | 19 | 18 |
| History of oral contraceptive use (%) c | 63 | 63 | 63 | 62 | 62 |
| Menopausal hormone use c | |||||
| Pre-menopausal (%) | 68 | 68 | 69 | 69 | 69 |
| Never used (%) | 16 | 16 | 16 | 16 | 15 |
| Past user (%) | 7 | 7 | 6 | 6 | 7 |
| Current user (%) | 9 | 9 | 9 | 9 | 9 |
| Family history of IBD (%) | 3 | 4 | 4 | 4 | 4 |
| Gluten intake (g/day) d | 3.1 (0.8) | 4.5 (0.6) | 5.5 (0.6) | 6.6 (0.8) | 8.8 (1.9) |
| Whole grains (g/day) d | 13.1 (14.4) | 14.9 (12.9) | 17.0 (13.2) | 20.0 (14.5) | 28.5 (20.5) |
| Refined grains (g/day) d | 38.2 (17.4) | 48.5 (13.1) | 55.4 (14.0) | 62.4 (15.7) | 74.8 (22.9) |
MET metabolic equivalent of task. NSAIDnon-steroidal anti-inflammatory drug. IBDinflammatory bowel disease. Missing data: physica activity (<1%), smoking status (3%), oral contraceptives (3.4%), menopausal hormones (3.4%).
All values other than age standardized to the age distribution of the study population. Values are mean (standard deviation) unless stated otherwise. Values may not sum to 100% du to rounding.
NSAIDs use in year 1990 for NHS and 1995 for NHSII.
For female participants only.
Energy-adjusted.
Through the end of follow up, we documented 337 CD cases and 447 UC cases over 5,115,265 person-years. Mean time from enrollment to diagnosis of IBD was 15.7 years (SD 6.5). We observed no heterogeneity in the association between gluten intake and risk of CD or UC across cohorts (all Pinteraction > 0.77) and therefore conducted pooled analysis by combining individual-level data from all cohorts. The risk of CD and UC according to quintiles of gluten intake are shown in Table 2 (pooled cohort) and Supplementary Tables 2–4 (individual cohorts). In our pooled analysis, we did not observe an association between cumulative average intake of gluten and risk of CD or UC in either age-adjusted or multivariable-adjusted analysis. For those in the highest quintile of gluten intake, the MV-adjusted HR (aHR) was 1.16 (95% CI 0.82–1.64; Ptrend = 0.41) for CD and 1.04 (95% CI 0.75–1.44; Ptrend = 0.64) for UC compared to those in the lowest quintile. Similarly, we found no association between simple-updated or baseline gluten intake and risk of IBD (all Ptrend > 0.37). There was no non-linear association between gluten intake and risk of CD and UC (all Pnon-linear > 0.46, Supplementary Figure 1).
Table 2.
Energy-adjusted gluten intake and risk of Crohn’s disease and ulcerative colitis in the pooled cohort.
| Quintiles of energy-adjusted gluten intake | Ptrend | |||||
|---|---|---|---|---|---|---|
| 1 (lowest) | 2 | 3 | 4 | 5 (highest) | ||
| CUMULATIVE AVERAGE | ||||||
| Crohn’s Disease | ||||||
| Cases | 61 | 66 | 71 | 67 | 72 | |
| Person-years | 926,287 | 1,043,311 | 1,067,128 | 1,066,781 | 1,011,758 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 0.96 (0.68–1.36) | 1.02 (0.72–1.43) | 0.96 (0.67–1.35) | 1.09 (0.77–1.53) | 0.66 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.97 (0.68–1.37) | 1.03 (0.73–1.46) | 0.99 (0.70–1.40) | 1.16 (0.82–1.64) | 0.41 |
| MV-adjusted HR + cereal fiber a,b | 1 (ref) | 0.99 (0.69–1.41) | 1.07 (0.75–1.53) | 1.04 (0.71–1.52) | 1.27 (0.84–1.92) | 0.29 |
| Gluten from whole grains a,c | 1 (ref) | 0.92 (0.64–1.32) | 0.95 (0.66–1.36) | 0.87 (0.59–1.29) | 0.96 (0.62–1.49) | 0.78 |
| Gluten from refined grains a,d | 1 (ref) | 0.97 (0.68–1.37) | 1.03 (0.73–1.46) | 0.99 (0.69–1.41) | 1.16 (0.81–1.67) | 0.45 |
| Ulcerative Colitis | ||||||
| Cases | 71 | 91 | 103 | 102 | 80 | |
| Person-years | 926,287 | 1,043,311 | 1,067,128 | 1,066,781 | 1,011,758 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.16 (0.85–1.59) | 1.29 (0.95–1.75) | 1.28 (0.94–1.73) | 1.06 (0.77–1.46) | 0.56 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.14 (0.83–1.55) | 1.25 (0.92–1.70) | 1.25 (0.92–1.70) | 1.04 (0.75–1.44) | 0.64 |
| MV-adjusted HR + cereal fiber a,b | 1 (ref) | 1.12 (0.81–1.53) | 1.21 (0.89–1.65) | 1.19 (0.86–1.64) | 0.96 (0.66–1.38) | 0.99 |
| Gluten from whole grains a,c | 1 (ref) | 1.17 (0.85–1.61) | 1.32 (0.95–1.82) | 1.34 (0.95–1.89) | 1.16 (0.77–1.74) | 0.33 |
| Gluten from refined grains a,d | 1 (ref) | 1.13 (0.83–1.54) | 1.24 (0.91–1.68) | 1.22 (0.90–1.68) | 1.00 (0.71–1.40) | 0.82 |
| SIMPLE UPDATED e | ||||||
| Crohn’s Disease | ||||||
| Cases | 64 | 70 | 63 | 62 | 78 | |
| Person-years | 1,018,233 | 1,025,127 | 1,027,732 | 1,021,087 | 1,023,086 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.10 (0.78–1.54) | 0.97 (0.68–1.37) | 0.96 (0.67–1.36) | 1.21 (0.87–1.68) | 0.50 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.11 (0.79–1.56) | 0.99 (0.70–1.40) | 0.98 (0.69–1.39) | 1.25 (0.89–1.75) | 0.37 |
| MV-adjusted HR + cereal fiber a,b | 1 (ref) | 1.12 (0.79–1.58) | 1.01 (0.71–1.44) | 1.01 (0.70–1.46) | 1.32 (0.91–1.91) | 0.30 |
| Gluten from whole grains a,c | 1 (ref) | 1.07 (0.76–1.51) | 0.94 (0.65–1.34) | 0.91 (0.62–1.32) | 1.11 (0.75–1.63) | 0.94 |
| Gluten from refined grains a,d | 1 (ref) | 1.11 (0.79–1.56) | 0.99 (0.70–1.40) | 0.98 (0.69–1.40) | 1.25 (0.88–1.76) | 0.40 |
| Ulcerative Colitis | ||||||
| Cases | 91 | 71 | 96 | 102 | 87 | |
| Person-years | 1,018,233 | 1,025,127 | 1,027,732 | 1,021,087 | 1,023,086 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 0.78 (0.57–1.06) | 1.05 (0.79–1.40) | 1.14 (0.86–1.51) | 0.96 (0.71–1.29) | 0.40 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.76 (0.56–1.04) | 1.02 (0.77–1.36) | 1.11 (0.83–1.47) | 0.94 (0.59–1.26) | 0.50 |
| MV-adjusted HR + cereal fiber a,b | 1 (ref) | 0.75 (0.55–1.03) | 1.00 (0.75–1.34) | 1.07 (0.80–1.44) | 0.89 (0.64–1.22) | 0.73 |
| Gluten from whole grains a,c | 1 (ref) | 0.77 (0.56–1.06) | 1.05 (0.78–1.42) | 1.16 (0.85–1.57) | 1.00 (0.71–1.41) | 0.29 |
| Gluten from refined grains a,d | 1 (ref) | 0.76 (0.55–1.03) | 1.01 (0.76–1.35) | 1.09 (0.82–1.45) | 0.91 (0.67–1.23) | 0.62 |
| BASELINE f | ||||||
| Crohn’s Disease | ||||||
| Cases | 70 | 74 | 54 | 72 | 67 | |
| Person-years | 989,017 | 1,021,634 | 1,030,932 | 1,037,266 | 1,036,417 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.04 (0.75–1.44) | 0.75 (0.52–1.07) | 1.01 (0.72–1.40) | 0.95 (0.68–1.32) | 0.70 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.05 (0.76–1.46) | 0.76 (0.53–1.08) | 1.03 (0.74–1.43) | 1.01 (0.72–1.42) | 0.98 |
| MV-adjusted HR + cereal fiber a,b | 1 (ref) | 1.05 (0.76–1.47) | 0.76 (0.53–1.09) | 1.04 (0.73–1.46) | 1.02 (0.70–1.49) | 0.98 |
| Gluten from whole grains a,c | 1 (ref) | 1.01 (0.72–1.40) | 0.70 (0.49–1.01) | 0.92 (0.65–1.31) | 0.86 (0.58–1.25) | 0.36 |
| Gluten from refined grains a,d | 1 (ref) | 1.05 (0.75–1.46) | 0.75 (0.53–1.08) | 1.02 (0.73–1.43) | 1.00 (0.70–1.41) | 0.91 |
| Ulcerative Colitis | ||||||
| Cases | 89 | 84 | 79 | 110 | 85 | |
| Person-years | 989,017 | 1,021,634 | 1,030,932 | 1,037,266 | 1,036,417 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 0.92 (0.68–1.24) | 0.86 (0.63–1.16) | 1.19 (0.90–1.58) | 0.93 (0.69–1.25) | 0.69 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.91 (0.67–1.23) | 0.84 (0.62–1.14) | 1.18 (0.89–1.56) | 0.93 (0.69–1.26) | 0.68 |
| MV-adjusted HR + cereal fiber a,b | 1 (ref) | 0.90 (0.66–1.21) | 0.83 (0.61–1.12) | 1.14 (0.85–1.52) | 0.88 (0.63–1.22) | 0.96 |
| Gluten from whole grains a,c | 1 (ref) | 0.92 (0.68–1.25) | 0.86 (0.63–1.18) | 1.21 (0.90–1.64) | 0.97 (0.69–1.37) | 0.48 |
| Gluten from refined grains a,d | 1 (ref) | 0.91 (0.67–1.22) | 0.84 (0.62–1.14) | 1.16 (0.87–1.54) | 0.90 (0.66–1.23) | 0.83 |
CI Confidence interval, HR hazard ratio, MV multivariable.
Models adjusted for BMI (<25, 25–30, ≥30 kg/m2), smoking status (never, past, current), Alternate Healthy Eating Index score (quintiles), physical activity (quintiles), nonsteroidal anti-inflammatory drug use, history of appendectomy (yes/no), and family history of IBD (yes/no).
Additionally adjusted for cereal fiber.
Additionally adjusted for refined grains.
Additionally adjusted for whole grains.
Gluten intake updated every 4 years.
Gluten intake at baseline: 1986 (NHS, HPFS) and 1991 (NHSII).
Sensitivity analysis:
In sensitivity analyses, we adjusted for primary sources of gluten (Table 2). When adjusting for refined grains, and therefore with the variance of gluten intake explained by whole grain intake, there was no association between cumulative average gluten intake and risk of IBD [aHR for highest compared to lowest quintile: 0.96 (95% CI 0.62–1.49; Ptrend = 0.78) for CD and 1.16 (95% CI 0.77–1.74; Ptrend = 0.33) for UC]. Similarly, when adjusting for whole grains, and therefore with the variance of gluten intake explained by refined grain intake, there was no association between cumulative average gluten intake and risk of IBD [aHR for highest compared to lowest quintile: 1.16 (95% CI 0.81–1.67; Ptrend = 0.45) for CD and 1.00 (95% CI 0.71–1.40; Ptrend = 0.82) for UC]. Adjustment for cereal fiber did not materially alter the association between gluten intake and IBD risk (all Ptrend > 0.29). Similarly, there was no association between whole-or refined-grain-adjusted gluten and risk of CD or UC using the residual model (all Ptrend > 0.09).
Replacement of AHEI with daily servings of fruit and vegetables did not materially change our estimates (all Ptrend > 0.49), nor did additional adjustment for added sugar (all Ptrend > 0.21). Finally, 4and 8-year lagged analyses showed no relationship between gluten intake and risk of CD and UC (all Ptrend > 0.06; Table 3).
Table 3.
Risk of ulcerative colitis and Crohn’s disease for the pooled cohort, using cumulative averages of energyadjusted gluten as exposure with a 4-and 8-year lagged period.
| Quintiles of cumulative average of energy-adjusted gluten intake | Ptrend | |||||
|---|---|---|---|---|---|---|
| 1 (lowest) | 2 | 3 | 4 | 5 (highest) | ||
| 4 YEAR LAG | ||||||
| Crohn’s Disease | ||||||
| Cases | 52 | 63 | 57 | 65 | 55 | |
| Person-years | 782,870 | 875,670 | 895,211 | 896,738 | 856,872 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.12 (0.78–1.62) | 0.98 (0.68–1.44) | 1.13 (0.78–1.63) | 0.98 (0.67–1.44) | 0.94 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.13 (0.78–1.63) | 0.98 (0.67–1.43) | 1.15 (0.80–1.66) | 1.03 (0.70–1.52) | 0.85 |
| Gluten from whole grains a | 1 (ref) | 1.06 (0.72–1.54) | 0.88 (0.59–1.31) | 0.99 (0.65–1.49) | 0.81 (0.50–1.32) | 0.40 |
| Gluten from refined grains a | 1 (ref) | 1.13 (0.78–1.63) | 0.98 (0.67–1.44) | 1.16 (0.80–1.68) | 1.04 (0.70–1.56) | 0.81 |
| Ulcerative Colitis | ||||||
| Cases | 55 | 85 | 68 | 91 | 71 | |
| Person-years | 782,870 | 875,670 | 895,211 | 896,738 | 856,872 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.41 (1.00–1.98) | 1.10 (0.77–1.57) | 1.48 (1.06–2.07) | 1.22 (0.85–1.73) | 0.30 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.39 (0.99–1.96) | 1.08 (0.76–1.55) | 1.46 (1.04–2.05) | 1.23 (0.86–1.75) | 0.27 |
| Gluten from whole grains a | 1 (ref) | 1.48 (1.04–2.10) | 1.20 (0.82–1.76) | 1.69 (1.15–2.49) | 1.53 (0.97–2.40) | 0.06 |
| Gluten from refined grains a | 1 (ref) | 1.38 (0.98–1.94) | 1.07 (0.75–1.53) | 1.44 (1.02–2.02) | 1.19 (0.82–1.72) | 0.37 |
| 8 YEAR LAG | ||||||
| Crohn’s Disease | ||||||
| Cases | 32 | 52 | 42 | 64 | 39 | |
| Person-years | 637,569 | 708,298 | 725,682 | 728,401 | 699,336 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.52 (0.98–2.36) | 1.21 (0.77–1.93) | 1.85 (1.21–2.84) | 1.17 (0.73–1.86) | 0.31 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.52 (0.98–2.37) | 1.23 (0.77–1.95) | 1.87 (1.22–2.87) | 1.23 (0.77–1.98) | 0.21 |
| Gluten from whole grains a | 1 (ref) | 1.45 (0.92–2.27) | 1.13 (0.69–1.83) | 1.65 (1.02–2.68) | 1.02 (0.57–1.83) | 0.70 |
| Gluten from refined grains a | 1 (ref) | 1.51 (0.97–2.36) | 1.21 (0.76–1.93) | 1.83 (1.19–2.83) | 1.19 (0.73–1.94) | 0.28 |
| Ulcerative Colitis | ||||||
| Cases | 51 | 67 | 49 | 71 | 50 | |
| Person-years | 637,569 | 708,298 | 725,682 | 728,401 | 699,336 | |
| Age-adjusted HR (95% CI) | 1 (ref) | 1.21 (0.84–1.75) | 0.86 (0.58–1.28) | 1.26 (0.88–1.81) | 0.93 (0.63–1.37) | 0.83 |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.19 (0.82–1.71) | 0.84 (0.56–1.24) | 1.24 (0.86–1.79) | 0.93 (0.63–1.39) | 0.87 |
| Gluten from whole grains a | 1 (ref) | 1.29 (0.88–1.88) | 0.96 (0.63–1.47) | 1.50 (0.98–2.29) | 1.23 (0.75–2.04) | 0.29 |
| Gluten from refined grains a | 1 (ref) | 1.19 (0.82–1.71) | 0.84 (0.56–1.24) | 1.24 (0.86–1.79) | 0.93 (0.62–1.40) | 0.88 |
Models are adjusted for same variables as in Table 2.
Exploratory analysis:
We examined the association between gluten intake and risk of IBD according to strata of baseline age, BMI, smoking status and family history of IBD (Tables 4 and 5). We did not observe any effect modification according to BMI, smoking status, or family history in CD or UC (all Pinteraction > 0.27). However, the association between gluten intake and risk of CD appeared to be modified by age at baseline (Pinteraction = 0.049). For participants aged ≥ 45 years, those in the highest quintile of gluten intake had an aHR of 1.49 (95% CI: 0.85–2.63; Ptrend = 0.06) for risk of CD compared to those in the lowest quintile. This association reached significance after additional adjustment for whole grains and cereal fiber (Ptrend = 0.02 for each), but not after additional adjustment for refined grains (Ptrend = 0.54). The associations were similar when using refined-grain-adjusted gluten (Ptrend = 0.53) and slightly attenuated when using whole grain-adjusted gluten (Ptrend = 0.06). Among those < 45 years at baseline, we observed no association between gluten intake and risk of CD (aHR for the highest quintile = 0.92 (95% CI 0.59–1.42), Ptrend = 0.35), including after additional adjustment for refined grains, whole grains or cereal fiber (all Ptrend > 0.21). There was no evidence of effect modification according to baseline age for UC (Pinteraction = 0.83).
Table 4.
Gluten intake and risk of Crohn’s disease according to baseline age, body mass index, smoking status, and family history.
| Quintiles of cumulative average energy-adjusted gluten intake | Ptrend | Pinteraction | |||||
|---|---|---|---|---|---|---|---|
| 1 (lowest) | 2 | 3 | 4 | 5 (highest) | |||
| Age < 45 | 0.049 | ||||||
| Cases | 41 | 41 | 36 | 30 | 41 | ||
| Person-years | 514,656 | 585,627 | 599,177 | 602,690 | 575,310 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.88 (0.57–1.36) | 0.76 (0.48–1.19) | 0.63 (0.39–1.01) | 0.88 (0.57–1.36) | 0.27 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.88 (0.57–1.35) | 0.75 (0.48–1.18) | 0.63 (0.39–1.01) | 0.92 (0.59–1.42) | 0.35 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.88 (0.57–1.37) | 0.76 (0.47–1.21) | 0.64 (0.38–1.06) | 0.94 (0.55–1.60) | 0.42 | |
| Gluten from whole grains a | 1 (ref) | 0.85 (0.54–1.33) | 0.71 (0.44–1.15) | 0.58 (0.34–0.99) | 0.81 (0.45–1.47) | 0.21 | |
| Gluten from refined grains a | 1 (ref) | 0.86 (0.56–1.33) | 0.73 (0.47–1.15) | 0.61 (0.37–0.98) | 0.86 (0.54–1.36) | 0.23 | |
| Age ≥ 45 | |||||||
| Cases | 20 | 25 | 35 | 37 | 31 | ||
| Person-years | 411,631 | 457,684 | 467,951 | 464,091 | 436,448 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.11 (0.62–2.00) | 1.53 (0.88–2.66) | 1.64 (0.95–2.82) | 1.45 (0.83–2.55) | 0.08 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.09 (0.61–1.97) | 1.52 (0.88–2.64) | 1.64 (0.95–2.83) | 1.49 (0.85–2.63) | 0.06 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.15 (0.63–2.09) | 1.64 (0.93–2.91) | 1.84 (1.02–3.31) | 1.81 (0.93–3.50) | 0.02 | |
| Gluten from whole grains a | 1 (ref) | 1.01 (0.56–1.83) | 1.32 (0.75–2.33) | 1.33 (0.74–2.39) | 1.09 (0.57–2.11) | 0.54 | |
| Gluten from refined grains a | 1 (ref) | 1.12 (0.62–2.03) | 1.58 (0.91–2.76) | 1.75 (1.00–3.04) | 1.69 (0.93–3.07) | 0.02 | |
| Body mass index < 25 kg/m 2 | 0.63 | ||||||
| Cases | 37 | 34 | 38 | 42 | 49 | ||
| Person-years | 522,076 | 595,745 | 624,113 | 647,720 | 651,772 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.79 (0.50–1.26) | 0.84 (0.53–1.32) | 0.89 (0.57–1.39) | 1.03 (0.67–1.58) | 0.65 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.79 (0.49–1.26) | 0.85 (0.54–1.34) | 0.91 (0.58–1.41) | 1.11 (0.72–1.70) | 0.44 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.81 (0.50–1.30) | 0.89 (0.56–1.42) | 0.96 (0.60–1.56) | 1.22 (0.73–2.06) | 0.33 | |
| Gluten from whole grains a | 1 (ref) | 0.77 (0.48–1.24) | 0.82 (0.51–1.32) | 0.85 (0.52–1.40) | 1.01 (0.58–1.74) | 0.83 | |
| Gluten from refined grains a | 1 (ref) | 0.79 (0.49–1.26) | 0.86 (0.54–1.35) | 0.91 (0.58–1.43) | 1.11 (0.71–1.76) | 0.46 | |
| Body mass index ≥ 25 kg/m 2 | |||||||
| Cases | 24 | 32 | 33 | 25 | 23 | ||
| Person-years | 404,212 | 447,566 | 443,015 | 419,061 | 359,985 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.21 (0.71–2.05) | 1.26 (0.74–2.13) | 1.01 (0.57–1.77) | 1.08 (0.61–1.93) | 0.96 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.19 (0.70–2.02) | 1.19 (0.70–2.03) | 1.02 (0.58–1.80) | 1.10 (0.62–1.96) | 0.98 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.23 (0.71–2.11) | 1.26 (0.72–2.21) | 1.10 (0.59–2.05) | 1.24 (0.61–2.51) | 0.72 | |
| Gluten from whole grains a | 1 (ref) | 1.09 (0.63–1.87) | 1.02 (0.57–1.80) | 0.82 (0.43–1.55) | 0.78 (0.37–1.64) | 0.34 | |
| Gluten from refined grains a | 1 (ref) | 1.19 (0.69–2.03) | 1.19 (0.70–2.04) | 1.02 (0.57–1.81) | 1.09 (0.59–2.00) | 0.999 | |
| Never smoker | 0.49 | ||||||
| Cases | 31 | 31 | 31 | 34 | 35 | ||
| Person-years | 491,693 | 575,498 | 598,206 | 609,022 | 593,707 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.86 (0.52–1.42) | 0.84 (0.51–1.39) | 0.88 (0.54–1.44) | 0.93 (0.57–1.52) | 0.87 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.82 (0.50–1.35) | 0.81 (0.49–1.34) | 0.86 (0.53–1.40) | 0.92 (0.56–1.50) | 0.87 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.85 (0.51–1.41) | 0.86 (0.51–1.44) | 0.93 (0.54–1.59) | 1.05 (0.57–1.90) | 0.79 | |
| Gluten from whole grains a | 1 (ref) | 0.79 (0.48–1.31) | 0.76 (0.45–1.27) | 0.78 (0.46–1.32) | 0.78 (0.43–1.42) | 0.48 | |
| Gluten from refined grains a | 1 (ref) | 0.81 (0.49–1.35) | 0.81 (0.49–1.34) | 0.85 (0.52–1.40) | 0.91 (0.54–1.52) | 0.82 | |
| Past or current smoker | |||||||
| Cases | 30 | 35 | 40 | 33 | 37 | ||
| Person-years | 434,594 | 467,813 | 468,922 | 457,759 | 418,051 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.07 (0.66–1.75) | 1.23 (0.76–1.97) | 1.03 (0.63–1.70) | 1.27 (0.78–2.06) | 0.43 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.09 (0.67–1.78) | 1.23 (0.77–1.99) | 1.05 (0.63–1.72) | 1.32 (0.81–2.14) | 0.36 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.11 (0.68–1.83) | 1.28 (0.78–2.10) | 1.11 (0.65–1.89) | 1.44 (0.81–2.58) | 0.29 | |
| Gluten from whole grains a | 1 (ref) | 1.03 (0.62–1.71) | 1.13 (0.68–1.88) | 0.92 (0.52–1.62) | 1.08 (0.57–2.05) | 0.98 | |
| Gluten from refined grains a | 1 (ref) | 1.09 (0.67–1.79) | 1.24 (0.77–2.01) | 1.06 (0.64–1.76) | 1.34 (0.81–2.24) | 0.35 | |
| No family history | 0.29 | ||||||
| Cases | 51 | 52 | 61 | 58 | 61 | ||
| Person-years | 892,124 | 1,000,070 | 1,020,764 | 1,022,061 | 968,681 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.91 (0.62–1.34) | 1.05 (0.72–1.52) | 1.00 (0.68–1.46) | 1.10 (0.76–1.59) | 0.50 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.91 (0.62–1.35) | 1.07 (0.73–1.55) | 1.03 (0.71–1.51) | 1.17 (0.80–1.70) | 0.30 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.95 (0.64–1.41) | 1.14 (0.77–1.68) | 1.14 (0.76–1.72) | 1.38 (0.87–2.17) | 0.12 | |
| Gluten from whole grains a | 1 (ref) | 0.87 (0.59–1.28) | 0.97 (0.66–1.44) | 0.90 (0.60–1.37) | 0.95 (0.60–1.52) | 0.93 | |
| Gluten from refined grains a | 1 (ref) | 0.92 (0.62–1.35) | 1.07 (0.74–1.56) | 1.04 (0.71–1.52) | 1.18 (0.80–1.75) | 0.31 | |
| Positive family history | |||||||
| Cases | 10 | 14 | 10 | 9 | 11 | ||
| Person-years | 34,163 | 43,241 | 46,364 | 44,720 | 43,077 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.06 (0.45–2.52) | 0.66 (0.26–1.66) | 0.60 (0.23–1.56) | 0.71 (0.28–1.80) | 0.24 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.94 (0.39–2.29) | 0.62 (0.24–1.61) | 0.53 (0.19–1.42) | 0.74 (0.29–1.89) | 0.29 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.90 (0.37–2.21) | 0.59 (0.22–1.55) | 0.47 (0.16–1.37) | 0.64 (0.22–1.83) | 0.22 | |
| Gluten from whole grains a | 1 (ref) | 0.91 (0.36–2.27) | 0.58 (0.21–1.65) | 0.48 (0.15–1.52) | 0.64 (0.18–2.35) | 0.29 | |
| Gluten from refined grains a | 1 (ref) | 0.93 (0.38–2.27) | 0.61 (0.23–1.59) | 0.51 (0.18–1.41) | 0.70 (0.26–1.91) | 0.27 | |
Models are adjusted for the same variables as Table 2 minus the strata variable.
Table 5.
Gluten intake and risk of ulcerative colitis according to baseline age, body mass index, smoking status, and family history.
| Quintiles of cumulative average energy-adjusted gluten intake | Ptrend | Pinteraction | |||||
|---|---|---|---|---|---|---|---|
| 1 (lowest) | 2 | 3 | 4 | 5 (highest) | |||
| Age < 45 | 0.83 | ||||||
| Cases | 48 | 52 | 65 | 59 | 53 | ||
| Person-years | 514,656 | 585,627 | 599,177 | 602,690 | 575,310 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.98 (0.66–1.46) | 1.19 (0.82–1.73) | 1.07 (0.73–1.57) | 1.02 (0.69–1.50) | 0.81 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.98 (0.66–1.45) | 1.18 (0.81–1.72) | 1.07 (0.73–1.57) | 1.02 (0.68–1.51) | 0.80 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.95 (0.64–1.41) | 1.12 (0.76–1.64) | 0.99 (0.66–1.47) | 0.89 (0.57–1.38) | 0.71 | |
| Gluten from whole grains a | 1 (ref) | 1.03 (0.69–1.55) | 1.28 (0.85–1.93) | 1.21 (0.77–1.88) | 1.22 (0.73–2.06) | 0.34 | |
| Gluten from refined grains a | 1 (ref) | 0.97 (0.65–1.43) | 1.15 (0.79–1.67) | 1.02 (0.69–1.51) | 0.94 (0.62–1.41) | 0.88 | |
| Age ≥ 45 | |||||||
| Cases | 23 | 39 | 38 | 43 | 27 | ||
| Person-years | 411,631 | 457,684 | 467,951 | 464,091 | 436,448 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.56 (0.93–2.62) | 1.46 (0.87–2.46) | 1.69 (1.02–2.80) | 1.12 (0.64–1.95) | 0.66 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.51 (0.90–2.53) | 1.43 (0.85–2.40) | 1.68 (1.01–2.80) | 1.13 (0.65–1.98) | 0.57 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.49 (0.89–2.52) | 1.41 (0.82–2.40) | 1.64 (0.96–2.82) | 1.09 (0.58–2.05) | 0.66 | |
| Gluten from whole grains a | 1 (ref) | 1.48 (0.88–2.51) | 1.39 (0.81–2.40) | 1.62 (0.93–2.84) | 1.07 (0.55–2.08) | 0.75 | |
| Gluten from refined grains a | 1 (ref) | 1.51 (0.90–2.53) | 1.43 (0.85–2.41) | 1.69 (1.01–2.83) | 1.14 (0.63–2.05) | 0.54 | |
| Body mass index < 25 kg/m 2 | 0.90 | ||||||
| Cases | 48 | 49 | 67 | 64 | 56 | ||
| Person-years | 522,076 | 595,745 | 624,113 | 647,720 | 651,772 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.93 (0.62–1.39) | 1.20 (0.83–1.74) | 1.10 (0.76–1.60) | 0.96 (0.65–1.41) | 0.89 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.92 (0.62–1.37) | 1.20 (0.82–1.74) | 1.13 (0.77–1.64) | 1.01 (0.68–1.49) | 0.63 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.90 (0.60–1.35) | 1.15 (0.79–1.69) | 1.06 (0.71–1.58) | 0.91 (0.58–1.43) | 0.98 | |
| Gluten from whole grains a | 1 (ref) | 0.95 (0.63–1.43) | 1.27 (0.85–1.89) | 1.22 (0.80–1.87) | 1.14 (0.70–1.87) | 0.34 | |
| Gluten from refined grains a | 1 (ref) | 0.92 (0.61–1.37) | 1.19 (0.81–1.73) | 1.11 (0.76–1.63) | 0.98 (0.65–1.47) | 0.74 | |
| Body mass index ≥ 25 kg/m 2 | |||||||
| Cases | 23 | 42 | 36 | 38 | 24 | ||
| Person-years | 404,212 | 447,566 | 443,015 | 419,061 | 359,985 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.64 (0.99–2.74) | 1.45 (0.86–2.46) | 1.60 (0.95–2.69) | 1.17 (0.66–2.09) | 0.67 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.61 (0.96–2.68) | 1.44 (0.85–2.43) | 1.56 (0.93–2.64) | 1.15 (0.65–2.05) | 0.72 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.56 (0.93–2.62) | 1.36 (0.79–2.35) | 1.45 (0.83–2.54) | 1.02 (0.52–1.98) | 0.97 | |
| Gluten from whole grains a | 1 (ref) | 1.60 (0.94–2.72) | 1.43 (0.81–2.53) | 1.55 (0.85–2.85) | 1.14 (0.55–2.38) | 0.81 | |
| Gluten from refined grains a | 1 (ref) | 1.57 (0.94–2.63) | 1.39 (0.81–2.36) | 1.48 (0.87–2.52) | 1.05 (0.58–1.92) | 0.98 | |
| Never smoker | 0.27 | ||||||
| Cases | 29 | 38 | 49 | 53 | 43 | ||
| Person-years | 491,693 | 575,498 | 598,206 | 609,022 | 593,707 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.16 (0.72–1.89) | 1.43 (0.90–2.26) | 1.49 (0.94–2.34) | 1.26 (0.78–2.01) | 0.21 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.15 (0.71–1.87) | 1.39 (0.87–2.20) | 1.46 (0.92–2.30) | 1.25 (0.78–2.01) | 0.22 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.12 (0.69–1.83) | 1.33 (0.83–2.13) | 1.38 (0.85–2.22) | 1.13 (0.66–1.93) | 0.46 | |
| Gluten from whole grains a | 1 (ref) | 1.20 (0.73–1.96) | 1.50 (0.92–2.44) | 1.63 (0.98–2.73) | 1.49 (0.82–2.69) | 0.10 | |
| Gluten from refined grains a | 1 (ref) | 1.14 (0.70–1.85) | 1.37 (0.86–2.18) | 1.43 (0.90–2.27) | 1.20 (0.73–1.97) | 0.29 | |
| Past or current smoker | |||||||
| Cases | 42 | 53 | 54 | 49 | 37 | ||
| Person-years | 434,594 | 467,813 | 468,922 | 457,759 | 418,051 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.20 (0.80–1.80) | 1.21 (0.81–1.82) | 1.13 (0.75–1.72) | 0.94 (0.60–1.47) | 0.75 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.18 (0.79–1.78) | 1.19 (0.79–1.79) | 1.11 (0.73–1.67) | 0.92 (0.59–1.43) | 0.66 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.15 (0.76–1.73) | 1.13 (0.75–1.72) | 1.03 (0.66–1.60) | 0.82 (0.49–1.36) | 0.42 | |
| Gluten from whole grains a | 1 (ref) | 1.18 (0.77–1.80) | 1.19 (0.76–1.84) | 1.10 (0.68–1.78) | 0.91 (0.51–1.62) | 0.74 | |
| Gluten from refined grains a | 1 (ref) | 1.17 (0.78–1.76) | 1.16 (0.77–1.75) | 1.07 (0.70–1.62) | 0.86 (0.54–1.37) | 0.48 | |
| No family history | 0.56 | ||||||
| Cases | 60 | 79 | 94 | 85 | 71 | ||
| Person-years | 892,124 | 1,000,070 | 1,020,764 | 1,022,061 | 968,681 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 1.21 (0.87–1.70) | 1.40 (1.01–1.94) | 1.26 (0.91–1.76) | 1.12 (0.79–1.57) | 0.55 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 1.21 (0.86–1.69) | 1.40 (1.01–1.94) | 1.27 (0.91–1.77) | 1.13 (0.80–1.60) | 0.48 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 1.20 (0.85–1.68) | 1.37 (0.98–1.91) | 1.23 (0.86–1.75) | 1.07 (0.72–1.60) | 0.69 | |
| Gluten from whole grains a | 1 (ref) | 1.22 (0.86–1.72) | 1.42 (1.00–2.00) | 1.29 (0.89–1.88) | 1.17 (0.76–1.80) | 0.46 | |
| Gluten from refined grains a | 1 (ref) | 1.20 (0.86–1.69) | 1.38 (1.00–1.92) | 1.25 (0.89–1.75) | 1.10 (0.77–1.58) | 0.60 | |
| Positive family history | |||||||
| Cases | 11 | 12 | 9 | 17 | 9 | ||
| Person-years | 34,163 | 43,241 | 46,364 | 44,720 | 43,077 | ||
| Age-adjusted HR (95% CI) | 1 (ref) | 0.89 (0.38–2.10) | 0.52 (0.21–1.33) | 1.23 (0.55–2.76) | 0.58 (0.23–1.46) | 0.57 | |
| MV-adjusted HR (95% CI) a | 1 (ref) | 0.80 (0.34–1.91) | 0.52 (0.20–1.32) | 1.26 (0.55–2.86) | 0.60 (0.24–1.52) | 0.71 | |
| MV-adjusted HR + cereal fiber a | 1 (ref) | 0.75 (0.31–1.81) | 0.44 (0.17–1.14) | 0.96 (0.40–2.29) | 0.38 (0.13–1.08) | 0.20 | |
| Gluten from whole grains a | 1 (ref) | 0.90 (0.36–2.20) | 0.63 (0.23–1.73) | 1.64 (0.63–4.31) | 0.90 (0.27–2.98) | 0.59 | |
| Gluten from refined grains a | 1 (ref) | 0.80 (0.34–1.93) | 0.49 (0.19–1.25) | 1.18 (0.52–2.69) | 0.49 (0.18–1.30) | 0.46 | |
Models are adjusted for the same varaibles as Table 2 minus the strata variable.
DISCUSSION:
In three large adult US prospective cohorts, we found no association between long-term intake of gluten and risk of subsequent IBD. The lack of an association was consistent across various exposure time-windows and subgroups defined by IBD risk factors.
Our study represents the first investigation of the relationship between long-term gluten intake and risk of incident IBD. Interestingly, prior studies have shown that many patients with IBD report non-celiac gluten sensitivity and improved symptoms with GFD 10,11. These observations raise important questions regarding the biological mechanisms that underpin the relationship between gluten and gastrointestinal symptoms. In an animal model of colitis, mice fed gluten had increased intestinal mucosal inflammation, permeability, and bacterial translocation when compared to mice fed standard diet13. In contrast, among healthy participants without celiac disease, consumption of high amount of gluten did not result in malabsorption or an acute inflammatory response in intestinal biopsies 26. Therefore, as previously suggested, it is possible that non-gluten-components, such as FODMAPs or non-gluten wheat proteins, may explain gastrointestinal symptoms commonly attributed to gluten7,27.
GFDs have gained popularity in those without celiac disease 6. However, GFD may result in changes in diet quality. For example, we have previously demonstrated that avoidance of gluten from whole but not refined grain may affect cardiovascular disease risk 14. Dietary grain source has also been implicated in systemic inflammation and intestinal microbiota composition. A diet high in refined grain has been associated with increased levels of C-reactive protein (CRP), interleukin 6 (IL-6) and soluble intracellular cell adhesion molecule 1 (sICAM-1) 28,29. In contrast, a diet high in whole grain is associated with a reduction in tumor necrosis factor-alpha (TNF-alpha), IL-6 and CRP 30,31 and increased abundance of anti-inflammatory short chain fatty acid (SCFA)-producing microbes, reflected by increased Firmicutes to Bacteroidetes ratios32,33. Two randomized dietary intervention studies demonstrated that changes in circulating inflammatory markers accompanying whole-grain diet coincided with alterations in gut microbiota composition 31,33. Additionally, Faecalibacterium prausnitzii and Prevotella copri, previously implicated in CD pathogenesis, are increased in abundance with whole grain consumption and decreased with refined grain consumption30,34. Thus, empiric gluten avoidance, particularly from whole grains may have significant negative biological and health consequences.
In exploratory analysis we found that increased consumption of gluten from refined but not whole grains was associated with increased risk of CD among older adults. This may be related to the pro-inflammatory effects of refined grains rather than gluten itself. In support of this, when adjusting our models for refined grain, the association between gluten and CD risk in those aged ≥ 45 years was fully attenuated (Table 4), while refined grain itself was associated with increased CD risk in this group (aHR = 1.78, 95% CI 1.06 – 2.99 for highest quintile, Ptrend = 0.03). That these findings were limited to older adults may be related to a greater role of environment in older-onset CD compared to younger-onset disease35. Aging has been linked to marked reduction in microbial diversity and immunosenescence 36,37. Therefore, it is plausible that the known pro-inflammatory effect of refined grain on the gut microbiome and immune function is significantly augmented in older adults38.
Our study has several strengths. Dietary data were collected prospectively every 4 years over nearly 30 years, accounting for long-term changes in dietary intake and minimizing the possibility of recall or selection biases. We confirmed all cases of CD and UC through medical record review, minimizing the possibility of outcome misclassification. Additionally, we accounted for important confounders using detailed and validated information on lifestyle, medication and dietary factors. Finally, our study spanned a broad age range, with mean baseline age of 36 years in NHSII to 53 years in NHS and HPFS, allowing us to examine the effects of gluten intake according to younger versus older age.
Our study also has several limitations. There is measurement error with the use of SFFQ for estimating nutrient intake. However, our validation study demonstrated moderate to good correlation between SFFQ and diet records for estimating gluten intake5. SFFQ data could not capture if participants were on strict GFD, thus we could not examine the relationship between a GFD and risk of IBD. As most gluten consumption in the US comes from wheat products, we cannot fully assess the impact of gluten from sources such as barley and rye. We also did not consider minor sources of gluten such as soy sauce or condiments, though we expect these sources to have minimal contribution to total gluten intake. We also note that the baseline age in our cohort is higher than the average age of IBD onset in the US, thus younger-onset IBD may be under-represented. However, baseline age for NHSII participants ranged from 26–46 years which allowed us to also capture younger-onset cases. Participants in our cohorts are also health care professionals and are largely white and female, thus generalizability may be limited. We also acknowledge that the positive association in exploratory analysis may represent a type 1 error related to multiple comparisons or residual confounding due to unavailability of important risk factors such as antibiotics use in our cohorts.
CONCLUSION:
Long-term intake of gluten did not confer an increased risk of developing IBD in adults US, even among high-risk participants with a family history of IBD. Our findings do not support the theory that gluten contributes to IBD development. This is important because of the established health benefits of a diet rich in whole grains. Our observation that gluten from refined grains may be associated with an increased risk of CD in older adults warrants confirmation in further prospective studies. Thus, gluten should not be empirically avoided in persons without celiac disease for the purpose of preventing IBD.
Supplementary Material
WHAT YOU NEED TO KNOW.
BACKGROUND:
Diet plays a role in inflammatory bowel disease (IBD) pathogenesis. Some patients with IBD report gastrointestinal symptoms with gluten intake. However, the relationship between gluten intake and risk of IBD is unknown.
FINDINGS:
Our study suggests that gluten intake does not confer increased risk for development of Crohn’s disease or ulcerative colitis.
IMPLICATIONS FOR PATIENT CARE:
Empiric gluten avoidance, which may have adverse health effects, should not be recommended for the purpose of preventing IBD.
Acknowledgement
We would like to thank the participants and staff of the NHS, NHSII and HPFS for their valuable contributions.
Grant Support
Funded by UM1 CA186107 NHS cohort infrastructure grant, U01 CA176726 NHSII cohort infrastructure grant, U01 CA167552 HPFS cohort infrastructure grant. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This study was funded by a senior research award from the Crohn’s and Colitis Foundation to HK and a senior research award from the Crohn’s and Colitis Foundation to ATC. EL is funded from NIH T32 DK007191. Funding sources did not participate in study design, analysis, interpretation, drafting of manuscript, or submission process.
Abbreviations:
- (CD)
Crohn’s disease
- (FODMAP)
low fermentable oligosaccharides, disaccharides, monosaccharides and polyols
- (GFD)
gluten-free diet
- (HPFS)
Health Professionals Follow-up Study
- (IBD)
Inflammatory bowel disease
- (NHS)
Nurses’ Health Study
- (NHSII)
Nurses’ Health Study II
- (SFFQ)
semi-quantitative food frequency questionnaire
- (UC)
ulcerative colitis
Footnotes
Disclosures
All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: HK is supported by the American College of Gastroenterology Senior Research Award and the Beker Foundation; HK has received consulting fees from Abbvie and Takeda; HK has also received grant funding from Pfizer and Takeda; BL personal fees from Takeda, Innovate Biopharmaceuticals, and Anokion, outside the submitted work; AA reports advisory board participation for Gilead and Kyn Therapeutics; AA reports grants from Pfizer, outside the submitted work; ATC is the Stuart and Suzanne MGH Research Scholar; JFL reports funding from Janssen corporation for work unrelated to this manuscript; ATC has received consulting fees from Bayer Pharma AG, Pfizer Inc., and Boehringer Ingelheim for work unrelated to this manuscript. There are no other relationships or activities that could appear to have influenced the submitted work. All authors had full access to the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES:
- 1.Rubin DT, Ananthakrishnan AN, Siegel CA, et al. ACG Clinical Guideline: Ulcerative Colitis in Adults. Am J Gastroenterol. 2019;114(3):384–413. doi: 10.14309/ajg.0000000000000152 [DOI] [PubMed] [Google Scholar]
- 2.Baumgart DC, Sandborn WJ. Crohn’s disease. In: The Lancet. Vol 380. Lancet Publishing Group; 2012:1590–1605. doi: 10.1016/S0140-6736(12)60026-9 [DOI] [PubMed] [Google Scholar]
- 3.Jostins L, Ripke S, Weersma RK, et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature. 2012;491(7422):119–124. doi: 10.1038/nature11582 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Liu JZ, Van Sommeren S, Huang H, et al. Association analyses identify 38 susceptibility loci for inflammatory bowel disease and highlight shared genetic risk across populations. Nat Genet. 2015;47(9):979–986. doi: 10.1038/ng.3359 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Khalili H, Chan SSM, Lochhead P, et al. The role of diet in the aetiopathogenesis of inflammatory bowel disease. Nat Rev Gastroenterol Hepatol. 2018;15(9):525–535. doi: 10.1038/s41575-018-0022-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kim HS, Patel KG, Orosz E, et al. Time trends in the prevalence of celiac disease and gluten-free diet in the US population: Results from the national health and nutrition examination surveys 2009–2014. JAMA Intern Med. 2016;176(11):1716–1717. doi: 10.1001/jamainternmed.2016.5254 [DOI] [PubMed] [Google Scholar]
- 7.Biesiekierski JR, Peters SL, Newnham ED, et al. No effects of gluten in patients with self-reported non-celiac gluten sensitivity after dietary reduction of fermentable, poorly absorbed, short-chain carbohydrates. Gastroenterology. 2013;145(2). doi: 10.1053/j.gastro.2013.04.051 [DOI] [PubMed] [Google Scholar]
- 8.Niland B, Cash BD. Health Benefits and Adverse Effects of a Gluten-Free Diet in Non-Celiac Disease Patients. Gastroenterol Hepatol (N Y). 2018;14(2). [PMC free article] [PubMed] [Google Scholar]
- 9.Shah A, Walker M, Burger D, et al. Link Between Celiac Disease and Inflammatory Bowel Disease. J Clin Gastroenterol. May 2018:1. doi: 10.1097/mcg.0000000000001033 [DOI] [PubMed] [Google Scholar]
- 10.Limketkai BN, Sepulveda R, Hing T, et al. Prevalence and factors associated with gluten sensitivity in inflammatory bowel disease. Scand J Gastroenterol. 2018;53(2):147–151. doi: 10.1080/00365521.2017.1409364 [DOI] [PubMed] [Google Scholar]
- 11.Herfarth HH, Martin CF, Sandler RS, et al. Prevalence of a gluten-free diet and improvement of clinical symptoms in patients with inflammatory bowel diseases. Inflamm Bowel Dis. 2014;20(7):1194–1197. doi: 10.1097/MIB.0000000000000077 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rubio-Tapia A, Ludvigsson JF, Brantner TL, et al. The Prevalence of Celiac Disease in the United States. Am J Gastroenterol. 2012;107(10). https://journals.lww.com/ajg/Fulltext/2012/10000/The_Prevalence_of_Celiac_Disease_in_the_United.17.aspx. [DOI] [PubMed] [Google Scholar]
- 13.Menta L, Andrade ME, Leocardio P, et al. SUN-P076 Gluten intake increase bacterial translocation and aggravates intestinal inflammation in experimental colitis. Clin Nutr. 2018;37:S69. [Google Scholar]
- 14.Lebwohl B, Cao Y, Zong G, et al. Long term gluten consumption in adults without celiac disease and risk of coronary heart disease: prospective cohort study. BMJ. 2017;357:j1892. doi: 10.1136/bmj.j1892 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.De Munter JSL, Hu FB, Spiegelman D, et al. Whole Grain, Bran, and Germ Intake and Risk of Type 2 Diabetes: A Prospective Cohort Study and Systematic Review. PLoS Med | www. 2007;4:1385. doi: 10.1371/journal.pmed [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wu H, Flint AJ, Qi Q, et al. Association between dietary whole grain intake and risk of mortality: Two large prospective studies in US Men and Women. JAMA Intern Med. 2015;175(3):373–384. doi: 10.1001/jamainternmed.2014.6283 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lim H- S, Kim S- K, Hong S- J. Food Elimination Diet and Nutritional Deficiency in Patients with Inflammatory Bowel Disease. Clin Nutr Res. 2018;7(1):48. doi: 10.7762/cnr.2018.7.1.48 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Conviser JH, Fisher SD, McColley SA. Are children with chronic illnesses requiring dietary therapy at risk for disordered eating or eating disorders? A systematic review. Int J Eat Disord. 2018;51(3):187–213. doi: 10.1002/eat.22831 [DOI] [PubMed] [Google Scholar]
- 19.History | Nurses’ Health Study. https://www.nurseshealthstudy.org/about-nhs/history. Accessed March 3, 2020.
- 20.Health Professionals Follow-Up Study About the Study. https://sites.sph.harvard.edu/hpfs/about-the-study/. Accessed March 3, 2020.
- 21.Yuan C, Spiegelman D, Rimm EB, et al. Validity of a Dietary Questionnaire Assessed by Comparison With Multiple Weighed Dietary Records or 24-Hour Recalls. Am J Epidemiol. 2017;185(7):570–584. doi: 10.1093/aje/kww104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Khalili H, Huang ES, Ananthakrishnan AN, et al. Geographical variation and incidence of inflammatory bowel disease among US women. Gut. 2012;61(12):1686–1692. doi: 10.1136/gutjnl-2011-301574 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Liu P-H, Lebwohl B, Burke KE, et al. Dietary Gluten Intake and Risk of Microscopic Colitis Among US Women without Celiac Disease. Am J Gastroenterol. 2019;114(1):127–134. doi: 10.1038/s41395-018-0267-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Willett W, Stampfer MJ, Hegsted M, et al. Total Energy Intake: Implications for Epidemiologic Analyses. Am J Epidemiol. 1986;124(1). https://academic.oup.com/aje/article-abstract/124/1/17/150019. [DOI] [PubMed] [Google Scholar]
- 25.Durrleman S, Simon R. Flexible Regression Models with Cubic Splines. Stat Med. 1989;8(5):551–561. [DOI] [PubMed] [Google Scholar]
- 26.Levine RA, Briggs GW, Harding RS, et al. Prolonged gluten administration in normal subjects. N Engl J Med. 1966;274(20):1109–1114. [DOI] [PubMed] [Google Scholar]
- 27.Zevallos VF, Raker V, Tenzer S, et al. Nutritional Wheat Amylase-Trypsin Inhibitors Promote Intestinal Inflammation via Activation of Myeloid Cells. Gastroenterology. 2017;152(5):1110–1113. doi: 10.1053/j.gastro.2016.12.006 [DOI] [PubMed] [Google Scholar]
- 28.Schulze MB, Hoffmann K, Manson JE, et al. Dietary pattern, inflammation, and incidence of type 2 diabetes in women. Am J Clin Nutr. 2005;82:675–684. https://academic.oup.com/ajcn/article-abstract/82/3/675/4863018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nettleton JA, Steffen LM, Mayer-Davis EJ, et al. Dietary patterns are associated with biochemical markers of inflammation and endothelial activation in the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Clin Nutr. 2006;83:1369–1379. https://academic.oup.com/ajcn/article-abstract/83/6/1369/4633049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Roager HM, Vogt JK, Kristensen M, et al. Whole grain-rich diet reduces body weight and systemic low-grade inflammation without inducing major changes of the gut microbiome: A randomised cross-over trial. Gut. 2019;68(1):83–93. doi: 10.1136/gutjnl-2017-314786 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Vitaglione P, Mennella I, Ferracane R, et al. Whole-grain wheat consumption reduces inflammation in a randomized controlled trial on overweight and obese subjects with unhealthy dietary and lifestyle behaviors: Role of polyphenols bound to cereal dietary fiber. Am J Clin Nutr. 2015;101(2):251–261. doi: 10.3945/ajcn.114.088120 [DOI] [PubMed] [Google Scholar]
- 32.Vanegas SM, Meydani M, Barnett JB, et al. Substituting whole grains for refined grains in a 6-wk randomized trial has a modest effect on gut microbiota and immune and inflammatory markers of healthy adults. Am J Clin Nutr. 2017;105(3):635–650. doi: 10.3945/ajcn.116.146928 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Martínez I, Lattimer JM, Hubach KL, et al. Gut microbiome composition is linked to whole grain-induced immunological improvements. ISME J. 2013;7(2):269–280. doi: 10.1038/ismej.2012.104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sokol H, né dicte Pigneur B, Watterlot L, et al. Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients. PNAS. 2008;105(43):16731–16736. www.pnas.orgcgidoi10.1073pnas.0804812105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ananthakrishnan AN, Huang H, Nguyen DD, et al. Differential effect of genetic burden on disease phenotypes in crohn’s disease and ulcerative colitis: Analysis of a north american cohort. Am J Gastroenterol. 2014;109(3):395–400. doi: 10.1038/ajg.2013.464 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Cevenini E, Monti D, Franceschi C. Inflamm-ageing. Curr Opin Clin Nutr Metab Care. 2013;16(1):14–20. doi: 10.1097/MCO.0b013e32835ada13 [DOI] [PubMed] [Google Scholar]
- 37.Claesson MJ, Jeffery IB, Conde S, et al. Gut microbiota composition correlates with diet and health in the elderly. Nature. 2012;488(7410):178–184. doi: 10.1038/nature11319 [DOI] [PubMed] [Google Scholar]
- 38.Ghosh TS, Rampelli S, Jeffery IB, et al. Mediterranean diet intervention alters the gut microbiome in older people reducing frailty and improving health status: The NU-AGE 1-year dietary intervention across five European countries. Gut. 2020;69:v–1228. doi: 10.1136/gutjnl-2019-319654 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.