Introduction Chronic inflammatory diseases (CIDs) (Crohn’s disease, ulcerative colitis, psoriasis, psoriatic arthritis, rheumatoid arthritis and multiple sclerosis) are diseases of the immune system that have some shared genetic and environmental predisposing factors, but still few studies have investigated the effects of lifestyle on disease risk of several CIDs. The primary aim of this prospective cohort study is to investigate the impact of fibre, red meat and processed meat on risk of late-onset CID, with the perspective that results of this study can contribute in supporting future diet recommendations for effective personalised prevention.
Methods and analysis The study will use data from 57 053 persons from the prospective Danish cohort study ‘Diet, Cancer and Health’ together with National Health Registry data. The follow-up period is from December 1993 to December 2018. Questionnaire data on diet and lifestyle were collected at entry to the Diet, Cancer and Health study. The outcome CID is defined as having a diagnosis of one of the CIDs registered in the National Patient Registry or, for multiple sclerosis, in the Danish Multiple Sclerosis Registry during follow-up and being treated with a drug used for the specific disease. The major outcome of the analyses will be to detect variability in risk of late onset of any CID and, if power allows, disease risk of late onset of each CID diagnosis between persons with different fibre and red meat, and processed meat intake. The outcome will be adjusted for age, sex, body mass index, physical activity, energy, alcohol, fermented dairy products, education, smoking status, hormone replacement therapy and comorbidity.
Ethics and dissemination The study is approved by the Danish Data Protection Agency (2012-58-0018). The core study is an open register-based cohort study. The study does not need approval from the Ethics committee or Institutional Review Board by Danish law. Study findings will be disseminated through peer-reviewed journals, patient associations and presentations at international conferences.
Trial registration number NCT03456206; Post-results.
- chronic inflammatory diseases
- disease risk
- red meat
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Strengths and limitations of this study
The linkage to Danish national registers will ensure almost complete follow-up of the study population, as the Danish registries are considered of high validity and completeness.
The big sample size will enable a sufficient power of the total late-onset chronic inflammatory disease (CID) group even when taking loss to emigration and competing risk of death into account.
The study includes several CIDs.
Risk of low specificity of the diagnostic codes and treatment codes as criteria for identifying late-onset CID cases.
Prospective studies including younger age groups are necessary to reveal the generalisability of the results.
The chronic inflammatory diseases (CIDs) can be considered as systemic diseases which primarily affect one organ such as the intestine (inflammatory bowel disease [IBD]: Crohn’s disease [CD] and colitis ulcerosa [UC]), skin (psoriasis [PsO]), joints (rheumatoid arthritis [RA] and psoriatic arthritis [PsA]) or the brain (multiple sclerosis [MS]).
The peak of disease onset is in the adult phase of life. The diseases have a large impact on the patients and their families’ quality of life due to lack of causative treatment, and on the society due to the absence from work and on healthcare economy due to lack of preventative measures.1–7 The CIDs have a high prevalence, with IBD and MS affecting, respectively, 0.5% and 0.1% of the population in the Western world.8–12 Studies from across the world have reported prevalence estimates of RA and PsO ranging from 0.3% to 1.0% and 0.7% to 3.2%, respectively,2 8 13–15 while the prevalence of PsA is estimated as 0.04%–0.25% across countries.14 16 Furthermore, one-third of patients with RA are diagnosed at >60 years of age17 and the incidence of late-onset IBD and MS has been reported to increase.18 19
The CIDs have some shared genetic20–24 and environmental (eg, smoking, gut microbiome) predisposing factors, and causes of the high incidence and prevalence point to such environmental factors.8 25–40 Therefore, further research of the associations between potential modifiable environmental risk factors and risk of CID is important.
It has been demonstrated that a high level of red meat consumption is a risk factor for the development of inflammatory polyarthritis (including RA).39 40 A high-fibre intake has been associated with low risk of IBD.27 Furthermore, several studies of the impact of dietary factors on MS point towards an impact of meat preservation such as smoking or addition of nitrites.41–44 Studies have also suggested other lifestyle factors such as physical activity,45–47 body mass index (BMI),1 47–49 smoking status50–56 and other dietary factors such as alcohol,1 57–59 dairy products38 and glycaemic index60 61 to be associated with CIDs. Furthermore, studies have shown that the interactions among the microbiota, female sexual hormones and immunity are associated with the development of autoimmune diseases.62
There are many potential mechanisms whereby the environment such as diet may affect the immune system (figure 1). Recently, we provided the hypothesis that intake of high fibre/low red and processed meat may protect against CID.61 Figure 1 presents a model, proposed in a previous study,8 whereby a diet high in meat and low in fibres may impact inflammation.
Aims and hypotheses
Based on previous evidence, we set out to prospectively identify dietary factors that have an impact on disease risk of late-onset CID.
The primary aim of this prospective cohort study is to investigate the impact of fibre, red meat and processed meat on disease risk outcomes of late-onset CID in the ‘Diet, Cancer and Health’ (DCH) cohort. The overall perspectives are that results from this study will be contributing in supporting future diet recommendations for effective personalised prevention of individuals identified to be at high risk.
The main hypothesis is that ‘the risk of late-onset CID will be significantly lower among those with a high fibre/low red meat, and processed meat intake compared with those with a low fibre/high red meat and processed meat intake.’ The hypothesis is illustrated in figure 1.
Based on previous research on a shared aetiology in CIDs, we hypothesise that ‘the suggested association between high fibre/low red meat, and processed meat intake and risk of developing CID is applicable for each of the CID diagnoses’.
The secondary aim of this prospective cohort study is to investigate whether risk of late-onset CID in the DCH cohort is affected by other dietary and lifestyle differences, and therefore to adjust for the potential confounders: age, sex, BMI, physical activity, energy, alcohol intake, intake of fermented dairy products, education after basic school, hormone replacement therapy (HRT), smoking status and comorbidity.
Methods and analyses
Design and setting
This study is an observational study using prospective registry follow-up data. We will use the Danish cohort ‘DCH’, and the follow-up period will be from the date of entry in the DCH cohort (between December 1993 and May 1997) until December 2018.
The DCH study is an ongoing Danish cohort study designed to investigate the relation between diet, lifestyle and disease risk.63 The cohort consists of 57 053 persons, recruited between December 1993 and May 1997. All the subjects (50–64 years of age) gave detailed information on diet (Food Frequency Questionnaire [FFQ]) and other lifestyle data. Questionnaire data on diet and lifestyle were collected at study entry.63 Data from the DCH cohort will be combined with Danish health registries (the National Patient Registry [NPR], the Danish Civil Registration System [CRS], The Danish National Prescription Registry [DNPR] and the Danish Multiple Sclerosis Registry [DMSR]).
Participant characteristics and eligibility criteria
Criteria for inclusion: the population to be studied include participants in the DCH cohort.63 In short, the criteria for invitation to the DCH cohort were: age between 50 and 64, born in Denmark, and no diagnosis of cancer registered in the Danish Cancer Registry. All persons fulfilling these criteria and living in the areas of Copenhagen and Aarhus were invited.63
Criteria for exclusion: participants registered in NPR or DMSR with a CID primary diagnosis from a department with relevant area of specialisation in the period between 1977 and entry to the DCH cohort will be excluded regardless of whether the person receive medical treatment for CID or not.
The Danish health registries
We will extract data from the four national registries in those time periods that is possible for each registry: NPR from 1977 to 2018, CRS from 1977 to 2018, DNPR from 1994 to 2018 and DMSR from 1977 to 2018. An overview of the information obtained from the different registries is presented in table 1.
The NPR will be used to identify patients with CID during follow-up and, in addition, patients with CID before study entry. The NPR contains data on all patients admitted to Danish hospitals since 1977. The register covers both inpatient and outpatient records and indicates the main medical reason for diagnostic procedures or treatment (since 1971 according to the eight version and 1994 according to the tenth version of the International Classification of Diseases [ICD-8 and ICD-10]). ICD-8 and ICD-10 codes will be used to identify cases diagnosed before entry to the DCH cohort study. Information on the departments with the relevant areas of specialisation will be obtained from the NPR and used to identify cases as described in the Data analysis section.
The DMSR will be used to identify patients with MS during follow-up, and in addition, patients with MS before study entry.
The DNPR will be used to obtain information on the medical treatment according to the Anatomical Therapeutic Chemical classification (ATC) code.
From the CRS, we will extract follow-up information on civil status, death and immigration.
Data will be linked by the unique identification number assigned to all residents in Denmark at birth or first immigration which provides a unique opportunity to link information about diagnoses, medications, etc., at the individual level.
Outcome and exposures
Outcome: in this study, the outcome late-onset CID is defined as one of the following diseases: CD, UC, PsO, PsA, RA or MS. This outcome is defined by fulfilling the following two criteria: (1) having the CID disease in NPR (except MS) from a department with relevant area of specialisation, or in the DMSR (MS) during the follow-up period and (2) being treated (irrespective of the number of treatments) with a drug used for the specific disease, and the treatment being registered either in the DNPR or from a department with relevant area of specialisation (criterion 2 will not apply to MS). The outcome late-onset CIDs with the associated ICD-8 and ICD-10 codes and the ATC codes in DNPR and treatment codes in the NPR are specified in table 2. The date and year of the diagnosis are defined as the date and year of the diagnosis in NPR and/or DMSR. If a patient has several CID diseases, only time to the first diagnosis will be included in the analyses.
Exposure and possible confounders: information on exposure in this study is defined as high-fibre intake and high red meat, and processed meat intake, and also other dietary and lifestyle factors.
In the DCH study, information on diet and lifestyle exposure was collected at enrolment using questionnaires as has been described in detail elsewhere.63 In short, in the FFQ diet consumption was assessed in 12 categories of predefined responses, ranking from ‘never’ to ‘eight times or more per day over the past 12 months’. The daily intake was then calculated by FoodCalc.63 Both intake of dietary fibre and red meat and processed meat were measured as continues variable in g/day. The initially collected data on diet and lifestyle exposure of the DCH cohort is used as the baseline information on exposure in this study.
Information on dietary and lifestyle exposures used for this study is specified in table 3. Fibres, red meat and processed meat are defined based on the classification of the food items in the DCH study,63 and with inspiration from other studies using dietary data from the DCH cohort.64–66 Red meat is defined as fresh and minced meat (unprocessed) from beef, veal, pork and lamb and excluding poultry, fish and eggs. Processed meat consists of red meat, poultry and fish items that have undergone processing such as smoking, salting or curing. This includes various kinds of sausages, salami, smoked or cooked ham, poultry or fish, other cold cuts, bacon and liver pate. Poultry includes chicken and turkey both unprocessed and processed, such as various cold cuts of chicken and turkey. Fish includes all unprocessed and processed fish as well as shellfish. Total meat is defined as the total of red (unprocessed) meat, red processed meat together with poultry and fish.64–66 Fibres are defined as fibres from fibrous food items from the FFQ.
Information on possible confounders will also be obtained from the questionnaire data at enrolment in relation to sex, age, education after basic school, BMI, physical activity, energy, HRT, comorbidity, smoking status, alcohol intake and also intake of fermented dairy products. The Charlsons Comorbidity Index67 will be used to classify comorbidity among the study participants. ICD-8 and ICD-10 codes from the NPR will be used to calculate the Charlson score, using the updated Charlson Comorbidity Index.68
Primary exposure variable: the primary exposure variables will be analysed in tertiles. Based on the hypothesis that the risk of CID and late-onset CID will be lower among those with a high fibre/low red meat, and processed meat intake compared to those with a low fibre/high red meat, and processed meat intake, it is expected that
The upper tertile of the sample (33.3% of the total sample), based on the ratio of fibre/meat intake, is associated with lower risk of CID and late-onset CID.
The lower tertile of the sample (33.3% of the total sample) with respect to intake of red meat and processed meat and the upper tertile of the sample (33.3% of the total sample) with respect to intake of dietary fibres are independently associated with lower risk of CID and late-onset CID, and a potential interaction between them may further lower the risk of CID and late-onset CID.
Other (exploratory) exposure variables:
Other dietary and lifestyle factors independently or combined will be analysed, and are presented in table 3.
Statistical analysis plan
The data obtained from this study will be used to investigate our ability to predict risk of late-onset CID, based on whether a diet high in fibre and low in red meat, and processed meat is a predictive factor. Furthermore, data on other lifestyle factors and dietary factors obtained from the FFQ will be used to investigate whether these factors potentially could give rise to confounding of the association between fibre/meat intake and risk of late-onset CID.
Descriptive analyses for categorical variables will be presented as frequencies, and differences between late-onset CID cases and non-cases will be evaluated by χ2 test. Continuous variables will be tabulated as medians (with quartiles, Q1 and Q3) and non-parametric tests on the equality of medians will be used to test for differences between groups. P values below 0.05 will be considered statistically significant. Negative binomial regression will be applied to calculate incidence rates per 1000 patient years and incidence rate ratios between exposures. To investigate the risk of and time to CID diagnosis, the Fine-Gray competing risk regression model will be applied, and thus competing risk of death will be taken into account while handling emigration as censoring and reporting cumulative incidences and sub-HRs and the corresponding p values for late-onset CID associated with specified food substitutions. Regressions will be carried out as both crude regressions, only including the exposure, and adjusted for age, sex, lifestyle factors and selected comorbidities. Sensitivity analyses of the outcome variable and use of medication will be carried out.
Logistic regression will be conducted to examining the association between fibre/meat intake and already being diagnosed with a CID diagnose. Analyses will be conducted using Stata V.15.69
If power of each of the disease groups allows, there will be conducted subanalyses on each of the late-onset CID diagnoses separately with the overall aim of testing if the hypothesised association between diet factors and development of late-onset CID is applicable for all the late-onset CIDs or if some of the late-onset CIDs deviate from this association. Moreover, analyses omitting MS from the overall definition of CID will be performed.
Information on the departments with the relevant areas of specialisation will be used to evaluate the robustness of the diagnosis codes from the NPR and to identify the cases as we will only accept cases that have a diagnosis from a relevant department. The DMSR is highly validated,70 and data from the registry will be used to ensure complete follow-up of valid diagnoses of MS.
Power considerations, late-onset CID: the age-adjusted (age between 45 and 69) incidences per 100 000 years of the CIDs in Denmark are 10.1–10.6 for CD71; 27.5–32.7 for UC71; 200–225 for PsO15; 28–35 for PsA16; 4.4–7.5 for MS19 72–75 and 28.1–57.6 for RA.76 77 In the cohort of approximately 57 000 participants and a follow-up period of approximately 20 years, these incidences correspond to an estimate of approximately 2000 cases of CID in the cohort. To estimate the power of the study, we therefore expect approximately 2000 cases of CID, that is, an incidence of about 3% and about one-third of the cohort in the low risk and two-thirds in the high-risk exposure. Assuming a fixed follow-up of 20 years and a relevant effect size of 2% CID risk in the low-risk exposure compared with 3% in the high-risk exposure, this would require 11 667 participants to obtain a power of 0.90 in an (unadjusted) χ2 test. Hence, with a total cohort size of approximately 57 000, this should ensure a sufficient power even when taking loss to emigration and competing risk of death (resulting in less than 20-year follow-up for some participants) into account.
Patient and public involvement
This study is register based and there will be no patients or public involvement in the study.
Strengths and limitations of the study
A strength of this study is that it is not limited to one disease, as it includes several CIDs. Another strength is that the linkage to Danish health registries will ensure almost complete follow-up of the study population, as the Danish health registries are considered the internationally most comprehensive with high validity.78–81 Furthermore, in this study, we have chosen very restrictive criteria for defining the late-onset CID cases by requiring that cases fulfil both diagnostic and treatment criteria, or that the cases were registered in the DMSR. This approach ensures that a high proportion of the identified cases really had late-onset CID. Therefore, some ‘real’ CID cases might not have been identified, hence lowering the sensitivity.70 82 83
A possible limitation of this study could have been the validity of the diagnostic codes from the NPR and the treatment codes from the DNPR as criteria for identifying late-onset CID cases.70 82 But, as described in the analysis plan, this potential limitation will be sought eliminated by using the information on the hospital departments, and for MS, using the DMSR for identifying MS cases. In addition, usually dietary habits do not change much during life. Therefore, the exposure time to the diet is long in 50+age group and a possible impact of diet is possibly located much earlier in life. Therefore, we have included an analysis among those who at entry to the DCH cohort already had a CID diagnosis to examine if their low/high intake of dietary fibre, red meat and processed meat is associated with having CID. We are well aware that such an analysis might be impacted by bias by indication and that the results should be interpreted with this in mind.
The FFQ applied in the present study has been used in the large, European prospective cohort study ‘The European Prospective Investigation into Cancer and Nutrition’ (EPIC),84 85 and it has been used and evaluated in the Danish population, with results that demonstrate consistency.86–88 However, the FFQ is not without limitations, with respect to the lack of information on portion sizes89 90 or to underestimation and overestimation of intake of unhealthy and healthy food.91 Any imprecision of the FFQ due to standardised portion sizes or incorrect reporting of food intake, will lead to large CIs, which potentially can lead to null results.8 On the other hand, studies suggest that specification of a standard portion size may not introduce a large error in the estimation of food and nutrient intake.92 Furthermore, the FFQ has been validated as being appropriate for use in studies that examine relationships between diet and risk of disease.91
Another limitation of this study is the validity of the information on exposure as this information is collected at study entry, which potentially can be several years before any outcome appears, and measurement error in the FFQ may occur if the participants change their diet over time. Furthermore, there is a potential risk of recall bias according to the information on exposure, as this information relies on the participants’ ability to recall their dietary intake. In this study, however, it is expected that dietary and lifestyle patterns among adults are relatively stable over time, based on other longitudinal studies that showed minimal temporal changes.93 94
The MET score is an accepted standard measure of physical activity.95 However, the weaknesses of the MET score include a risk of adding random variation by applying an assumed intensity to include activities. Another weakness of the MET score is the implicit assumption that the intensity aspect of physical activity is important for the development of disease.95
The study population in this study is based on a cohort of middle-aged women and men living in urban areas. This could reduce the generalisability of study findings, as the incidence of CID diagnoses could be different for younger persons and persons from rural areas.
The disease groups may be heterogeneous regarding dietary and lifestyle factors. This study may not capture every dietary and lifestyle difference between the disease groups, due to insufficient power of each disease group, as described in the analysis plan. Replication of the results in other well-characterised populations using prospectively sampled dietary data will minimise the risk of potential type 2 errors. Study results should preferably be replicated in cohorts in other countries and other age groups for further evaluation of the robustness of the results.
This registry study is a cross-disciplinary collaboration that includes clinical specialists within neurology, dermatology, rheumatology, IBDs, prospective cohort study design and clinical registries.
We anticipate that the PROCID-DCH study will reveal factors of importance, including whether the diet is likely to interfere with the disease risk of late-onset CID.
The perspective is that significant results from this study will be sought replicated in other cohorts such as the EPIC-IBD84 and UK biobank96 with high-quality prospective lifestyle data. Successful replication indicates the robustness of the findings which is an important step on the road to developing clinical tools for effective personalised prevention of individuals at high risk.
Dissemination of results to the public and scientifically
Target journals include international journals within internal medicine. In addition to the scientific reporting of results, major findings with translational implications will also be communicated to categories of both health professionals, and targeted stakeholders including public health policy-makers, and to the general public through various media and news activities. Intellectual Property Rights to discoveries based on the outlined research belong to the University of Southern Denmark.
Contributors VA conceived the presented idea and achieved funding. AT (diet and environmental factors, epidemiology) designed the cohort, collected the data and made the data available for the present project. NFR, KHR, MS and VA wrote the first draft and contributed to the further preparation and final adjustments of the protocol. According to their respective specialisation, all other authors, ES (multiple sclerosis), MLH (rheumatology), BG (rheumatology) and AB (psoriasis) contributed to the project. All authors discussed the results and contributed to the final manuscript.
Funding This work is supported by Region of Southern Denmark grant number [17/33849] (Vibeke Andersen), and the Danish Rheumatism Association (R150-A4434-B760, Vibeke Andersen).
Competing interests BG declares to have received research funding from Abbvie, Biogen, Pfizer. MLH declares to have received research funding from BMS, MSD, Pfizer, Biogen, Samsung, CellTrion, Lilly and Novartis. AB has participated in the development of educational material for Biogen. All other authors declare no conflict of interest.
Ethics approval This is an open register-based cohort study. The study does not need approval from the local Ethics committee or Institutional Review Board by Danish law. The study was approved by the Danish Data Protection Agency (2012-58-0018).
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent for publication Obtained.