Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Department of Internal Medicine and Clinical Nutrition Institute of Medicine Sahlgrenska Academy, University of Gothenburg Gothenburg 2022 Nutritional impact on health in patients with Rheumatoid Arthritis © Erik Hulander 2022 erik.hulander@gu.se ISBN 978-91-8009-825-0 (PRINT) ISBN 978-91-8009-826-7 (PDF) “Qui se ultro morti offerant facilius reperiuntur quam qui dolorem patienter Printed in Borås, Sweden 2022 NENMÄRKVA E ferant” Printed by Stema Specialtryck AB Trycksak3041 0234 - Julius Caesar, 100-44 BCE Commentarii de bello Gallico, VII.77 S T Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Department of Internal Medicine and Clinical Nutrition, Institute of Medicine Sahlgrenska Academy, University of Gothenburg Gothenburg, Sweden ABSTRACT Objective: Rheumatoid Arthritis (RA) is the most common autoimmune rheumatic disease, affecting around 0.5-1% of the population. The aim of this thesis was to study dietary impact on markers of health in patients with RA. Methods: Data from the randomized controlled crossover trial Anti- inflammatory diet in Rheumatoid Arthritis (ADIRA) is used. The trial compares a Mediterranean-like diet intervention with a typical western diet in patients with RA (n = 47). Additionally, cross-sectional analyses were done on data obtained at screening pooled from the ADIRA-trial and a postprandial meal challenge trial in patients with RA (n = 30). Results: In the ADIRA-trial, apolipoprotein-B100/A1 ratio was improved, high density bound cholesterol increased and triglycerides decreased in the intervention compared to the control. Proinflammatory chemokines decreased compared to control, as well as erythrocyte sedimentation rate in participants with high compliance and no major medication changes. Body composition improved over time during both the intervention and the control diet periods. Developments in nutritional quality differed between the intervention and control diet periods, indicating a successful implementation of the dietary regimens. There was no relation between habitual nutritional quality and health outcomes in a pooled cross-sectional analysis. Conclusions: Comparing a Mediterranean-like diet to a typical western diet, dietary intake improved cardiovascular risk profile, and in a per protocol analysis, reduced inflammation. Further studies in more diverse populations are required to determine effects on long-term health outcomes. Keywords: dietary intervention, rheumatoid arthritis, health ISBN 978-91-8009-825-0 (PRINT) ISBN 978-91-8009-826-7 (PDF) SAMMANFATTNING PÅ SVENSKA Bakgrund: Reumatoid artrit (RA) är den vanligaste autoimmuna reumatiska sjukdomen, och påverkar cirka 0.5-1% av befolkningen. Även om farmakologisk behandling förbättrats avsevärt de senaste årtionden, är upprätthållen remission över tid inte vanligt. Patienter med RA efterfrågar ofta kostråd, men det saknas i dagsläget evidens för någon specifik nutritionsbehandling vid RA. Metoder: Den randomiserade och kontrollerade crossover-studien, Antiinflammatorisk Diet vid Reumatoid Artrit (ADIRA) ligger till grund för avhandlingen. Studien utvärderar effekten av en medelhavskost-liknande intervention gentemot en typisk västerländsk kost på patienter med RA (n = 47). I en tvärsnittsanalys med screeningdata från ADIRA samt en postprandiell måltidsstudie (n = 30) studeras dessutom samband mellan kostkvalitet och hälsoutfall. Resultat: I ADIRA-studien förbättrades apolipoprotein-B100/A1-kvoten, högdensitet lipoprotein-bundet kolesterol ökade och triglycerider minskade av intervention jämfört med kontroll. Proinflammatoriska kemokiner minskade av intervention jämfört mot kontroll, och likaså sänkan minskade hos patienter med hög följsamhet och utan större läkemedelsändringar. Kroppssammansättningen förbättrades oberoende av kostbehandling. Utveckling i kostkvalitet skiljde sig däremot åt mellan intervention- och kontrollkost, vilket tyder på en framgångsrik implementering av respektive kost. I den poolade tvärsnittsanalysen sågs inga signifikanta samband mellan habituell kostkvalitet och hälsoutfall. Slutsatser: En medelhavskost-liknande intervention förbättrade riskprofil för hjärt-kärlsjukdom och minskade inflammation i en per-protokoll-analys jämfört med en västerländsk kontrollkost. Ytterligare studier krävs i en bredare patientpopulation och under längre tid för att utvärdera långsiktiga hälsoeffekter. LIST OF PAPERS This thesis is based on the following studies, referred to in the text by their Roman numerals. I. Hulander E, Bärebring L, Turesson Wadell A, Gjertsson I, Calder PC, Winkvist A, Lindqvist HM. Diet intervention improves cardiovascular profile in patients with rheumatoid arthritis: results from the randomized controlled cross-over trial ADIRA. Nutr J. 2021 Jan 23;20(1):9. doi: 10.1186/s12937-021- 00663-y. II. Hulander E, Bärebring L, Turesson Wadell A, Gjertsson I, Calder PC, Winkvist A, Lindqvist HM. Proposed Anti-Inflammatory Diet Reduces Inflammation in Compliant, Weight-Stable Patients with Rheumatoid Arthritis in a Randomized Controlled Crossover Trial. J Nutr. 2021 Dec 3;151(12):3856-3864. doi: 10.1093/jn/nxab313. III. Hulander E, Lindqvist HM, Wadell AT, Gjertsson I, Winkvist A, Bärebring L. Improvements in Body Composition after a Proposed Anti- Inflammatory Diet Are Modified by Employment Status in Weight-Stable Patients with Rheumatoid Arthritis, a Randomized Controlled Crossover Trial. Nutrients. 2022 Mar 2;14(5):1058. doi: 10.3390/nu14051058. IV. Hulander E, Lindqvist HM, Turesson Wadell A, Gjertsson I, Winkvist A, Bärebring L. Associations between nutritional quality of habitual diet, concurrent health characteristics and response to a dietary intervention in patients with rheumatoid arthritis Manuscript. i CONTENT ABBREVIATIONS ............................................................................................. VI 1 RHEUMATOID ARTHRITIS............................................................................ 1 1.1 Prevalence ............................................................................................. 1 1.2 Classification ......................................................................................... 1 1.3 disease activity ...................................................................................... 4 1.4 Malnutrition and body composition ...................................................... 6 1.5 Cardiovascular disease .......................................................................... 9 1.5.1 Blood lipid transportation .............................................................. 9 1.5.2 CVD Risk factors ........................................................................ 11 1.6 Inflammatory markers ......................................................................... 13 2 TREATMENT OF RA .................................................................................. 16 2.1 Dietary intake ...................................................................................... 16 2.1.1 Diet and CVD risk factors in RA ................................................ 17 2.1.2 Diet and inflammation in RA ...................................................... 18 2.1.3 Diet and body composition in RA ............................................... 18 2.1.4 Need of further studies ................................................................ 18 AIM ................................................................................................................ 21 3 METHODS ................................................................................................. 22 3.1 Data sources ........................................................................................ 22 3.2 The Anti-inflammatory Diet In Rheumatoid Arthritis (ADIRA) trial. 22 3.2.1 Study design & participant selection in the ADIRA-trial ............ 22 3.2.2 Dietary intervention ..................................................................... 26 3.2.3 Outcomes from the ADIRA-trial included in this thesis ............. 30 3.2.4 Contribution to the ADIRA-trial by the PhD-candidate .............. 32 3.3 The Postprandial Inflammation in Rheumatoid Arthritis (PIRA) ....... 33 3.3.1 Meal composition ........................................................................ 33 3.3.2 Recruiting participants with RA .................................................. 34 3.3.3 Screening of participants with RA .............................................. 34 ii iii 5.5.2 Participants .................................................................................. 58 3.3.4 Recruiting healthy controls ......................................................... 35 5.5.3 Recruitment ................................................................................. 59 3.3.5 Postprandial meal challenge for patients with RA ...................... 35 5.5.4 Statistical considerations ............................................................. 60 3.3.6 Postprandial meal challenges for healthy controls ...................... 37 5.5.5 Measurements of outcomes ......................................................... 61 3.3.7 Outcome measures in the PIRA-trial ........................................... 37 5.6 Conclusion & future perspectives ....................................................... 62 3.3.8 Outcomes from the PIRA-trial included in this thesis................. 37 ACKNOWLEDGEMENT .................................................................................... 64 3.3.9 Contribution to the PIRA-trial by the PhD-candidate ................. 37 REFERENCES .................................................................................................. 66 3.4 Common methodology in the ADIRA- and PIRA-trials ..................... 38 APPENDIX ...................................................................................................... 78 3.4.1 Assessment of nutritional quality ................................................ 38 3.4.2 Assessment of disease activity .................................................... 38 3.4.3 Assessment of body composition ................................................ 38 3.4.4 Assessment of physial activity .................................................... 38 3.5 Differing methods in the ADIRA- and PIRA-trials ............................ 40 3.6 Statistical analyses .............................................................................. 41 3.6.1 Carry-over effects ........................................................................ 42 3.6.2 Confounder analyses ................................................................... 42 3.6.3 Interaction analyses ..................................................................... 43 3.6.4 Per protocol and intention to treat analysis ................................. 45 4 RESULTS ................................................................................................... 46 4.1 Study population ................................................................................. 46 4.2 Paper I ................................................................................................. 47 4.3 Paper II ................................................................................................ 47 4.4 Paper III .............................................................................................. 49 4.5 Paper IV .............................................................................................. 49 5 DISCUSSION .............................................................................................. 52 5.1 cardiovascular risk, Paper I ................................................................. 52 5.2 inflammation, Paper II ........................................................................ 54 5.3 body composition, Paper III ................................................................ 55 5.4 nutritional quality index, Paper IV ...................................................... 56 5.5 The ADIRA study design .................................................................... 58 5.5.1 Dietary intervention foods ........................................................... 58 iv v IDL Intermediate density lipoprotein ABBREVIATIONS IGF-1 Insulin-like growth factor-1 IL Interleukin ACR American College of Rheumatology LDL Low density lipoprotein ADIRA Anti-inflammatory diet in rheumatoid arthritis LDL-C Low density lipoprotein-bound cholesterol anti-CCP Anti-cyclic citrullinated peptide NF-κB Nuclear factor kappa-light-chain-enhancer of APO-A1 Apolipoprotein-A1 activated b-cells APO-B Apoliprotein-B NMR 1H Nuclear Magnetic Resonance APO-B100 Apoliprotein-B100 Non-HDL-C Non high density lipoprotein-bound cholesterol APO-B48 Apolipoprotein-B48 NRF Nutrient rich foods index BCAA Branched chain amino acids PBMC Peripheral blood mononuclear cells bDMARD Biological disease modifying anti-rheumatic drugs PIRA Postprandial inflammation in rheumatoid arthritis BIS Bioelectrical impedance spectroscopy PREDIMED Prevención con dieta mediterránea BMI Body mass index RA Rheumatoid arthritis CETP Cholesterol ester transfer protein RBC Red blood cells CRP C-reactive protein RDI Recommended daily intake csDMARD conventional synthetic disease modifying anti- RF Rheumatoid factor rheumatic drugs SJC Swollen joint count CVD Cardiovascular disease SRQ Swedish Rheumatology Quality Register CXCL C-X-C motif chemokine ligand TG Triglycerides DAS28 28-joints disease activity score erythrocyte TJC Tender joint count sedimentation rate TNF-α Tumor necrosis factor alpha DMARD Disease modifying anti-rheumatic drugs VAS Visual analog scale DNA deoxyribonucleic acid DXA Dual energy X-ray ESR Erythrocyte sedimentation rate EULAR European League Against Rheumatism FAACT Functional assessment of anorexia/cachexia therapy FFM Fat free mass FFMI Fat free mass index FM Fat mass FMI Fat mass index GH Global health HAQ Health assessment questionnaire disability index HbA1c Glycated hemoglobin HDL High density lipoprotein HDL-C High density lipoprotein-bound cholesterol vi vii Erik Hulander 1 RHEUMATOID ARTHRITIS 1.1 PREVALENCE Autoimmune diseases affect approximately 4-5% of the global population, and rheumatoid arthritis (RA) is the most prevalent autoimmune rheumatic disease (20). Around 0.5-1% of the global population is expected to be directly affected by RA (21). Symptoms of RA are typically pain, swelling, and a reduced function in peripheral joints (22). During continued disease activity over time, joint destruction often occurs, further diminishing the functional capacity of afflicted individuals. Genetic risk factors, female sex, smoking and advanced age contribute to the development of RA (22). Recent large scale cohort studies have also linked obesity to a higher risk of developing RA (23-25). 1.2 CLASSIFICATION Today, diagnosis of RA is based on the 2010 American College of Rheumatology (ACR) and European League Against Rheumatism (EULAR) classification criteria (26). In brief, when a person presents with at least one inflamed joint not explained by any other diagnosis, without radiographic results typical for RA, a scoring system is applied. This is based on the number of tender or swollen joints (Figure 1) in combination with symptom duration and biochemical markers (Figure 2). A score of six or above classifies the disease as RA. The 2010 ACR and EULAR classification has a higher sensitivity, and thus captures more patients, than the previously used 1987 ACR criteria (27). 1 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Figure 1. In total, 30 joints are examined for classification of RA according to 2010 EULAR/ACR criteria, shown here with red circles. 3 2 Figure 2. Classification criteria of RA 1Cut off values for heightened CRP and ESR is determined by local standards. Abbreviations: anti-CCP; Anti-cyclic citrullinated peptide; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; RF, Rheumatoid Factor. Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 1.3 DISEASE ACTIVITY In patients with RA, the disease activity often varies over time and in order to assess disease activity, several indexes have been proposed (28, 29). Typically, a combination of the patient’s perceived health and an assessment of joint status coupled with biochemical markers of inflammation are used. The most common and accepted measure of disease activity is the 28-joints disease activity score erythrocyte sedimentation rate (DAS28). The DAS28-index is based on an assessment of how many of 28 prespecified joints that are swollen and/or tender, erythrocyte sedimentation rate (ESR)-levels and a patient- reported global health assessment on a visual analog scale (VAS) (range 0-100 mm) (Figure 3). The symptoms of RA has a major impact on quality of life on inflicted individuals, who score lower compared to the general population (as assessed by the 36-item short form health survey (30)) (31). Not only is the general quality of life reduced, disease activity and comorbidities also reduce the life span. The main causes of death appear to be similar between patients with RA and the general population (i.e. cardiovascular-, oncological and respiratory diseases), but disease development is accelerated and, overall, the life expectancy is reduced (32). Managing patients with RA is further costly; the main societal economic burden of RA is related to early retirement and sick Figure 3. In total, 28 joints are examined to assess disease activity in patients with leaves (33), this holds true also today, even though pharmacological treatment RA, shown with red circles here has improved tremendously during the past decades with the broad application of biological disease modifying drugs (bDMARD) (34). 5 4 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 1.4 MALNUTRITION AND BODY COMPOSITION malnutrition, with inflammatory stimuli, is commonly classified as cachexia, and is characterized by loss of lean mass (36). Inadequate nutritional intake, or inadequate dietary quality, in health and The prevalence of cachexia in patients with RA, commonly referred to as disease, may lead to loss of weight and muscle mass, a term referred to as rheumatoid cachexia, is somewhat unclear due to a diversity in diagnostic malnutrition. The diagnostic criteria for malnutrition are based either on a low criteria (Table 2). Between reports, the prevalence of rheumatoid cachexia body mass index (BMI), or a combination of weight loss and low fat free mass varies between 1% to 54% depending on methodology and patient population index (FFMI) (35) (Table 1). (37). Results from a meta-analysis indicate that, if using the most commonly used criteria (FFMI below 10th percentile in combination with a FMI above 25th percentile of a reference population), it appears to have affected about a third of patients with RA (37). Table 1. Diagnosis of malnutrition in patients identified at risk by nutritional screening It is not self-evident why cachexia appears in patients with RA. Explanations that have been put forward are related to joint impairment that reduces physical Either BMI <18.5 kg/m2 activity and promotes catabolism and fat accumulation, a chronic Or Weight loss >10%, or 5% inflammatory response that induces catabolism, side effects of glucocorticoids within 3 months as well as an unhealthy diet in general. Further, incidence of RA is increased in the elderly, in particular in middle aged and older females (38). With And either: advanced age, muscle mass and physical functions often decline regardless of BMI < 20 kg/m2 if age < 70 underlying disease, a phenomenon referred to as sarcopenia (39). There is thus years an overlap between these definitions; many cachectic patients are sarcopenic, BMI < 22 kg/m2 if age > 70 but sarcopenic patients are not necessarily cachectic. years In patients with RA, unfavorable body composition (FFMI <25th percentile and a higher than average FMI) is related to higher total serum cholesterol and Or: 2 oxidized low density lipoprotein (LDL) (40), and lower volume of lean mass FFMI < 15 kg/m for is related to higher levels of biomarkers of inflammation (41). Data from females inpatients in a Swedish hospital has shown that BMI is not a reliable proxy for FFMI < 17 kg/m2 for males body fatness in patients with RA (42). Despite a higher or equal BMI compared Data from the European society for clinical nutrition and metabolism 2015 consensus to the general population, patients with RA presented with lower lean mass, statement on the diagnostic criteria for malnutrition (35). indicating a need to assess body composition directly. Abbreviations: BMI, Body mass index; FFMI, Fat free mass index. Dietary intake as well as the metabolism of nutrients can be affected by an individual´s physiology, and many illnesses create a different metabolic environment, a phenomenon commonly referred to as disease related malnutrition. This is perhaps most studied in oncology, where both the disease itself and side effects of treatment can affect dietary intake and metabolism of nutrients. Disease related malnutrition can further be classified into conditions with inflammatory stimuli, or non-inflammatory states. Disease related 6 7 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Table 2. Criteria proposed and used to diagnose rheumatoid cachexia Bokhorst et al., (43). Engvall et al., Elkan et al., 1.5 CARDIOVASCULAR DISEASE (41). (40). weight loss ≥ 5% within 12 months FFMI < 10th FFMI < 10th 1.5.1 BLOOD LIPID TRANSPORTATION and ≥ 3 of the following: percentile and percentile and FMI > 25th FMI > 50th Lipids are transported in the circulation in small particles (diameter about 5 –  Low handgrip strength percentile percentile 1200 nm) with a hydrophilic outside and lipophilic inside called lipoproteins. The lipoproteins are composed of proteins that give structural and functional (< 3rd tertile compared properties called apolipoproteins as well as phospholipids, cholesterol (in the to reference) form of cholesteryl esters and free cholesterol) and triglycerides (TG).  Low FFMI (< 10th percentile compared to There are different distinct classes of lipoproteins with different functions, a reference) simplified schematic explanation is presented in figure 4. Chylomicrons carry  Low appetite (FFACT the apolipoprotein-B48 (APO-B48) and transport dietary fat from the intestines questionnaire or VAS to tissues and the liver (mainly TG), the remaining lipids (cholesterol and scale <50) phospholipids) are transferred to high density lipoprotein (HDL) particles. The  Fatigue (VAS >50) chylomicron remnant is then cleared from the circulation by the liver. The main structural protein for HDL-particles, apolipoprotein-A1 (APO-A1), is  Biochemical markers produced both in the liver and in the intestines. The primary function of HDL (ESR >10 mm/1h, particles is to collect and transport lipids back to the liver. CRP >8 mg/L or anemia) The liver produces TG-rich very low density lipoprotein (VLDL)-particles that primarily transport TG to tissues. Gradually, by unloading lipids, the VLDL- particle is transformed to intermediate density lipoprotein (IDL) and Abbreviations: CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; subsequently to low density lipoprotein (LDL). The LDL-particle is taken up FAACT; Functional assessment of anorexia/cachexia therapy; FFMI, Fat free mass either in tissue or in the liver. The VLDL-, IDL- and LDL-particles shares the index; FMI, Fat mass index; VAS, Visual analog scale. same basic structural protein, apoliprotein-B100 (APO-B100). VLDL-, IDL-, LDL-particles, and to a lesser extent chylomicrons, are considered to be atherogenic. HDL-particles on the other hand are considered to be atheroprotective. Some investigations simply quantify apolipoprotein-B (APO-B) as measure of both APO-B48 and APO-B100. In practice, in the fasted state, APO-B concentration primarily reflect the APO-B100 concentration. Elevated blood lipids, measured as either TG, total cholesterol or LDL-bound cholesterol (LDL-C) are considered to be important risk factors for future cardiovascular disease (CVD). 8 9 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander low density lipoprotein; LDL-R, Low density lipoprotein receptor; LPL, Lipoprotein lipase; LRP, LDL receptor-related protein; SR-B1, Scavenger receptor, class B type 1; TG, Triglycerides; VLDL, Very low density lipoprotein. Illustration is based on images from Servier Medical Art, licensed under a Creative Commons Attribution 3.0 Unported License. 1.5.2 CVD RISK FACTORS CVD is the most common cause of death in patients with RA (32). The most important modifiable risk factors for atherosclerotic CVD are high circulatory APO-B concentration, elevated blood pressure, smoking and diabetes (44). Chronic inflammation further induces CVD progression, and a relative risk increase of 50% compared to the general population is recommended to assume, when assessing risk of CVD in patients with RA (45). Current recommendations are to assess cardiovascular risk at least once every 5 years in patients with RA, and to measure blood lipid concentrations (45). The main dietary component affecting elevated blood pressure is sodium intake, thus the total salt intake is recommended not to exceed 5 g/day. Additionally, increasing the intake of potassium appears to lower blood pressure (46). The concentration of LDL particles has long been described as an independent causal risk factor for CVD (47). A reduction in LDL-C has shown to decrease the risk of CVD irrespective of treatment alternative (48). In patients with RA, the lipid metabolism deviates somewhat compared to the general population; inflammation, acute and chronic, lowers the concentration of cholesterol in the circulation (49) (Figure 5). As documented by Myasoedova et al. (50), a relative decrease in LDL-C and total cholesterol can be seen even in the years leading up to a diagnosis of RA. Since most cardiovascular risk assessments focus on lipid levels, this can lead to contradictory findings, where patients with a heightened inflammatory state present with low lipid levels. Normalization of inflammation, often achieved by disease modifying anti- rheumatic drugs (DMARD), typically result in an increase in cholesterol levels (Figure 5). Elevated TG concentration is another established risk factor for future cardiovascular disease. Lowering of TG concentration in blood infers a lower Figure 4. Schematic overview of the blood lipid and lipoprotein metabolism risk for future CVD events, although the effect is not as clear as for lowering LDL-C (51). Both lowering TG and LDL-C are associated with a reduced risk A1, Apolipoprotein A1; B-100, Apolipoprotein B-100; B-48, Apolipoprotein B-48; of cardiovascular event, but data suggest that this risk decrease is mediated by CETP, Cholesteryl ester transfer protein; HDL, High density lipoprotein; IDL, APO-B concentrations (52). Recent data further suggest that, when adjusting Intermediate density lipoprotein; LCAT, Lecithin–cholesterol acyltransferase; LDL, for APO-B, other lipid parameters no longer significantly predict future 10 11 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander cardiovascular events (53). Thus, when possible, CVD risk assessment benefits by including apolipoprotein concentrations. 1.6 INFLAMMATORY MARKERS While LDL-C has traditionally been the most important marker, recent recommendations point to include non HDL bound cholesterol (non-HDL-C) Patients with RA are characterized by a state of elevated systemic chronic (essentially a proxy for APO-B carrying lipoproteins) as a potentially superior inflammation. Current guidelines recommend quantifying C-reactive protein risk predictor compared to LDL-C (54). (CRP) and ESR as clinical markers of inflammation in patients with RA. CRP is produced and secreted primarily from the liver, and is a pentameric protein composed of five identical monomers. CRP was discovered as a substance (Fraction C) that increased during infection and reacted with the antigenic polysaccharide of streptococcus pneumoniae (55). The half-life of CRP in plasma is about 19 hours, following a tissue damage or infection. CRP production can increase rapidly and typically peaks at 48 hours post injury. CRP is described as having both anti-inflammatory and pro-inflammatory effects; in parts this is hypothesized to depend on the conformation, where pentameric CRP is suggested to dissolve into monomeric form in local tissue, and thereby potentiate the inflammatory response. Most assays, however, do not clearly distinguish between pentameric and monomeric CRP, consequently CRP is commonly interpreted as a general marker of inflammation. ESR is a long-standing standardized test that together with CRP is the recommended clinical measure to assess acute phase reactants in patients with RA. ESR is an indicator of the relative density between red blood cells (RBC)s and the surrounding plasma, measuring how far the RBCs condense to the bottom of a test tube containing whole blood and an anticoagulant. As recently reviewed (56), RBCs have a negatively charged surface and therefore naturally repel each other. Many plasma proteins have a positive charge, and the increase of plasma proteins alleviate the electrical charge and thereby allow erythrocytes to aggregate and form a rouleaux formation, which allows for a Figure 5. Blood lipids decrease under concurrent inflammation. Anti-inflammatory quicker sedimentation. Among plasma proteins, fibrinogen and treatment that dampen the inflammation, inversely affects the blood lipid levels. immunoglobulins have been pointed out as a specifically important factors in Image redrawn and adapted from Choy et al. (49). this rouleaux formation. The sedimentation rate is further positively correlated to the concentration of RBCs in the blood, and negatively correlated to the plasma transfer protein albumin. Hyperlipidemia is also known to increase the ESR. ESR is therefore appropriately viewed upon as a non-specific marker of inflammation with limited diagnostic value, and with a reference range that varies according to sex and age (Table 3). Importantly, deviations from an individual’s habitual ESR-value during healthy conditions may confer more information than an individual measurement (57). 12 13 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Table 3. Reference range of ESR by age and sex lymphocytes, fibroblasts) as well as endothelial cells and platelets (61). Often, the chemokine receptors are non-specific and can bind to several different Age <50 years Age >50 and <70 years individual chemokines (58). Females1 <21 mm/1h <30 mm/1h Males <13 mm/1h <20 mm/1h 1Contraceptives and pregnancy produce values to around 30 mm/1h (57). The intercellular communication, and thus the response of the immune system, is further regulated by signaling-proteins called cytokines. Cytokines are typically categorized either by their function, or by their structural similarity. Some basic functional classes of cytokines are interferons (antiviral proteins) chemokines (proteins that direct movement of immune cells), anti- inflammatory cytokines (ex. interleukin(IL)-10, transforming growth factor- β), and pro-inflammatory proteins (ex. IL-1, IL-6, tumor necrosis factor α (TNF-α)). However, many cytokines are pleiotropic (i.e., have several different functions) and can act both in an endocrine, autocrine or paracrine manner and be involved in a multitude of processes. (58) Two cytokines with a central place in the treatment of patients with RA is IL- 6 and TNF-α. IL-6 is secreted in a wide range of immune cells as well as endothelial and epithelial cells and is a pleiotropic protein (59), but is generally viewed upon as a proinflammatory molecule. TNF-α is produced mainly by macrophages and T-cells, but many other cells also produce this molecule, such as neutrophiles, B-cells, NK-cells, monocytes, dendritic cells and also epithelial and endothelial cells (59). TNF- α exist both as a membrane bound protein and in a soluble form in the circulation with a half-life of around 14 minutes (59). TNF-α is generally considered a proinflammatory marker that controls apoptosis and promotes inflammation, often explained by its effect on the transcriptor factor nuclear factor kappa-light-chain-enhancer of activated b-cells (NF-κB) (59). Inhibition of TNF-α, by antibody therapy, is a common and often effective pharmacological treatment in patients with RA. Primarily IL-6, but also TNF- α, induce expression of CRP in liver cells (60). Chemokines are further a subclass of soluble cytokines that share structural similarities, the primary function of chemokines is to direct to movement of immune cells, who move towards a higher concentration (58). Primary source of chemokines are immune cells (macrophages, dendritic cells, T- 14 15 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 2 TREATMENT OF RA Pharmacological treatment of patients with RA has improved dramatically during recent decades with the arrival of potent anti-inflammatory bDMARD such as TNF- and IL-6 inhibitors. Treatment with bDMARDS are often applied in combination with metotrexate which is also overall the most common treatment (62). Patients responding to bDMARDs substantially reduce their cardiovascular risk profile (63). But not all patients are responders, and non-responders can present a residual risk of cardiovascular incidence up to double that of the general population (63). Furthermore, remission over time is a rare occurrence; data suggest that only about one in four patients with RA retains remission over a 12-month period (64). While the potential impact of smoking cessation and physical activity on health outcomes for patients with RA is clear, there is no evidence for any specific dietary treatment (45). An overall healthy diet and lifestyle is recommended. 2.1 DIETARY INTAKE Figure 6. The omega-3 fatty acid eicosapentaenoic acid act competitively against There is a high prevalence of belief among patients with RA that dietary intake arachidonic acid and produce less potent proinflammatory molecules. Figure redrawn and adapted from Tosi et al. (76). can affect the disease severity (65-68). Dietary changes, instigated without professional medical advice, appear to be common (69). Following a strict and Fiber intake is another dietary component likely to have physiological effects. unsupervised diet may however result in nutritional deficiencies, thereby For example, an adequate fiber intake is indicated to promote satiety, intestinal leading to a further diminished nutritional status. There are several indications epithelial integrity, improved glucose metabolism as well as lowering the that diet matters; higher intake of fish (70) and unsaturated fats (71) have been circulating cholesterol concentration (77). associated to lower disease activity in patients with RA. Likewise, overall diet quality is inversely associated with inflammation in patients with RA (72). 2.1.1 DIET AND CVD RISK FACTORS IN RA There are several nutrients with proposed immunomodulatory effects. Omega- Aside from an adequate fiber and omega-3 fatty acid intake, several other 3 fatty acids is perhaps the most well described nutrient that mediate the dietary factors affect blood lipids. The central dietary advice for patients with immune system (73), the main function is the production of less hyperlipidemia is, except for weight normalization, to improve their dietary fat proinflammatory eicosanoids in competition with omega-6 fatty acids (figure quality. Several saturated fatty acids are indicated to decrease the LDL- 6), thereby facilitating a resolution or alleviation of inflammation. A higher receptor activity whereas several unsaturated fatty acids have the opposite proportion of omega-3 fatty acids in blood has also been correlated to lower effect, thereby altering the concentration of LDL-C in circulation (78). In concentrations of proinflammatory, and higher concentrations of anti- patients with RA, one intervention study has seen lowering of cholesterol inflammatory cytokines in the general population (74). In addition to effects concentrations in blood after a vegan exclusion diet (5) (presumably on inflammation, omega-3 fatty acids decrease the circulating TG combining high fiber intake with improved dietary fat quality). This trial was concentrations by increased lipid oxidation and decreased lipid production in however troubled by simultaneous weight loss, which naturally affect the lipid the liver (75). transportation in blood (79). Other Mediterranean-like diet interventions have yet failed to lower blood lipid levels in patients with RA (1, 16). 16 17 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 2.1.2 DIET AND INFLAMMATION IN RA When assessing dietary effects on inflammation, it is rather clear that obesity instigates chronic inflammation, and that weight reduction among the obese has anti-inflammatory effects regardless of diet composition (80), by reducing proinflammatory cytokines, and improving insulin sensitivity. The whole-diet intervention trials performed on patients with RA have often been of small sample size, and quite often the intervention has both resulted in weight loss and a high dropout rate (Table 4). To date, there is only one controlled whole-diet intervention study that has seen improvement in one of the clinically validated biomarkers of inflammation (CRP or ESR), without unadjusted and uneven weight loss between study arms. The study by Sarzi- puttini et al (13) tried a hypoallergenic diet higher in unsaturated fatty acids compared to a control diet. The intervention was devoid of wheat, egg, milk, strawberries, tomato, chocolate, dried fruit, acid fruits as well as crustacean products, and low in red meat. Instead of these foods, the intervention diet was high in hydrolyzed milk, cornbread and corn flour, fresh pineapple and cooked apples. Both diets were designed to be weight-normalizing for the participants, and when analyzed adjusted for BMI-developments, ESR (but not CRP) was reduced by the intervention 2.1.3 DIET AND BODY COMPOSITION IN RA In terms of dietary intervention on patients with RA aimed at improving body composition, data are rather scarce. One previous trial examined the effects of an amino acid supplement enriched with β-hydroxy-β-methylbutyrate compared to placebo, but saw increased lean mass over time regardless of treatment (81). Another trial assessed creatine supplementation and found that patients with RA, just like patients without RA, increased in lean mass following creatine supplementation, but saw no functional benefit (82). As of today, to the best of our knowledge, there has been no diet intervention assessing effects on body composition in patients with RA. 2.1.4 NEED OF FURTHER STUDIES As to date, the potential impact of diet on health outcomes in patients with RA is rather unclear, and remains to be determined. The dietary impact on cardiovascular risk profile, inflammation and body composition under controlled, weight stable conditions, remains to be determined. There is a need to assess efficacy of dietary interventions in patients with RA. 19 18 Table 4. Previous food-based intervention studies on the biomarkers of inflammation CRP and ESR in patients with RA Author Intervention Design Length Weight N Dropout CRP ESR change Included (I / C)1 Barnard et al., 2022 Vegan exclusion diet with Randomized crossover trial 16w 44 27% (5) reintroduction vs I/C No info placebo supplement Dawczynski C et al., Omega-3 enriched dairy products randomized, double-blind, placebo- 12w No 45 13% No 2009 (6) controlled crossover study info. I/C No info info. Dennisov et al., 1992 Hypoallergic anti-inflammatory non-randomized parallel trial 4w No info 92 0% / 0% No (9) diet vs no diet intervention info. Elkan AC et al., 2008 Vegan vs balanced non-vegan diet randomized parallel trial 1 year 66 29% / 21% No (10) info. Hansen GV et al., high-protein, fish och vegetarian randomized parallel trial, single 6 mo NS. 109 26% No 1996 (11) foods, omega-3, selenium & blinded I/C No info info. vitamin A,C & E supplements Holst-Jensen SE et Peptide based / Elemental diet vs randomized parallel trial, blinded 4 w, 30 13% / 7% al., 1998 (12) habitual diet assessor (+ 6 mo)1 Kavanagh et al., 1995 Modified elemental diet & food randomized parallel trial 24 w ↓ at 4w 47 46% / 78% (14) reintroduction vs habitual diet NS at supplemented with elemental 24w. nutrition Kjelsen-Kragh et al., Fasting and gluten-free vegan or randomized parallel trial, blinded 13m 53 0% / 0% 1991 (15) vegetarian vs "normal" assessor Kremer et al., 1985 Mufa- & pufa rich diet with marine randomized double blind parallel 12 w NS 52 15% / 16% No (17) oil, vs control resembling average trial info. population intake Lindqvist et al., 2018 1 meal of blue mussel vs randomized crossover trial 11 w NS. 39 41% (19) chicken/ham/beef, 5 days/week I/C No info McKellar et al., 2007 Mediterranean diet cooking classes non-randomized parallel trial 6 w, NS. 130 0% / 0% (1) versus brochure on healthy eating (+ 3 mo)1 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander AIM The aim of this thesis was to explore the relation between dietary intake and biological markers of health in patients with RA. The specific aim of each paper in this thesis was as follows: I. To study if dietary intervention can alter cardiovascular risk profile in patients with RA, and if so, what type of characteristics that predispose for response II. To study effects of dietary intervention on markers of inflammation in patients with RA III. To study if body composition is altered by a dietary intervention in patients with RA and what patient characteristics that predispose for treatment response IV. To explore the role of nutritional quality of habitual diet in relation to health outcomes in patients with RA, and the development in nutritional quality during a diet intervention study 21 20 Table 4 continued. Previous food-based intervention studies on the biomarkers of inflammation CRP and ESR in patients with RA Author Intervention Design Length Weight N Dropout CRP ESR change Included (I / C)1 McKellar et al., 2007 Mediterranean diet cooking classes versus non-randomized parallel 6 w, NS. 130 0% / 0% (1) brochure on healthy eating trial (+ 3 mo)1 Nenonen et al., 1998 raw food vegan vs habitual diet randomized parallel trial 3 mo 43 14% / 5% (2) Panush RS et al., hypoallergenic diet vs active comparator randomized double blind 10 w No info. 33 21% No 1983 (7) parallel trial I/C No info. info Sarzi-Puttini et al., hypoallergenic diet vs active comparator randomized double blind 24 w NS. 50 12% / 2000 (13) parallel trial 16% Skoldstam et al., Mediterranean diet vs habitual diet randomized parallel trial 12w 56 10% / 7% 2003 (16) van de Laar MA et Allergen-free enteral nutrition versus enteral randomized double blind 12 w 94 16% al., 1992 (18) nutrition w lactoproteins & azo-dyes parallel trial /18% Green color denotes decreases, gray indicates no significant change. 1: Follow-up after completed study. Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3 METHODS 3.1 DATA SOURCES This thesis is based on data from two clinical intervention studies performed in patients with RA, one whole diet intervention study and one postprandial meal challenge study. 3.2 THE ANTI-INFLAMMATORY DIET IN Figure 7. Median (25p, 75p) duration of intervention diet period were 9.6 (9.3, 9.9) weeks, intervention diet periods 9.6 (9.0, 9.9) weeks, the washout time between RHEUMATOID ARTHRITIS (ADIRA) TRIAL dietary treatment was 16.9 (10, 17) weeks. The ADIRA-trial aimed at determining the efficacy of combining foods with Potential participants, with diagnosis of RA listed at Sahlgrenska University potential anti-inflammatory effects into a proposed anti-inflammatory Hospital, were identified in the Swedish Rheumatology Quality Register portfolio diet, compared to a typical western diet (nutritionally similar to the (SRQ) (n = 1091). Those residing on an address where the company mat.se average nutritional intake in Sweden). The ADIRA-trial was approved by the (procured to provide study foods) delivered food were invited to participate (n local ethical committee (976-16 and T519-17), registered in ClinicalTrials.gov = 774). In total, 113 (15%) patients responded, and out of those 66 fulfilled (NCT02941055), was carried out from February 2017 until May 2018. The pre-screening criteria and were thus called to screening. Inclusion criteria main outcome in the ADIRA-trial was a change in DAS28 as a consequence prescreening were age ≥18 years and ≥2 years disease duration. Exclusion of altered dietary intake. criteria at screening were any known life threatening diseases, pregnancy or lactation, allergies to any of the foods in the study, or inability to confirm an The outcomes included in the papers forming this thesis are secondary apprehension of study instructions. Out of these, 50 patients were invited and outcomes from the ADIRA-trial; biological markers of inflammation, risk randomized by a computer-generated list (allocation ratio 1:1) to take part in factors for CVD and body composition. Additionally, this thesis has taken the study. Data from 47 patients who completed at least one diet period, and advantage of data from screening in the ADIRA-trial, and evaluated markers 44 who completed two diet periods were available for analysis from the of inflammation, body composition and DAS28 in cross-sectional analyses. ADIRA-trial. (Figure 8). 3.2.1 STUDY DESIGN & PARTICIPANT SELECTION IN THE ADIRA-TRIAL The ADIRA-trial was designed as a randomized controlled crossover trial (Figure 7); sequence was computer-generated and assigned by random at screening. A sample size of 38 was needed for 90% power to detect a change in DAS28 of 0.6 units (α = 0.05). To account for dropout, 50 participants were recruited. 22 23 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Table 5. Participant data material used from the ADIRA-trial Total With BIS Without With high measurements1 major compliance3 medication changes2 Completed Intervention 45 41 38 43 one diet period Control 46 41 38 39 Completed both diet 44 40 36 35 periods 1 Participants with metal implants were excluded from measurements of body composition. 2 Without new or discontinued DMARD or glucocorticoid treatment during any of the diet periods. 3 Participants who reported > 80% compliance. Abbreviations: BIS, Bioelectrical impedance spectroscopy . Figure 8. Flowchart according to the CONSORT recommendation on patients in the ADIRA-trial. Medication, prescribed and self-administered, were recorded by participants at screening, and during the study visits after each diet period. Interval, dose, and Participants were interviewed midway into each dietary period, and answered time of each substance was compiled. The reported medication usage after questions on how many of the prescribed meals had been consumed during the each visit was compared to usage at screening to identify changes. All drugs past week. For each day a basic compliance score was calculated; consuming were then categorized by substance class. In statistical analysis, changes in all of a meal (breakfast, the main meal, and snacks) gave 100%, some parts of medication were defined as complete cessation or a newly instated drug. a meal 50%, and nothing of the meals 0%. Every meal was given the same weight, and an average score over the past week was calculated for each In order not to convolute the dataset too much, changes in timing, dosage or participant. A cut-off at > 80% was used to define participants with high other medications than conventional synthetic (cs)DMARDs, bDMARDs or compliance. glucocorticoids were ignored in the statistical analysis. If a participant was started on a new, or completely discontinued any of these drugs during any of In general, the self-reported compliance was high, with a median (25p, 75p) of the diet periods, all data from that individual was deemed as potentially 100% (93%, 100%) during intervention and 100% (87%, 100%) during control affected by altered medication. Changes in medication were considered in a among those completing at least one diet period. The proportion of participants sensitivity analysis when assessing cardiovascular risk factors (Paper I) and in with high compliance (i.e., >80%) was 96% during the intervention diet period the main analysis of biological markers of inflammation (Paper II). and 85% during control diet period (Table 5). Changes in csDMARD-treatment were noted for: azathioprine, ciklosporin, metotrexate and sulfasalazine. Changes in bDMARD-treatment were seen in: abatacept, certolizumab pegol, etanercept, golimumab, infliximab, rituximab. 24 25 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Changes in glucocorticoid usage was noted for: betamethasone, prednisolone Table 6. Average daily nutritional contents of the supplied foods and unspecified injections. Intervention Control 3.2.2 DIETARY INTERVENTION Energy (kcal) 1100 1120 The intervention in the ADIRA-trial was facilitated by home-delivery of foods Protein (g) 48 65 and recipes of easy to prepare meals, along with basic dietary guidance. Fat (g) 45 32 Carbohydrates (g) 114 136 During the intervention diet period, participants were asked to limit red meat Fiber (g) 20 10 to ≤ 3 times/week, consume ≥ 5 servings of fruit, berries or vegetables daily. Iron (mg) 9 8 Further, choosing oil and margarine over butter, selecting low-fat dairy Beta-Carotene (µg) 1768 148 products and whole grain products was encouraged. During the control diet Retinol equivalent 191 133 period, participants were asked to eat red meat ≥ 5 times/week, consume ≤ 5 Vitamin D (µg) 10 1 servings of fruits, berries or vegetables daily. Further, cooking in butter and Vitamin E (mg) 13 6 the use of whole-fat dairy products was encouraged and use of probiotics was Riboflavin(mg) 0.9 1.6 dissuaded. Niacin equivalent 23 29 Vitamin B12 (µg) 6 3 During both diet periods, participants received a partial food plan covering Magnesium (mg) 324 161 1100 kcal/day (Table 6), during 5 days per week, allowing for 2 days per week Potassium (mg) 2474 2137 of self-sufficing intake. During the intervention diet period, foods were fatty SFA (g) 9 12 fish, legumes, whole grain products, nuts, fruits, berries and vegetables, in MUFA (g) 15 12 contrast to the control diet period where the focus was on refined grains, red PUFA (g) 17 5 meat, chicken and protein-rich snacks (protein bars, quark-yoghurt mix or Cholesterol (mg) 69 150 protein pudding) (Table 7). Monosaccharides (g) 32 23 Disaccharides (g) 23 34 To avoid potential impact from varying climate conditions and seasonal habits, the intervention and control diet periods were designed to be evenly balanced EPA (Fatty acid 20:5)(g) 0.7 0.0 over the year (Figure 9). The intervention periods occurred only slightly more DPA (Fatty acid 22:5)(g) 0.4 0.0 frequent during spring, and consequently the control periods were slightly DHA (Fatty acid 22:6)(g) 1.3 0.0 more frequent during autumn. Selenium (µg) 45 27 Vitamin K (µg) 85 14 Iodine (µg) 55 108 Starch (g) 17 55 Copper (mg) 0.7 0.4 Antioxidants (mmol) 11 4 Macro- and micronutrients where difference was ≥ 10% between the intervention and the control diet plan. Individual fatty acids are excluded, as well as sodium which was not reliably accounted for. Values are calculated in and exported from DietistNet. Abbreviations: DHA, Docosahexaenoic acid; DPA, Docosapentaenoic acid; EPA, Eicosapentaenoic acid; MUFA, Monounsaturated fatty acids; PUFA, Polyunsaturated fatty acids; SFA, Saturated fatty acids. 26 27 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Table 7. Overview of the home-delivered foods in the ADIRA-trial Intervention Control Breakfast Probiotic shot, frozen berries or  Orange juice and pomegranate and a variation of; white bread with Low fat sour milk and nuts butter and cheese, or   Fiber enriched oat  Protein rich yoghurt porridge, skimmed milk, with cornflakes and walnuts orange juice  Low fat yoghurt, whole grain muesli Hot meal  Fish 2-5 (average 3.8)  Red meat 3-4 (average 3.5) times /week times/week  Legume based vegetarian  Chicken 1-2 (average 1.5) food 0-3 (average 1.2) Figure 9. The number of active patients per week stratified by treatment. Active times/week times/week weeks are defined as the week a participant had the first home delivery of foods until Whole grain products and the week that post-period measurements occurred.  vegetables in every meal Snack  2 fruits per day (apple and  A portion of quark-yoghurt banana) mix per day, or  A protein bar, or  a portion of protein pudding 29 28 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.2.3 OUTCOMES FROM THE ADIRA-TRIAL INCLUDED IN THIS THESIS 3.2.3.1 PAPER I Clinical markers of blood lipid transport (total- HDL- and LDL-cholesterol as well as TG-concentration) were analyzed in fresh samples in the fasting state before and after each diet period. The samples were analyzed according to clinical routine at the Sahlgrenska University Hospital, by enzymatic colorimetry using a Cobas 8000 instrument from Roche Diagnostica, Scandinavia AB. To explore lipid transport particles numbers, APO-B100 and APO-A1 concentration, and cholesterol and TG concentrations in particles, frozen samples were analyzed in a serie by 1H Nuclear Magnetic Resonance (NMR) analysis (83) at the Swedish NMR Centre. Fatty acid composition in plasma was analyzed externally in a serie from frozen samples using gas chromatography (84) at the University of Southampton, United Kingdom. Blood pressure was measured by nurses at the Department of Clinical Rheumatology Research Center before and after each diet period. 3.2.3.2 PAPER II Clinical biomarkers of inflammation were assessed by routine analysis of CRP and ESR at the Sahlgrenska university Hospital in fresh samples before and after each dietary period. An exploratory analysis of inflammation-related proteins in serum samples was analyzed externally by Olink Proteomics AB, Uppsala, Sweden using the Olink® Target 96 Inflammation panel (85). The analysis of inflammation-related proteins was done on a subset of serum samples where standardized handled serum samples were available (n=26) (Figure 10). 30 31 Figure 10. Serum samples used for quantification of inflammation-related proteins Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.2.3.3 PAPER III 3.3 THE POSTPRANDIAL INFLAMMATION IN Body composition was measured by bioelectrical impedance spectroscopy before and after each diet period in the fasted state after 5 minutes in a supine RHEUMATOID ARTHRITIS (PIRA) position using ImpediMed SFB7 (ImpediMed, Brisbane, Australia) and The PIRA-trial is a randomized controlled cross-over trial studying the disposable electrodes from the Fresenius Kabi Body Composition Monitor postprandial effects of intake of meals of different composition (Figure 11). product line. Weight was measured in habitual clothing and 1 kg was The PIRA study was approved by the regional ethical review board in subtracted to account for clothing. Gothenburg (Dnr 2019-05242) and registered on ClinicalTrials.gov (NCT04247009). Recruitment for PIRA began in January 2020, and clinical Metabolites examined for relation to body composition, branched chain amino study visits began in February 2020, but had to be halted in March 2020 due acids (BCAA) and insulin-like growth factor-1 (IGF-1), were analyzed in a to the COVID-19 pandemic. The trial was later resumed in August 2021 and serie from frozen serum samples. BCAA concentration was quantified by completed in November 2021. NMR analysis, Albumin and IGF-1 were quantified in serum following routine procedures by the clinical laboratory at Sahlgrenska University Hospital. 3.2.3.4 PAPER IV In addition to outcomes evaluated in paper I, II and III, a nutritional index score based on NRF11.3 (86, 87), as described further below, was calculated and used as an index of nutritional quality in paper IV. Data was also acquired from screening, where patients filled questionnaires on demographical background, employment status and of habitual physical activity. Waist to hip ratio and weight were measured to the closes 0.5 cm. Figure 11. The randomized controlled cross over trial PIRA Participants also disclosed their medication lists. The primary aim was to evaluate if isocaloric meals composed of different 3.2.4 CONTRIBUTION TO THE ADIRA-TRIAL BY THE PHD- protein and fat sources would affect inflammation and metabolic response CANDIDATE differently in patients with RA, and if patients with RA respond differently The PhD-candidate has collected and analyzed data from the ADIRA-trial, but compared to healthy controls. The main protein source was either minced meat was not originally involved in conceiving of or developing the study. (mix of pork and beef), salmon or a soy protein product (Figure 11). Patients with RA were planned to consume all three meals, and healthy controls only the meal of minced meat. 3.3.1 MEAL COMPOSITION The meals were developed and carefully calculated and standardized to differ at maximum 1 gram in protein, fat and carbohydrate between meals. In order to best accommodate standardization, burger-based meals were developed. Each meal consisted of two burgers, white bread, salad, cucumber and tomato along with hamburger dressing, calculated energy contents were totaling at 700 kcal/meal. Differences in water and fiber content were ignored. All meals were fried with standardized amount of canola oil and heated to the prespecified internal temperatures of 75 ̊ for minced beef, 52 ̊ for salmon and 70 ̊ for the soy-protein based dish. 32 33 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Analysis of the chemical composition of the meals was done externally  Obesity (> 30 kg/m2) (Eurofins Food & Feed Testing, Sweden) to validate the nutrition content.  Diagnosis of cancer, inflammatory bowel disease, celiac disease, Differences in amount of fat in the calculated meals versus the analysis were diabetes compensated by enriching the hamburger dressing with varying amounts of canola oil. Differences in carbohydrates and protein content was left  Allergy or intolerance to any of the foods in the study uncorrected.  Pregnancy or breastfeeding  Use of any blood lipid lowering medication, glucocorticoids or IL-6- 3.3.2 RECRUITING PARTICIPANTS WITH RA inhibitor during the past 4 weeks Female patients with confirmed diagnosis of RA (ICD-code M05.9 or M05.8)  No DMARD-changes for the past 3 months listed at the Sahlgrenska University Hospital was identified in SRQ. Inclusion  Smoking criteria for this database export was ≥ 2 years disease duration, an age span of 20-70 years and no ongoing treatment with IL-6 inhibition during the past four  Hemoglobin < 100 g/L weeks. Presumptive participants (n=934) were sent letters of invitation, of  HbA1c levels above age-standardized reference which 83 responded. Responders were double checked for diagnosis and those breastfeeding or pregnant, or who self-reported obesity (>30 kg/m2), smokers After excluding participants on the above criteria, eligible patients (n=37) were or with intolerance to any of the trial foods were excluded. Eventually, 42 invited to participate in the postprandial meal challenges. patients were screened. 3.3.4 RECRUITING HEALTHY CONTROLS 3.3.3 SCREENING OF PARTICIPANTS WITH RA Healthy controls, defined as not having a rheumatic diagnosis, were advertised During screening, weight, height, waist to hip ratio, body composition, ESR, for in social media and by word of mouth. Inclusion criteria were female sex CRP, hemoglobin concentration and HbA1c were measured in the non-fasting and age in the range of 20-70 years. Out of those responding to advertisement, state, and food records were administered. controls were matched to the patients with RA based on age, BMI and physical activity level. The same exclusion criteria as previously described for patients Weight and body composition were measured on a void bladder in with RA were applied also to controls. No medication or diagnoses were standardized hospital gown without shoes by Multi-frequency bio-electrical checked for, and joint status was not examined, but otherwise controls went impedance analysis (Tanita MC-180 MA, Tanita, Tokyo, Japan), dual energy through the same procedure as patients with RA. X-ray (DXA) technique using Luna Prodigy (Lunar Prodigy, enCORE software 12.30.008, GE Health Care, Madison, WI, USA) and bioelectrical 3.3.5 POSTPRANDIAL MEAL CHALLENGE FOR PATIENTS impedance spectroscopy (BIS) using ImpediMed SFB7 (ImpediMed, WITH RA Brisbane, Australia). Height was measured to the closest 1 cm and waist-hip ratio to the closest 0.5 cm. Participants came to the study center in the fasted state and were randomly assigned a meal sequence. Spot urine was collected, and blood drawn before CRP, ESR, hemoglobin and glycated hemoglobin (HbA1c) were measured in the meal. Twenty minutes were allotted for consumption of the meal and fresh samples according to the clinical routine at the Sahlgrenska University additional blood was then drawn after 30, 60, 120, 180 and 300 minutes Hospital. DAS28-ESR was estimated by nurses at the department of clinical (Figure 12). RBCs were collected only at the first meal challenge, serum, rheumatology research center at the Sahlgrenska University Hospital. plasma, blood, peripheral blood mononuclear cells (PBMC) and urine samples Participants were also asked to complete 4-day food records. The participants’ were extracted during all visits. health status and medication were further checked during interviews and from patient journals, those fulfilling any of the following exclusion criteria were excluded; 34 35 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.3.6 POSTPRANDIAL MEAL CHALLENGES FOR HEALTHY CONTROLS Healthy controls were only served the meal of minced meat, and in contrast to patients with RA, no PBMC-samples were collected and no whole blood was analyzed, but otherwise the same protocol was used (Figure 12). 3.3.7 OUTCOME MEASURES IN THE PIRA-TRIAL The main outcome of the PIRA-trial was the postprandial development of IL- 6 in serum. Secondary to IL-6, inflammatory and metabolic markers such as CRP, TG, apolipoproteins and gene expression analysis of inflammation- related genes. Serum and urine will also be used for analyses of metabolomics. 3.3.8 OUTCOMES FROM THE PIRA-TRIAL INCLUDED IN THIS THESIS Due to the COVID-19 pandemic and subsequent trial delays, only the data from screening that were performed in a similar way in the ADIRA-trial, analyzed in a cross-sectional dataset, are included in this thesis; body composition as measured by BIS, CRP, DAS28-ESR, ESR and dietary intake as recorded in 4-day food records. 3.3.9 CONTRIBUTION TO THE PIRA-TRIAL BY THE PHD- CANDIDATE The PhD-candidate worked on forming the hypothesis, meal development, selection of biomarkers to analyze, the power calculation, ethical approval application, study protocols as well as the recruitment and execution of the PIRA-trial. The trial unfortunately had to be halted due to the pandemic, in the second run, the candidate worked with planning the organization of preparations and meal challenges, re-recruitment of patients with RA, and processing of biological samples. 37 36 Figure 12. Times of postprandial sample collection during the meal challenges Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.4 COMMON METHODOLOGY IN THE ADIRA- Table 8. Composite nutritional score according to the NRF11.3 AND PIRA-TRIALS Score cut-off Maximum value 3.4.1 ASSESSMENT OF NUTRITIONAL QUALITY Positive scoring nutrients Women Men Assessment of nutritional quality was performed based on the dietary intake reported in the food records (3-day in ADIRA and 4-day food records in Calcium (mg) 800 800 1 PIRA). A composite score of the intake of some key nutrients, Nutrient Rich Foods Index (NRF)-11.3, previously identified to be of importance to a Fibre (g) 25 35 No limit Swedish population (87), was used. In brief, for each participant, the percentage score of recommended daily intake (RDI) of key nutrients were Folate (µg) 3001/4002 300 1 summed, and the percentage score of negative nutrients were subtracted, leaving a final composite score (Table 8). In paper III, this was computed on Iron (mg) 91/152 9 1 density values, i.e. adjusting all nutritional intake to an assumed 2000 kcal/day for females and 2500 kcal/day for males. In paper IV, it was calculated both as Magnesium (mg) 280 350 1 density and as absolute score, and the absolute score was used for the main analysis. Potassium (g) 3.1 3.5 1 3.4.2 ASSESSMENT OF DISEASE ACTIVITY Protein (E%)3 15 15 1.33 In both studies, joint examinations were performed by the same research nurses at Clinical Rheumatological Research Centre at Sahlgrenska University Vitamin A (retinol 700 900 1 Hospital. equivalents) 3.4.3 ASSESSMENT OF BODY COMPOSITION Vitamin C (mg) 75 75 1 Assessment of body composition was performed with several methods in the PIRA-trial, and only by BIS in the ADIRA trial. The BIS-measurements were Vitamin D (µg) 10 10 1 also not done in exactly the same way in the two trials. However, the same machine, with the same type of electrodes and a similar protocol of a 5-15- Vitamin E (mg) 8 10 1 minute supine position prior to measurement were applied in the ADIRA and PIRA trials. Negative scoring nutrients Score cut-off 3.4.4 ASSESSMENT OF PHYSIAL ACTIVITY Added sugars (E%) 10 10 No limit Physical activity was assessed as described in paper III in both trials. Based on Saturated fat (E%) 10 10 No limit scales between 1 and 5 on both habitual physical activity and intentional physical exercise, a physical activity index between 1 and 4 was calculated, Sodium (g) 2.4 2.4 No limit resembling what was previously tried and validated by Wareham et al. (88). 1 Dietary recommendations for post-menopausal women. 2 Dietary recommendations for women of reproductive age 3 Based on the mean of recommended 10-20 E% of energy intake, values above 20 E% were capped. Abbreviations: NRF11.3, Nutrient rich foods index 11.3 38 39 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.5 DIFFERING METHODS IN THE ADIRA- AND 3.6 STATISTICAL ANALYSES PIRA-TRIALS When analyzing the outcomes in the ADIRA-trial, the main analysis was a Most procedures were similar between the ADIRA- and the PIRA-trial during linear ANCOVA mixed model regression analysis (Table 9). Each outcome screening, but the recruitment differed in some key aspects. Most notable, in was adjusted for the run-in value prior to each dietary period, other fixed the PIRA-trial only females were recruited. Further, in PIRA acceptable age variables were period (first or second) and treatment .(intervention or control). span was from 20 to 70 years, where as in ADIRA it was set from 18 to 75 Each participant was included as random intercept. For each outcome residuals years. There was also cut-off for BMI when recruiting patients to the PIRA- trial, so that no obese patients were included in the trial. No such BMI related were inspected, and when needed, the outcome was transformed to comply criterion was present in the ADIRA-trial. with model assumptions. Furthermore, presumptive participants prescribed some specific drugs (such as Table 9. Statistical methods used in the included papers IL-6 inhibitors) were excluded in the PIRA-trial. Some more subtle differences include the time span of food records; in PIRA 4-day food records were used, Main outcomes and in ADIRA 3-day food records were used. In ADIRA, body composition was measured by BIS-measurement by multiple operators in the fasted state, Paper I Paper II Paper III Paper IV whereas in PIRA, DXA and BIS measurements were combined and measured by a single dedicated operator, in standardized clothing on a void bladder in linear X X X X the non-fasted state. In the ADIRA-trial, all patients with RA who resided in ANCOVA areas where home-delivery by the company mat.se was possible, were invited. mixed model In the PIRA-study, all patients with a diagnosis of RA, who received medical care at the Sahlgrenska University Hospital, were invited to participate. Confounder X X X X analysis Interaction X X X X analysis Linear X regression analysis Baseline comparisons Fisher’s exact X X X test Kruskal Wallis X test Mann Whitney X X test 40 41 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.6.1 CARRY-OVER EFFECTS 3.6.3 INTERACTION ANALYSES Washout was deemed sufficient not to influence outcome in paper I, carry-over Interaction analyses were done on dichotomized data (above or below median) effects were however explored in paper II and III. in papers I and III, and in paper II it was tested for carry-over effects (Table 10). In paper IV, habitual nutritional quality index was tested modelled as a In paper II, carry-over effects were evaluated by testing the interaction between continuous variable against developments in health outcomes (Table 10). diet period (first or second) and treatment (intervention or control) in the ANCOVA linear mixed model for CRP and ESR, with a preset cutoff for P- values at < 0.20. This analysis did not indicate any carry-over effect. In paper III, carry-over effect was tested by using students’ t-test on delta fat free mass (FFM), comparing developments between first and second period , within control and within intervention diet periods, with a cut-off for P-values at < 0.05. This analysis did not indicate any significant carry-over effect. 3.6.2 CONFOUNDER ANALYSES Confounder analysis in the ANCOVA mixed model was done in papers I, II and III. Any covariate exerting a change in effect estimate (Beta) >10% was included as a fixed effect in the ANCOVA linear regression mixed model. In paper I, II and III, the following variables were tested for confounding:  Age  Nicotine use (yes/no)  Sex  Dietary quality (index  Body mass index between 0 to 12) (BMI)  Educational level (index between 1 to 5). Similarly, in paper IV, any of the following variables that altered the effect estimate (Beta) >10% in a crude model were included as covariate in the final cross sectional linear regression:  Study site (ADIRA / PIRA)  Age (in years)  Biological sex (male / female)  Physical activity (a simplified index ranging 1-3) 42 43 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 3.6.4 PER PROTOCOL AND INTENTION TO TREAT ANALYSIS A per protocol analysis is based on the participants who complete a given intervention with acceptable compliance. In an intention to treat analysis, all available data included in the analysis, typically this should include all patients randomized to an intervention. There is however some uncertainty on how to properly handle missing data in an intention to treat analysis. In this thesis, per protocol analysis has been used and results are presented both on 1) all participants where data were available, and 2) participants who completed the intervention with high compliance and without major changes in medication. 44 45 Table 10. Interaction analyses performed in the papers in this thesis Paper I Paper II Paper III Paper IV Outcome variables  Total cholesterol  CRP  FFM  APO-B100/APO- A1 quote  LDL-C  ESR  FM  CRP  HDL-C  Fat mass %  DAS28  TG  ESR  Fat mass %  Non-HDL-C Variables tested for 1. BMI (normal weight vs 1. Diet period 1. Habitual dietary quality 1. Habitual nutritional effect modification above normal weight) 2. Physical activity (questionnaire of quality (NRF11.3 score 2. Coronary risk (1 or 2) intentional exercise and daily activity) at screening) evaluation (SCORE 3. High FMI2 2015)1 4. Low FFMI2 3. Dietary quality index3 5. Low FFMI & high FMI2 4. Dietary fiber intake3 6. HAQ 5. Dietary fat intake3 7. Age 8. Educational level (below or above 2- years senior high school) 9. Albumin concentration 10. Employment status (employed / not employed) 1Swedish version of the systematic coronary risk score 2015, which indicates the risk of death of CVD within the coming 10 years (3). 2 Below 25th percentile for FFMI or above 75th percentile for FMI, compared to reference values from the general population (4). 3 Based on food frequency questionnaires at screening, ranked according to an index proposed by Swedish Food Agency to measure dietary quality (8) Abbreviations: LDL-C, Low density lipoprotein-bound cholesterol; HDL-C, High density lipoprotein-bound cholesterol; TG, Triglycerides; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; BMI, Body mass index; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; FFM, Fat free mass; FM, Fat mass; FMI, Fat mass index; FFMI, Fat free mass index; HAQ, Health assessment questionnaire disability index; APO-B100, Apoliprotein-B100; APO-A1, Apolipoprotein-A1; CRP, C-reactive protein; DAS28, 28-joints disease activity score erythrocyte sedimentation rate; ESR, Erythrocyte sedimentation rate; Non-HDL-C, Non high density lipoprotein-bound cholesterol; NRF, Nutrient rich foods index. Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 4 RESULTS 4.2 PAPER I Overall, the results of this thesis indicated that blood lipid profile, In paper I, examining blood lipid profile as a proxy for cardiovascular risk in apolipoprotein concentrations and inflammation were affected by dietary the ADIRA-trial, the intervention diet increased HDL-cholesterol and reduced intevention in patients with RA. Body composition improved over time non-HDL-cholesterol as well as TG concentration compared to control diet. regardless of diet allocation. Habitual nutritional quality was not tied to any of The increase in HDL-bound cholesterol occurred within the lower density the assessed outcomes a cross sectional analysis. HDL-particles (density range of 1.063 - 1.112 kg/L), and changes in TG primarily occurred in VLDL- and IDL-particles. Notably, Apolipoprotein The main outcome of the ADIRA-trial, the effects on DAS28, has been B100 decreased, and the APO-B100/APO-A1 ratio improved (Figure 13). published outside of this thesis (89). When using the same intention to treat Lastly, the fatty acid composition in plasma was altered between the and per protocol analyses as in the papers included in this thesis, the results on intervention and control period; saturated fatty acids were comparatively DAS28 are as follows. In the whole group (n = 47), there was no effect lower, whereas mono- and polyunsaturated fatty acids increased. between the intervention and the control diet period on DAS28 (mean: –0.289; 95% CI: –0.652, 0.075; P = 0.116). When excluding participants with new or There were also characteristics that modified effects between intervention discontinued DMARD of glucocorticoid treatment (previously unpublished compared to control; HDL-C increased only in those with low risk of data), DAS28 was lower after intervention compared to after the control diet cardiovascular death (according to SCORE 2015). TG were reduced only in period (mean: -0.487; 95% CI: -0.841, -0.133; P = 0.009) (n=38). Similarly, those either overweight, with a low habitual fiber intake, or a low habitual diet when analyzing only those without these major changes in medication and with quality at study start. Similarly, LDL-C was comparatively lower in those with high compliance to both diet periods (n = 29) (previously unpublished data), a high habitual fiber intake. DAS28 was lower after intervention compared to after the control diet period (mean: -0.505; 95% CI: -0.854, -0.156; P = 0.006). 4.3 PAPER II In paper II, the effects on markers of inflammation in the ADIRA-trial was evaluated in participants without changed medication. There was no effect on 4.1 STUDY POPULATION CRP or ESR in the group as a whole. When removing those with low compliance and who did not complete both diet periods, ESR was lower after In the ADIRA-trial, out of the 50 patients included, 47 completed one diet intervention compared to after control diet period (Figure 13) (mean –5.490 period and 44 completed both diet periods. Out of those completing at least mm/1h, 95% CI –10.310, –0.669 P = 0.027) (n = 29). one diet period, 77% were female, median age was 63 years. About three quarters of participants used a csDMARD, a third bDMARD and around one In a multiplex assay on a subset of the participants (n = 26), several chemokines in eight no DMARD at all. A tenth of participants changed their DMARD (C-X-C motif chemokine ligands (CXCL)-1, -5, and -6) and the protein treatment, and another tenth changed their glucocorticoid treatment during the TNFS14 was lower after intervention compared to after control diet period. trial. Most had a moderate (57%) or low (28%) disease activity, only a minority When removing those with low compliance, only the chemokines CXCL-1 and were in remission (6%) or had a high disease activity (9%). -6 and the protein GDNF were lower after intervention compared to after control. In the PIRA-trial, 30 patients supplied sufficient information (food records and body composition measurements) without acute illness, and were thus included in the cross sectional analysis. Participants in PIRA were all females with a median age of 60 years. A large part of these participants was either in remission (43%) or had low disease activity (30%), fewer had a moderate (20%) or high (7%) disease activity. 46 47 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 4.4 PAPER III In paper III, evaluating changes in body composition in the ADIRA-trial, there was no difference between intervention and control in any measurement of body composition (Figure 14), but an improvement in both groups over time regardless of treatment was detected. An interaction analysis revealed that employment status modified the response; those not working had significantly lower fat mass after the intervention diet period compared to after the control diet period. Those not working were different to the others in the trial in several ways at study start; they had a higher fat mass % and FMI at study start, a higher age, they were all female, they reported a higher nutritional quality and fewer were medicated with csDMARDs. 4.5 PAPER IV Paper IV revealed that the nutritional quality differed significantly between the intervention diet and the control diet periods within participants in the ADIRA- trial. Most participants improved in nutritional quality on the intervention diet, whereas the response to the control diet was more mixed (Figure 15). The pooled cross-sectional analyses, from data collected at screening in the ADIRA- and the PIRA-trial, indicated no relation between nutritional quality index and any of the health outcomes examined. 49 48 Figure 13. Changes in blood lipids, apoliproteins and markers of inflammation in the ADIRA-trial between intervention and control diet periods Values displayed are changes between after intervention diet period compared to after control diet period. Analyzed by a linear mixed ANCOVA model, with the same model and confounders as in each respective paper. ●Intention to treat; all available data from participants completing 1 diet period. ●Per protocol; completed ≥1 diet period without major changes in medication during the trial. ●Per protocol; completed both diet periods without major changes in medication and with high compliance. Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander Figure 15. Developments in nutritional quality during the control and the intervention diet periods 51 50 Figure 14. Changes in measures of body composition in the ADIRA-trial between intervention and control diet periods Values displayed are changes between after intervention diet period compared to after control diet period. Analyzed by a linear mixed ANCOVA model, with the same model and confounders as in each respective paper. ●Intention to treat; all available data from participants completing 1 diet period. ●Per protocol; completed ≥ 1 diet period without major changes in medication during the trial. ●Per protocol; completed both diet periods without major changes in medication and with high compliance. Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 5 DISCUSSION prolonged intervention similar to the ADIRA-trial could reduce cardiovascular events. The effect on DAS28 is published outside of this thesis, but will be clarified briefly. In the published paper on DAS28 (89), there was a significant effect in A recent large cohort study investigated the importance of LDL-C, non-HDL- a Wilcoxon signed rank analysis comparing DAS28 after intervention to after C, TG concentrations in relation to APO-B on risk of myocardial infarction control (n = 44). However, there was no effect in an ANCOVA linear mixed (53). In this study, neither non-HDL-C nor TG concentrations were associated model on the group as a whole (n = 47), nor was there any effect when with cardiovascular risk when adjusted for Apo-B concentrations. This was including only data from participants with unchanged DMARD and true both for primary prevention (n=389529) and for secondary prevention in glucocorticoid treatment (n = 25). In comparison, in the papers that this thesis patients with established atherosclerosis (n=40430). is based upon, participants were excluded from analysis only if completely discontinuing or starting on new DMARD or glucocorticoid treatment, A previous systematic review found a decreased risk ratio of 7% (RR: 0.93, ignoring changes in dosage or timing. Consequently, this renders a larger 95% CI: 0.94, 0.99) for cardiovascular death by each decrease in 10 mg/dL of sample size. When using the same per protocol analyses as in the papers included in this thesis, DAS28 was lower after intervention compared to after APO-B (94), although the effect was limited to interventions that increased the control diet period, highlighting both the potent effects of anti-rheumatic expression of the LDL-receptor. The between-diet period effect seen in the medications, and the impact of participant selection criteria. ADIRA-trial on APO-B100 was about 4 mg/dL (mean: -4.455, 95% CI: -3.505, 6.846) (90), suggesting that the gap between the two diets conferred a modest change in cardiovascular risk. 5.1 CARDIOVASCULAR RISK, PAPER I In the ADIRA-trial, we found that TG- and non-HDL-C concentrations The mechanism behind the effect on blood lipid levels seen in the ADIRA-trial decreased, HDL-C increased and the Apo-B100/A1 quote improved after is not obvious, and the data do not allow for mechanistic conclusions. The intervention compared to after control diet (90). Our data do not allow us to intervention diet was relatively higher than the control diet in omega-3 fatty draw any conclusion on functional properties of lipoprotein particles, but we acids, which have previously (especially docosahexaenoic acid) been linked to can interpret the changes in lipid levels and apolipoproteins. decreased lipogenesis and increase beta-oxidation of fatty acids (75). Furthermore, the control diet was, relative to the intervention diet, high in These changes in blood lipids seen in the ADIRA-trial could be compared to saturated fatty acids, which can downregulate LDL-receptor expression and those in the largest RCT to date (n=7447) aimed at examining the effects of a activity (95). The intervention diet was also high in fiber; a current theory is Mediterranean diet supplemented with either extra virgin olive oil or nuts, the that higher fiber intake decreases blood cholesterol by binding to bile acids in Prevención con Dieta Mediterránea (PREDIMED) trial (91). This study has the intestine as well as increases formation of short chain fatty acids in the been criticized on the basis that it could be seen as the control participants also colon, which might reduce circulatory cholesterol by lowered hepatic synthesis consuming a type of Mediterranean diet (92). Still, in an intention to treat (96). The probiotic supplement used during the intervention diet might also analysis in the Mediterranean diet group supplemented with nuts (the have reduced the reuptake of cholesterol (97). intervention most similar to ADIRA), a reduced hazard ratio of cardiovascular events (myocardial infarction, stroke or CVD death) of 0.72 (95% CI; 0.54 – While significant effects were seen on markers indicating changes in blood 0.95) was found (91). An interim report about 12 weeks into PREDIMED lipid transportation, the behavior of HDL- and LDL-particles stretch beyond revealed increased HDL-C (mean: 0.04 mmol/L, 95% CI: 0.01, 0.07) (93), lower than the effect estimate seen in the ADIRA-trial of (Mean: 0.074 transporting lipids. As recently reviewed by Feingold et al. (98), smaller LDL- mmol/L, 95% CI: 0.000, 0.148) (93). However, this report also saw decreased particles have a lower affinity to the LDL-receptor, thereby allowing for longer glucose and insulin concentrations as well as a reduction in blood pressure. time in circulation. Additionally, the smaller LDL-particles more easily Thus, the effects in the PREDIMED-trial are not entirely comparable to the infiltrate the arterial walls and are more susceptible to oxidation. The HDL results of the ADIRA-trial, but lends the possibility to hypothesize that a particle also takes part in a range of mechanisms in addition to mere lipid transport, such as antioxidant, inflammatory and antithrombotic processes (99). 52 53 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander As an example, previous large scale RCT studies on cholesterol ester transfer been found in patients with RA compared to healthy controls (106). The protein (CETP) inhibitors, which raise HDL-C by blocking transportation of intervention diet in ADIRA was markedly higher in phytochemicals. Although cholesterol to Apo-B containing lipoproteins (essentially inhibiting some of we lack data on oxidative status of participants, it is possible that the difference the HDL particles’ functions) have generated either unfavorable (100-103), or in phytochemicals between the two diets conferred a benefit during the ambiguous results (104) on the risk of developing CVD. Assessing lipoprotein intervention diet through a reduction in oxidative stress. particle functionality or size was not available when handling data from the ADIRA-trial, thus, it might be worthwhile to bear in mind that we might not The difference in fiber content might also have affected inflammation. This have captured the complete impact on CVD risk. has been indicated in an uncontrolled trial on patients with RA that saw increased short-chain fatty acids, a product from the intestinal microflora, in Another major factor for CVD is heightened blood pressure; no effect on blood serum, coupled with decreased proinflammatory cytokines, during concurrent pressure was found in the ADIRA-trial. However, neither the intervention nor fiber-bar supplementation (107). the control diet was designed to alter sodium intake, which is the main dietary factor related to blood pressure. Results from the Olink multiplex assay indicated that effects were found in CXCL-1, and -6, and depending on participant selection CXCL-5, TNFSF14 5.2 INFLAMMATION, PAPER II or GDNF, without correction for multiple hypothesis testing. The number of hypothesis tests in the multiplex assay makes it highly unlikely that not at least When examining biomarkers of inflammation, we saw a significant reduction some of the results are due to type-1 errors. However, CXCL-1, -5 and -6 are in ESR among those who reported to have completed both diet periods with proteins that are related in structure and function. It could be reasonable to high compliance, but only a nonsignificant trend was seen in the whole group, postulate that the effects seen in these might reflect physiological changes. and no effect on CRP-concentration (105). CXCL-1, -5 and -6 belong to the cytokine class called chemokines, that direct movement of leukocytes to a site of infection or tissue damage. Previous data Most previous dietary intervention studies in patients with RA that indicate from diet interventions on these chemokines are rare. Animal studies have beneficial effects on ESR or CRP have done so during concurrent weight loss demonstrated increased concentrations of CXCL-1 and -5 in serum of (10, 12, 15, 16), or have not reported data on weight change (6, 9). To the best overweight and in hyperglycemic conditions (108), indicating that these of our knowledge, only one previous trial, investigating the response to a proteins are involved in both inflammation and metabolic regulation. hypoallergenic diet, found lowered inflammation when controlling for weight changes (13). An important strength of the ADIRA-trial is thus that no significant weight change occurred during or between the dietary intervention 5.3 BODY COMPOSITION, PAPER III periods. When examining changes in body composition, we demonstrated that body The available data do not allow for a mechanistic explanation of these results, composition among patients with RA can indeed be affected by participating nor does it allow us to pinpoint effects to specific foods since the intervention in a dietary intervention study, without affecting body weight (105). The was given as a portfolio diet. However, several key differences in design causality of improvements over time regardless of treatment is not clear, but in between the intervention diet and control diet could have influenced the results line with previous results from a study on dietary supplements in patients with such as dietary fatty acid, fiber-, phytochemical- and wholegrain intake. RA (81). There is no possibility to separate dietary effects from potential changes in other lifestyle related behavior (e.g., physical activity or sleep), The intervention diet was specifically higher in omega-3 fatty acid from marine even when were not actively encouraged during the study. foods compared to the control diet, which could have led to the formation of less proinflammatory eicosanoids. Chronic inflammation is also linked with We further identified an interaction in that we saw that a subgroup of not oxidative stress, and prolonged oxidative stress can lead to lipid oxidation, employed participants actually responded differently to the intervention and lipoprotein dysfunction and cell damage. Higher levels or reactive oxygen control diet periods, albeit the reasons for this remain elusive. At baseline, the species, deoxyribonucleic acid (DNA) damage and lipid peroxidation have non-employed group was significantly different in regard to body fatness, age, sex, medication and dietary intake. Likely, employment status could be a proxy 54 55 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander for several yet unknown factors. One could speculate that the higher body period, a decrease of about 1000 kcal/day compared to pre intervention, fatness in this subgroup made them more open to lose fat, and that the rendering the daily intake of virtually all nutrients below recommended levels. intervention diet, through less refined and fiber rich foods, could have affected Measuring dietary intake in free living individuals is rarely an exact science, satiety and nutrient metabolism, thereby reducing body fatness. as another example, there was no correlation between energy intake and weight change during the diet periods (data not shown). Under-reporting coupled with 5.4 NUTRITIONAL QUALITY INDEX, PAPER IV variations in daily intake are likely important factors behind this phenomenon. When examining health outcomes in a cross sectional analysis on pooled data An important limitation that applies specifically to paper IV is the handling of from the ADIRA- and PIRA-trials, no relation to the nutritional quality index the NRF11.3 index. This index was originally developed based on nutrient was revealed in any outcome. It could be that this patient group were already density per standardized serving size or per 100 kcal (86). The validation and sufficiently nourished, or it could be that the index was not capable of capturing modification of this index in a large Swedish cohort was based on Nordic the dietary intake well enough. One needs to keep in mind that the sample size recommendations on daily intake of nutrients, compared to energy-normalized was rather small, and that self-reported 3-4 days food records might not daily nutrient intake as estimated in food frequency questionnaires (87). In accurately enough capture habitual dietary intake. The accuracy of FFM paper IV, the analysis of developments during the ADIRA-trial is based on the assessed by BIS is also not very high (109), and it is plausible to assume that same index of nutrients identified as important in relation to longevity by Strid not all variation in FFM is captured by this method in a cross-sectional setting. et al. (87), but instead of energy density, absolute intake was assessed. Arguably, this approach answers a slightly different question than an energy When assessing the developments within the ADIRA-trial, it was clear that normalized calculation would do. However, in the cross sectional analyses, the nutritional quality of the intervention diet differed to the control diet, indicating nutritional quality index was not related to any health outcomes, regardless if that at least in terms of self-reported intake, the dietary intervention and control calculated as nutrient density or as nutrient adequacy adjusted for energy diet were successfully implemented. Some previous studies, such as the intake. dietary intervention by Skoldstam et al. (16), instead of using an active comparator, recruited selectively based on habitual dietary intake in order to achieve a control group with a habitual diet resembling a typical western diet. A drawback with a passive control design, however, is the risk of introducing variations between intervention and control beyond the designed altered dietary intake. In this aspect, the ADIRA-trial likely more accurately compared two different dietary regimens. As further outlined in paper IV, there were several interactions between habitual nutritional quality and health outcomes in the ADIRA-trial. In a dietary intervention trial, it would seem plausible to expect a varying response dependent on the baseline dietary intake. However, none of the subgroup analyses, based on habitual nutritional quality, indicated significant results between the intervention and the control diet. Thus, it remains hard to draw any conclusions. While most individuals improved in nutritional quality during the intervention diet period, the response to the control diet was more diverse (Figure 15). Further, there was at least one outlier who strongly decreased in nutritional quality during the intervention diet periods. In this particular case, the calculated energy intake averaged at 600 kcal/day after the intervention diet 56 57 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander 5.5 THE ADIRA STUDY DESIGN 5.5.3 RECRUITMENT Another parameter in the ADIRA-trial potentially affecting external validity is 5.5.1 DIETARY INTERVENTION FOODS the recruitment process; 15% of those invited by letter agreed to participate, One key difference between the ADIRA-trial and most other Mediterranean- and a selection bias is likely. One could expect that participants that reply to like diet interventions, is that no extra virgin olive oil was used. Extra virgin letters of invitation have an above-average interest in diet and health- olive oil is suggested to induce a range of anti-inflammatory effects due to its encouraging lifestyle approaches. It could be postulated that this selection also high content of polyphenols (110), and the European Food Safety Authority predispose to a higher compliance than that of the average patient with RA. has approved a health claim that extra virgin olive oil protects LDL-particles This situation warrants the use of an active comparator design in order to from oxidation (111). In contrast, canola oil, produced locally in Sweden, was compare the two different dietary approaches. This does however also hinder used in the ADIRA-trial. Even though data on immunomodulatory potential of us from speculating on the effect of a dietary intervention compared to the canola oil are rather scarce, it does not appear to be comparable to refined olive oil (112), or extra virgin olive oil (113) in other patient populations. The fatty habitual diet of patients with RA in a clinical setting. acid composition, including omega-3 fatty acids, is however excellent and there are documented positive effects on the blood lipid profile from intake of The mean DAS28 value of the 50 participants entering the ADIRA-trial was at canola oil (114). This component of the intervention could thus in part confer 3.8. There are no systematic national statistics on the median value of DAS28 the beneficial effect on blood lipids observed in the ADIRA-trial. in patients with RA in Sweden (personal communication, SRQ), but in general, a value above 3.2 instigates pharmacological intervention with the aim of The probiotic strain used in the ADIRA-trial, Lactobacillus plantarum 299v/ reaching remission (123). It could thus seem questionable to include patients DSM 9843, has not previously been tried on patients with RA. However, in an that needed medical intervention. However, a rheumatologist examined each uncontrolled trial on men with coronary artery disease, a decreased level of patient at inclusion and only those assessed as adequately medicated were proinflammatory cytokines (115) and gene expressions in PBMCs (116) was included in the study. noted. Naturally, the lack of a proper control group and differing patient characteristics call for a cautious interpretation. Previous interventions with In general, as described in a wide range of datasets, for extreme measurements probiotic strains indicating beneficial effects on patients with RA include within a variable that can vary, values tend to normalize over time, a statistical Bifidobacterium bifidum, Lactobacillus acidophilus and Lactobacillus casei concept referred to as regression to the mean (124). The ADIRA-trial recruited combined (117, 118), as well as Lactobacillus casei 01 (119) who all reduced patients that were estimated to have a DAS28 score of at least 2.6, which CRP. In the ADIRA-trial, we lack data to specifically confer effects to the indeed is a variable that can vary over time. The rationale was to recruit probiotic supplementation. patients with ongoing inflammation, so as to be able to produce a decrease. This approach also has a downside; while the differences between intervention 5.5.2 PARTICIPANTS and control diet periods in the group as a whole were not significant (89), there The participants in the ADIRA-trial were mostly older females, while was indeed an improvement in DAS28 between pre and post values regardless incidence of RA is increased in this group, the results might not be of treatment during the first diet period (period 1 mean difference; -0.42, P < representable to different patient groups. Another example is socioeconomic 0.001, n= 47), but no such improvement was seen in the second diet period factors; educational level appears to relate to health outcome and life (period 2 mean difference -0.01, P = 0.947, n = 44), analyzed by a paired t- expectancy among patients with RA (120). In the ADIRA-trial, educational test. This pattern is compatible with the concept of a regression to the mean level was generally high; 49% of those who completed at least one diet period effect and could affect all outcomes analyzed. However, the randomized study had a university level education. This is similar to the general population design and the statistical analysis where only the differences between dietary average of 44% (121), but perhaps slightly higher than the average among treatments are interpreted, adjusted for diet period, should counteract this females in a similar age range where 42% has a university level education phenomenon in the published papers. Further, while ESR concentration is a (122). These factors could limit the external validity. significant factor in DAS28, the improvements during this first period was mainly related to improvements in tender joints (data not shown). This is a 58 59 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander subjective factor that perhaps could be affected by psychological well being, correction answers a largely irrelevant question, since focus shifts from each which could have improved by being part of a clinical trial. individual parameter to an assessment of whether two groups are identical in all possible aspects (125). In the papers included in this thesis, it was decided 5.5.4 STATISTICAL CONSIDERATIONS to not correct for multiple hypothesis test in order not to be overly conservative All the articles in this thesis have been produced without correction for and introduce type-2 errors, and hence results need to be interpreted multiple hypothesis tests, and as such are mainly to be viewed as indicative accordingly. and exploratory rather than conclusive; more studies are needed to replicate and verify our findings. Another common line of thought seems to be that it is harder to find significant findings in smaller datasets, thereby concluding that significant results from In paper II, there was an exploratory analysis on a wide range of biomarkers smaller datasets are robust. However, while it might be harder to find true related to inflammation. There was a hypothesis that IL-6 or TNF could be differences in small datasets, the risk of chance findings from clustering by affected by the intervention, but there was no hypothesis that chemokines chance also increases with a limited number of participants. It would be would be affected. While the results seem plausible, the multi-testing approach reasonable to replicate our results with appropriate statistical power for each needs to be considered. The chance of type-1 error can be computed by (1 – outcome of interest. In the ADIRA-trial, the power calculation was performed (1-α)n, where n is the number of hypothesis tests performed. In the case of the with the aim of investigating DAS28, an outcome variable not included in the multiplex assay, 72 tests were done in two group selections, totaling in 144 papers this thesis is based upon; all outcomes included are secondary tests, each with an α-value of 0.05. This translates to about 99.9% probability outcomes. To verify our results, a larger study sample would likely be needed. of finding at least one marker that is significantly different by pure chance (Figure 16). Still, there is a scarcity of dietary intervention studies in patients with RA. The robust study design of the ADIRA-trial, in an area where not many clinical trials have been performed, motivates for exploration of secondary outcomes despite the risk of introducing type-1 errors. 5.5.5 MEASUREMENTS OF OUTCOMES In the ADIRA-trial, measurements were taken before and after each dietary period, which allowed for an adequate evaluation of effects. It would also have been interesting to take multiple measurements over time, analyzing the trend within each diet period as repeated correlated measures or perhaps as area under the curve. With this approach, the statistical analysis could have gained in strength, and it would have been possible to perform mediator analyses. There was no objective measure of individual compliance in the ADIRA-trial; Figure 16. The risk of type-1 error by increasing number of hypothesis tests interviews with participants were performed mid-period and food records were assessed before and after each diet period. Interview with study staff about There are several methods to counteract multiple hypothesis tests, and the most one´s compliance to the study protocol could lead to overly positive results. popular is probably the Bonferroni correction. This method suggests dividing Food records are subject to two phases of bias; first by the participant to α by the number of hypothesis tests. The most appropriate dataset in this thesis accurately report intake and to maintain a habitual diet while recording to apply correction for multiple hypothesis tests is arguable the multiplex assay everything eaten, secondly by the dietitian to interpret types and amounts of performed in paper II. By using a Bonferroni correction, the significance level foods when not precisely detailed. It would thus be interesting to analyze would be set to α < 0.000347, and no results from the multiplex assay would developments in objective biomarkers of compliance and response, such as for be deemed significant. However, it has also been argued that the Bonferroni example alkylresorcinols for whole grain intake (126), or fatty acid 61 60 Nutritional impact on health in patients with Rheumatoid Arthritis Erik Hulander composition of red blood cells, or carotenoids for vegetable intake. Surely, foods. Arguable, the postprandial state might be more representable to the some of these markers will be revealed in forthcoming articles based on the habitual state of humans. ADIRA-trial. However, we did already assess changes in plasma fatty acid composition, which indicated that, on a group level, the dietary regimens Furthermore, in addition to examining dietary effects during weight stability, significantly affected the composition of fatty acids in the circulation. it would also be interesting to perform a study aimed at optimizing body Even though we believe that the washout period was sufficient in length to composition in obese patients with RA. Ideally by a combined diet and exercise facilitate a normalization of habitual intake, metabolic traces of a nutritional regimen. This would allow for determining the clinical relevance of metabolic intervention could well span for a prolonged time period. As an example, TG normalization, perhaps in comparison to the effect of pharmacological age in human adipocytes has a turnover rate of approximately 1.6 years in interventions. homeostatic conditions (127). It is thus possible that stronger effects could have been presented if ADIRA had a parallel study design, eliminating any potential for carry-over effects. However, this would require a larger sample size. In the ADIRA-trial, all eligible participants in the region were invited, as such it would not have been plausible to employ a parallel design without incorporating several study sites or simplifying the study protocol. In order to reduce interindividual variation and thereby maximize statistical power in the ADIRA-trial, without employing a multi-center design, a cross-over design was chosen. 5.6 CONCLUSION & FUTURE PERSPECTIVES Diet matters. Dietary intake can modify CVD risk profile, biomarkers of inflammation, and body composition in patients with RA. Even though the data presented in this thesis suffer from multiple hypothesis tests, it reveals new novel findings, and indications of efficacy that merits further replication and longer term evaluation. In terms of determining the long term benefit of dietary optimization for patients with RA, an intervention in the clinical setting could be done. For example, newly diagnosed patients could be invited and randomized to dietary counselling or control in a parallel design. This would allow for a proper evaluation of the feasibility of upholding compliance over time, and long term outcomes. In addition to evaluating the long term consequences of dietary modification in patients with RA, it would also be interesting to shed light on the acute effects of dietary intake in the postprandial phase. During screening, several patients in the ADIRA- and PIRA-trials reported feeling worse after intake of certain foods (data not published), which they habitually avoided. It would thus be interesting to evaluate the postprandial state after intake of these types of 62 63 Nutritional impact on health in patients with Rheumatoid Arthritis ACKNOWLEDGEMENT All the Co-authors to the papers. Without constructive feedback, none of the papers would have been what they are. First and foremost, I extend my gratitude to all the patients who voluntarily My family; my parents, and my older siblings Ragnar, Anne, Eva and applied to participate in the clinical trials PIRA and ADIRA. Rickard. Secondly, I would like to extend my sincerest gratitude to my supervisors All of my personal friends, who have supported me in a multitude of ways Helen Lindqvist, Linnea Bärebring, Inger Gjertsson and Anna Winkvist during these years, and lifted my spirit in the time of a global pandemic and who welcomed me to the research group, for believing in me, for supporting me with interest, expertise and patience all along the way. Thank you! cancelled study plans. I would like to thank Michael Hoppe, my neighbor in the office corridor, for all the valuable help when starting up the PIRA-trial, ordering lab equipment, getting started with DXA-measurements, and so on... To all colleagues at the department of internal medicine and clinical nutrition, it’s a great environment with lovely intelligent persons. To my fellow PhD-student colleagues at the department of internal medicine and clinical nutrition, you are all priceless! Jörgen Isgaard for being a cohesive and positive force at the department of internal medicine and clinical nutrition. My old friend Patrik Hansson, with whom I share several research interests, for moral support and discussions about future possibilities. My old friend Zhihui Wang in Oslo, for encouragements, friendship and an inspiring mindset. Alexandra Abramsson, whom I met through the PhD-student mentoring program, for being there and helping me bounce ideas. Anders Pederson at the NMR-center for quantification of lipids and amino acids in serum. Eva Flenner at the department of clinical chemistry at the Sahlgrenska University Hospital, who despite a high workload and many other tasks at hand, answered my many questions on analyses. Elisabet Bergh Börgdal at Skåne University Hospital in Lund, for always looking out for the best of her employees, and who, for example, let me visit Gothenburg for several days back in 2018 on paid leave. 64 65 Nutritional impact on health in patients with Rheumatoid Arthritis REFERENCES patients and effects of a diet adjusted in energy intake, fish-meal, and antioxidants. Scandinavian journal of rheumatology. 1996;25(5):325-30. 12. Holst-Jensen SE, Pfeiffer-Jensen M, Monsrud M, Tarp U, Buus A, Hessov I, et al. Treatment of rheumatoid arthritis with a peptide diet: a 1. McKellar G, Morrison E, McEntegart A, Hampson R, Tierney A, randomized, controlled trial. Scandinavian journal of rheumatology. Mackle G, et al. A pilot study of a Mediterranean-type diet intervention in female 1998;27(5):329-36. patients with rheumatoid arthritis living in areas of social deprivation in Glasgow. Annals of the rheumatic diseases. 2007;66(9):1239-43. 13. Sarzi-Puttini P, Comi D, Boccassini L, Muzzupappa S, Turiel M, Panni B, et al. Diet therapy for rheumatoid arthritis. A controlled double-blind 2. Nenonen MT, Helve TA, Rauma AL, Hanninen OO. Uncooked, study of two different dietary regimens. Scandinavian journal of rheumatology. lactobacilli-rich, vegan food and rheumatoid arthritis. British journal of 2000;29(5):302-7. rheumatology. 1998;37(3):274-81. 14. Kavanaghi R, Workman E, Nash P, Smith M, Hazleman BL, 3. Karjalainen T, Adiels M, Björck L, Cooney M-T, Graham I, Perk Hunter JO. The effects of elemental diet and subsequent food reintroduction on J, et al. An evaluation of the performance of SCORE Sweden 2015 in estimating rheumatoid arthritis. British journal of rheumatology. 1995;34(3):270-3. cardiovascular risk:The Northern Sweden MONICA Study 1999–2014. 15. Kjeldsen-Kragh J, Haugen M, Borchgrevink CF, Laerum E, Eek European Journal of Preventive Cardiology. 2017;24(1):103-10. M, Mowinkel P, et al. Controlled trial of fasting and one-year vegetarian diet in 4. Schutz Y, Kyle UU, Pichard C. Fat-free mass index and fat mass rheumatoid arthritis. Lancet (London, England). 1991;338(8772):899-902. index percentiles in Caucasians aged 18-98 y. International journal of obesity 16. Skoldstam L, Hagfors L, Johansson G. An experimental study of and related metabolic disorders : journal of the International Association for the a Mediterranean diet intervention for patients with rheumatoid arthritis. Annals Study of Obesity. 2002;26(7):953-60. of the rheumatic diseases. 2003;62(3):208-14. 5. Barnard ND, Levin S, Crosby L, Flores R, Holubkov R, 17. Kremer JM, Bigauoette J, Michalek AV, Timchalk MA, Lininger Kahleova H. A Randomized, Crossover Trial of a Nutritional Intervention for L, Rynes RI, et al. Effects of manipulation of dietary fatty acids on clinical Rheumatoid Arthritis. American Journal of Lifestyle manifestations of rheumatoid arthritis. Lancet (London, England). Medicine.0(0):15598276221081819. 1985;1(8422):184-7. 6. Dawczynski C, Schubert R, Hein G, Muller A, Eidner T, 18. van de Laar MA, van der Korst JK. Food intolerance in Vogelsang H, et al. Long-term moderate intervention with n-3 long-chain PUFA- rheumatoid arthritis. I. A double blind, controlled trial of the clinical effects of supplemented dairy products: effects on pathophysiological biomarkers in elimination of milk allergens and azo dyes. Annals of the rheumatic diseases. patients with rheumatoid arthritis. The British journal of nutrition. 1992;51(3):298-302. 2009;101(10):1517-26. 19. Lindqvist HM, Gjertsson I, Eneljung T, Winkvist A. Influence of 7. Panush RS, Carter RL, Katz P, Kowsari B, Longley S, Finnie S. Blue Mussel (Mytilus edulis) Intake on Disease Activity in Female Patients with Diet therapy for rheumatoid arthritis. Arthritis and rheumatism. Rheumatoid Arthritis: The MIRA Randomized Cross-Over Dietary Intervention. 1983;26(4):462-71. Nutrients. 2018;10(4). 8. Becker W. Indikatorer för bra matvanor : resultat från 20. Hayter SM, Cook MC. Updated assessment of the prevalence, intervjuundersökningar 2005 och 2006. Uppsala: Livsmedelsverket; 2007. spectrum and case definition of autoimmune disease. Autoimmunity reviews. 2012;11(10):754-65. 9. Denissov LN, Sharafetdinov K, Samsonov MA. On the medicinal efficacy of dietetic therapy in patients with rheumatoid arthritis. Int J Clin 21. Minichiello E, Semerano L, Boissier M-C. Time trends in the Pharmacol Res. 1992;12(1):19-25. incidence, prevalence, and severity of rheumatoid arthritis: A systematic literature review. Joint Bone Spine. 2016;83(6):625-30. 10. Elkan AC, Sjoberg B, Kolsrud B, Ringertz B, Hafstrom I, Frostegard J. Gluten-free vegan diet induces decreased LDL and oxidized LDL 22. Smolen JS, Aletaha D, Barton A, Burmester GR, Emery P, levels and raised atheroprotective natural antibodies against phosphorylcholine Firestein GS, et al. Rheumatoid arthritis. Nature Reviews Disease Primers. in patients with rheumatoid arthritis: a randomized study. Arthritis research & 2018;4(1):18001. therapy. 2008;10(2):R34. 23. Marchand NE, Sparks JA, Tedeschi SK, Malspeis S, 11. Hansen GV, Nielsen L, Kluger E, Thysen M, Emmertsen H, Costenbader KH, Karlson EW, et al. Abdominal Obesity in Comparison with Stengaard-Pedersen K, et al. Nutritional status of Danish rheumatoid arthritis 67 66 Nutritional impact on health in patients with Rheumatoid Arthritis General Obesity and Risk of Developing Rheumatoid Arthritis in Women. The 35. Cederholm T, Bosaeus I, Barazzoni R, Bauer J, Van Gossum A, Journal of rheumatology. 2021;48(2):165-73. Klek S, et al. Diagnostic criteria for malnutrition - An ESPEN Consensus 24. Linauskas A, Overvad K, Symmons D, Johansen MB, Stengaard- Statement. Clinical nutrition (Edinburgh, Scotland). 2015;34(3):335-40. Pedersen K, de Thurah A. Body Fat Percentage, Waist Circumference, and 36. Cederholm T, Barazzoni R, Austin P, Ballmer P, Biolo G, Obesity As Risk Factors for Rheumatoid Arthritis: A Danish Cohort Study. Bischoff SC, et al. ESPEN guidelines on definitions and terminology of clinical Arthritis care & research. 2019;71(6):777-86. nutrition. Clinical nutrition (Edinburgh, Scotland). 2017;36(1):49-64. 25. Lu B, Hiraki LT, Sparks JA, Malspeis S, Chen CY, Awosogba JA, 37. Santo RCE, Fernandes KZ, Lora PS, Filippin LI, Xavier RM. et al. Being overweight or obese and risk of developing rheumatoid arthritis Prevalence of rheumatoid cachexia in rheumatoid arthritis: a systematic review among women: a prospective cohort study. Annals of the rheumatic diseases. and meta-analysis. Journal of cachexia, sarcopenia and muscle. 2018;9(5):816- 2014;73(11):1914-22. 25. 26. Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT, Bingham 38. Eriksson JK, Neovius M, Ernestam S, Lindblad S, Simard JF, III CO, et al. 2010 Rheumatoid arthritis classification criteria: An American Askling J. Incidence of Rheumatoid Arthritis in Sweden: A Nationwide College of Rheumatology/European League Against Rheumatism collaborative Population-Based Assessment of Incidence, Its Determinants, and Treatment initiative. Arthritis & Rheumatism. 2010;62(9):2569-81. Penetration. Arthritis care & research. 2013;65(6):870-8. 27. Radner H, Neogi T, Smolen JS, Aletaha D. Performance of the 39. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, Boirie Y, Cederholm T, 2010 ACR/EULAR classification criteria for rheumatoid arthritis: a systematic Landi F, et al. Sarcopenia: European consensus on definition and diagnosis: literature review. Annals of the rheumatic diseases. 2014;73(1):114-23. Report of the European Working Group on Sarcopenia in Older People. Age and 28. Radner H, Chatzidionysiou K, Nikiphorou E, Gossec L, Hyrich Ageing. 2010;39(4):412-23. KL, Zabalan C, et al. 2017 EULAR recommendations for a core data set to 40. Elkan A-C, Håkansson N, Frostegård J, Cederholm T, Hafström support observational research and clinical care in rheumatoid arthritis. Annals I. Rheumatoid cachexia is associated with dyslipidemia and low levels of of the rheumatic diseases. 2018;77(4):476-9. atheroprotective natural antibodies against phosphorylcholine but not with 29. England BR, Tiong BK, Bergman MJ, Curtis JR, Kazi S, Mikuls dietary fat in patients with rheumatoid arthritis: a cross-sectional study. Arthritis TR, et al. 2019 Update of the American College of Rheumatology Recommended research & therapy. 2009;11(2):R37. Rheumatoid Arthritis Disease Activity Measures. Arthritis care & research. 41. Engvall IL, Elkan AC, Tengstrand B, Cederholm T, Brismar K, 2019;71(12):1540-55. Hafstrom I. Cachexia in rheumatoid arthritis is associated with inflammatory 30. Ware JEJ, Sherbourne CD. The MOS 36-ltem Short-Form activity, physical disability, and low bioavailable insulin-like growth factor. Health Survey (SF-36): I. Conceptual Framework and Item Selection. Medical Scandinavian journal of rheumatology. 2008;37(5):321-8. Care. 1992;30(6):473-83. 42. Elkan AC, Engvall IL, Tengstrand B, Cederholm T, Hafström I. 31. Matcham F, Scott IC, Rayner L, Hotopf M, Kingsley GH, Norton Malnutrition in women with rheumatoid arthritis is not revealed by clinical S, et al. The impact of rheumatoid arthritis on quality-of-life assessed using the anthropometrical measurements or nutritional evaluation tools. European SF-36: A systematic review and meta-analysis. Seminars in Arthritis and journal of clinical nutrition. 2008;62(10):1239-47. Rheumatism. 2014;44(2):123-30. 43. van Bokhorst - de van der Schueren MAE, Konijn NPC, Bultink 32. Widdifield J, Paterson JM, Huang A, Bernatsky S. Causes of IEM, Lems WF, Earthman CP, van Tuyl LHD. Relevance of the new pre- Death in Rheumatoid Arthritis: How Do They Compare to the General cachexia and cachexia definitions for patients with rheumatoid arthritis. Clinical Population? Arthritis care & research. 2018;70(12):1748-55. Nutrition. 2012;31(6):1008-10. 33. Jacobsson LTH, Lindroth Y, Marsal L, Juran E, Bergström U, 44. Visseren FLJ, Mach F, Smulders YM, Carballo D, Koskinas KC, Kobelt G. Rheumatoid arthritis: what does it cost and what factors are driving Bäck M, et al. 2021 ESC Guidelines on cardiovascular disease prevention in those costs? Results of a survey in a community‐derived population in Malmö, clinical practice: Developed by the Task Force for cardiovascular disease Sweden. Scandinavian journal of rheumatology. 2007;36(3):179-83. prevention in clinical practice with representatives of the European Society of Cardiology and 12 medical societies With the special contribution of the 34. Hsieh P-H, Wu O, Geue C, McIntosh E, McInnes IB, Siebert S. European Association of Preventive Cardiology (EAPC). European Heart Economic burden of rheumatoid arthritis: a systematic review of literature in Journal. 2021;42(34):3227-337. biologic era. Annals of the rheumatic diseases. 2020;79(6):771-7. 45. Agca R, Heslinga SC, Rollefstad S, Heslinga M, McInnes IB, Peters MJ, et al. EULAR recommendations for cardiovascular disease risk 68 69 Nutritional impact on health in patients with Rheumatoid Arthritis management in patients with rheumatoid arthritis and other forms of 55. Tillett WS, Francis T. SEROLOGICAL REACTIONS IN inflammatory joint disorders: 2015/2016 update. Annals of the rheumatic PNEUMONIA WITH A NON-PROTEIN SOMATIC FRACTION OF diseases. 2017;76(1):17-28. PNEUMOCOCCUS. J Exp Med. 1930;52(4):561-71. 46. Aburto NJ, Hanson S, Gutierrez H, Hooper L, Elliott P, 56. Tishkowski K, Gupta V. Erythrocyte Sedimentation Rate. Cappuccio FP. Effect of increased potassium intake on cardiovascular risk StatPearls. Treasure Island (FL): StatPearls Publishing, Copyright © 2021, factors and disease: systematic review and meta-analyses. BMJ : British Medical StatPearls Publishing LLC.; 2021. Journal. 2013;346:f1378. 57. Theodorsson E, Berggren Söderlund M, Laurell C-B. Laurells 47. Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Klinisk kemi i praktisk medicin. Tionde upplagan ed: Lund : Studentlitteratur; Bruckert E, et al. Low-density lipoproteins cause atherosclerotic cardiovascular 2018. disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus 58. Punt J. Kuby immunology. Eighth edition ed. Stranford SA, Panel. European Heart Journal. 2017;38(32):2459-72. Jones PP, Owen JA, editors: New York : Macmillan Education; 2019. 48. Silverman MG, Ference BA, Im K, Wiviott SD, Giugliano RP, 59. Kany S, Vollrath JT, Relja B. Cytokines in Inflammatory Grundy SM, et al. Association Between Lowering LDL-C and Cardiovascular Disease. International journal of molecular sciences. 2019;20(23):6008. Risk Reduction Among Different Therapeutic Interventions: A Systematic Review 60. Ganapathi MK, Rzewnicki D, Samols D, Jiang SL, Kushner I. and Meta-analysis. Jama. 2016;316(12):1289-97. Effect of combinations of cytokines and hormones on synthesis of serum amyloid 49. Choy E, Ganeshalingam K, Semb AG, Szekanecz Z, A and C-reactive protein in Hep 3B cells. The Journal of Immunology. Nurmohamed M. Cardiovascular risk in rheumatoid arthritis: recent advances in 1991;147(4):1261-5. the understanding of the pivotal role of inflammation, risk predictors and the 61. Abbas AK. Basic immunology : functions and disorders of the impact of treatment. Rheumatology (Oxford, England). 2014;53(12):2143-54. immune system. Sixth edition ed. Lichtman AH, Pillai S, Baker DL, editors: 50. Myasoedova E, Crowson CS, Kremers HM, Fitz-Gibbon PD, Philadelphia : Elsevier; 2020. Therneau TM, Gabriel SE. Total cholesterol and LDL levels decrease before 62. Kvalitetsregister SR. Årsrapport 2020 2021 [updated 2021-05- rheumatoid arthritis. Annals of the rheumatic diseases. 2010;69(7):1310-4. 06. Available from: https://srq.nu/wp- 51. Marston NA, Giugliano RP, Im K, Silverman MG, O'Donoghue content/uploads/2021/05/A%CC%8Arsrapport-2020.pdf. ML, Wiviott SD, et al. Association Between Triglyceride Lowering and 63. Ljung L, Rantapää-Dahlqvist S, Jacobsson LT, Askling J. Reduction of Cardiovascular Risk Across Multiple Lipid-Lowering Therapeutic Response to biological treatment and subsequent risk of coronary events in Classes: A Systematic Review and Meta-Regression Analysis of Randomized rheumatoid arthritis. Annals of the rheumatic diseases. 2016;75(12):2087-94. Controlled Trials. Circulation. 2019;140(16):1308-17. 64. Yu C, Jin S, Wang Y, Jiang N, Wu C, Wang Q, et al. Remission 52. Ference BA, Kastelein JJP, Ray KK, Ginsberg HN, Chapman rate and predictors of remission in patients with rheumatoid arthritis under MJ, Packard CJ, et al. Association of Triglyceride-Lowering LPL Variants and treat-to-target strategy in real-world studies: a systematic review and meta- LDL-C–Lowering LDLR Variants With Risk of Coronary Heart Disease. Jama. analysis. Clinical rheumatology. 2019;38(3):727-38. 2019;321(4):364-73. 65. Tedeschi SK, Frits M, Cui J, Zhang ZZ, Mahmoud T, Iannaccone 53. Marston NA, Giugliano RP, Melloni GEM, Park J-G, Morrill V, C, et al. Diet and Rheumatoid Arthritis Symptoms: Survey Results From a Blazing MA, et al. Association of Apolipoprotein B–Containing Lipoproteins and Rheumatoid Arthritis Registry. Arthritis care & research. 2017;69(12):1920-5. Risk of Myocardial Infarction in Individuals With and Without Atherosclerosis: Distinguishing Between Particle Concentration, Type, and Content. JAMA 66. Haugen M, Kjeldsen-Kragh J, Nordvåg B, Førre Ø. Diet and Cardiology. 2022;7(3):250-6. disease symptoms in rheumatic diseases — Results of a questionnaire based survey. Clin Rheumatol. 1991;10(4):401-7. 54. Hollan I, Ronda N, Dessein P, Agewall S, Karpouzas G, Tamargo J, et al. Lipid management in rheumatoid arthritis: a position paper of 67. Berenbaum F, Chauvin P, Hudry C, Mathoret-Philibert F, the Working Group on Cardiovascular Pharmacotherapy of the European Poussiere M, De Chalus T, et al. Fears and beliefs in rheumatoid arthritis and Society of Cardiology. European Heart Journal - Cardiovascular spondyloarthritis: a qualitative study. PloS one. 2014;9(12):e114350. Pharmacotherapy. 2019;6(2):104-14. 68. Grygielska J, Klak A, Raciborski F, Manczak M. Nutrition and quality of life referring to physical abilities - a comparative analysis of a 71 70 Nutritional impact on health in patients with Rheumatoid Arthritis questionnaire study of patients with rheumatoid arthritis and osteoarthritis. 81. Marcora S, Lemmey A, Maddison P. Dietary treatment of Reumatologia. 2017;55(5):222-9. rheumatoid cachexia with beta-hydroxy-beta-methylbutyrate, glutamine and 69. Salminen E, Heikkilä S, Poussa T, Lagström H, Saario R, arginine: a randomised controlled trial. Clinical nutrition (Edinburgh, Salminen S. Female Patients Tend to Alter Their Diet Following the Diagnosis of Scotland). 2005;24(3):442-54. Rheumatoid Arthritis and Breast Cancer. Preventive Medicine. 2002;34(5):529- 82. Wilkinson TJ, Lemmey AB, Jones JG, Sheikh F, Ahmad YA, 35. Chitale S, et al. Can Creatine Supplementation Improve Body Composition and 70. Tedeschi SK, Bathon JM, Giles JT, Lin TC, Yoshida K, Solomon Objective Physical Function in Rheumatoid Arthritis Patients? A Randomized DH. Relationship Between Fish Consumption and Disease Activity in Controlled Trial. Arthritis care & research. 2016;68(6):729-37. Rheumatoid Arthritis. Arthritis care & research. 2018;70(3):327-32. 83. Jiménez B, Holmes E, Heude C, Tolson RF, Harvey N, Lodge 71. Edefonti V, Parpinel M, Ferraroni M, Boracchi P, Schioppo T, SL, et al. Quantitative Lipoprotein Subclass and Low Molecular Weight Scotti I, et al. A Posteriori Dietary Patterns and Rheumatoid Arthritis Disease Metabolite Analysis in Human Serum and Plasma by 1H NMR Spectroscopy in a Activity: A Beneficial Role of Vegetable and Animal Unsaturated Fatty Acids. Multilaboratory Trial. Analytical chemistry. 2018;90(20):11962-71. Nutrients. 2020;12(12). 84. Fisk HL, West AL, Childs CE, Burdge GC, Calder PC. The use 72. McGarrity-Yoder ME, Insel KC, Crane TE, Pace TWW. Diet of gas chromatography to analyze compositional changes of fatty acids in rat quality and disease activity in rheumatoid arthritis. Nutr Health. liver tissue during pregnancy. Journal of visualized experiments : JoVE. 2021:2601060211044311. 2014(85). 73. Calder PC. n-3 PUFA and inflammation: from membrane to 85. Assarsson E, Lundberg M, Holmquist G, Björkesten J, Thorsen nucleus and from bench to bedside. Proceedings of the Nutrition Society. SB, Ekman D, et al. Homogenous 96-plex PEA immunoassay exhibiting high 2020;79(4):404-16. sensitivity, specificity, and excellent scalability. PloS one. 2014;9(4):e95192. 74. Ferrucci L, Cherubini A, Bandinelli S, Bartali B, Corsi A, 86. Fulgoni VL, 3rd, Keast DR, Drewnowski A. Development and Lauretani F, et al. Relationship of Plasma Polyunsaturated Fatty Acids to validation of the nutrient-rich foods index: a tool to measure nutritional quality Circulating Inflammatory Markers. The Journal of Clinical Endocrinology & of foods. The Journal of nutrition. 2009;139(8):1549-54. Metabolism. 2006;91(2):439-46. 87. Strid A, Johansson I, Bianchi M, Sonesson U, Hallström E, 75. Davidson MH. Mechanisms for the Hypotriglyceridemic Effect Lindahl B, et al. Diets benefiting health and climate relate to longevity in of Marine Omega-3 Fatty Acids. The American Journal of Cardiology. northern Sweden. The American journal of clinical nutrition. 2021. 2006;98(4, Supplement 1):27-33. 88. Wareham NJ, Jakes RW, Rennie KL, Schuit J, Mitchell J, 76. Tosi F, Sartori F, Guarini P, Olivieri O, Martinelli N. Delta-5 Hennings S, et al. Validity and repeatability of a simple index derived from the and Delta-6 Desaturases: Crucial Enzymes in Polyunsaturated Fatty Acid- short physical activity questionnaire used in the European Prospective Related Pathways with Pleiotropic Influences in Health and Disease. In: Camps Investigation into Cancer and Nutrition (EPIC) study. Public health nutrition. J, editor. Oxidative Stress and Inflammation in Non-communicable Diseases - 2003;6(4):407-13. Molecular Mechanisms and Perspectives in Therapeutics. Cham: Springer 89. Vadell AKE, Barebring L, Hulander E, Gjertsson I, Lindqvist International Publishing; 2014. p. 61-81. HM, Winkvist A. Anti-inflammatory Diet In Rheumatoid Arthritis (ADIRA)-a 77. Barber TM, Kabisch S, Pfeiffer AFH, Weickert MO. The Health randomized, controlled crossover trial indicating effects on disease activity. The Benefits of Dietary Fibre. Nutrients. 2020;12(10):3209. American journal of clinical nutrition. 2020. 78. Dietschy JM. Dietary Fatty Acids and the Regulation of Plasma 90. Hulander E, Bärebring L, Turesson Wadell A, Gjertsson I, Low Density Lipoprotein Cholesterol Concentrations. The Journal of nutrition. Calder PC, Winkvist A, et al. Diet intervention improves cardiovascular profile 1998;128(2):444S-8S. in patients with rheumatoid arthritis: results from the randomized controlled cross-over trial ADIRA. Nutrition journal. 2021;20(1):9. 79. Hasan B, Nayfeh T, Alzuabi M, Wang Z, Kuchkuntla AR, Prokop LJ, et al. Weight Loss and Serum Lipids in Overweight and Obese Adults: A 91. Estruch R, Ros E, Salas-Salvadó J, Covas MI, Corella D, Arós Systematic Review and Meta-Analysis. The Journal of Clinical Endocrinology & F, et al. Primary Prevention of Cardiovascular Disease with a Mediterranean Metabolism. 2020;105(12):3695-703. Diet Supplemented with Extra-Virgin Olive Oil or Nuts. The New England journal of medicine. 2018;378(25):e34. 80. Bianchi VE. Weight loss is a critical factor to reduce inflammation. Clinical Nutrition ESPEN. 2018;28:21-35. 72 73 Nutritional impact on health in patients with Rheumatoid Arthritis 92. Appel LJ, Van Horn L. Did the PREDIMED trial test a 104. Bowman L, Hopewell JC, Chen F, Wallendszus K, Stevens W, Mediterranean diet? The New England journal of medicine. 2013;368(14):1353- Collins R, et al. Effects of Anacetrapib in Patients with Atherosclerotic Vascular 4. Disease. The New England journal of medicine. 2017;377(13):1217-27. 93. Estruch R, Martínez-González MA, Corella D, Salas-Salvadó J, 105. Hulander E, Bärebring L, Turesson Wadell A, Gjertsson I, Ruiz-Gutiérrez V, Covas MI, et al. Effects of a Mediterranean-style diet on Calder PC, Winkvist A, et al. Proposed Anti-Inflammatory Diet Reduces cardiovascular risk factors: a randomized trial. Annals of internal medicine. Inflammation in Compliant, Weight-Stable Patients with Rheumatoid Arthritis in 2006;145(1):1-11. a Randomized Controlled Crossover Trial. The Journal of nutrition. 2021. 94. Khan SU, Khan MU, Valavoor S, Khan MS, Okunrintemi V, 106. Mateen S, Moin S, Khan AQ, Zafar A, Fatima N. Increased Mamas MA, et al. Association of lowering apolipoprotein B with cardiovascular Reactive Oxygen Species Formation and Oxidative Stress in Rheumatoid outcomes across various lipid-lowering therapies: Systematic review and meta- Arthritis. PloS one. 2016;11(4):e0152925. analysis of trials. European Journal of Preventive Cardiology. 2020;27(12):1255-68. 107. Dürholz K, Hofmann J, Iljazovic A, Häger J, Lucas S, Sarter K, et al. Dietary Short-Term Fiber Interventions in Arthritis Patients Increase 95. Fernandez ML, West KL. Mechanisms by which dietary fatty Systemic SCFA Levels and Regulate Inflammation. Nutrients. 2020;12(10). acids modulate plasma lipids. The Journal of nutrition. 2005;135(9):2075-8. 108. Nunemaker CS, Chung HG, Verrilli GM, Corbin KL, Upadhye 96. Hara H, Haga S, Aoyama Y, Kiriyama S. Short-chain fatty acids A, Sharma PR. Increased serum CXCL1 and CXCL5 are linked to obesity, suppress cholesterol synthesis in rat liver and intestine. The Journal of nutrition. hyperglycemia, and impaired islet function. Journal of Endocrinology. 1999;129(5):942-8. 2014;222(2):267-76. 97. Huang Y, Wang X, Wang J, Wu F, Sui Y, Yang L, et al. 109. Lindqvist HM, Wallengren O, Eriksson A, Hulander E, Winkvist Lactobacillus plantarum strains as potential probiotic cultures with cholesterol- A, Bärebring L. Validity of bioimpedance for assessment of fat-free mass in lowering activity. J Dairy Sci. 2013;96(5):2746-53. women with Rheumatoid Arthritis compared to non-rheumatic controls. Clinical 98. Feingold KR, Grunfeld C. The Effect of Inflammation and Nutrition ESPEN. 2021. Infection on Lipids and Lipoproteins. In: Feingold KR, Anawalt B, Boyce A, 110. Aparicio-Soto M, Sánchez-Hidalgo M, Rosillo M, Castejón ML, Chrousos G, de Herder WW, Dhatariya K, et al., editors. Endotext. South Alarcón-de-la-Lastra C. Extra virgin olive oil: a key functional food for Dartmouth (MA): MDText.com, Inc. prevention of immune-inflammatory diseases. Food & function. Copyright © 2000-2022, MDText.com, Inc.; 2000. 2016;7(11):4492-505. 99. Ben-Aicha S, Badimon L, Vilahur G. Advances in HDL: Much 111. EFSA Panel on Dietetic Products N, Allergies. Scientific More than Lipid Transporters. Int J Mol Sci. 2020;21(3). Opinion on the substantiation of health claims related to polyphenols in olive and protection of LDL particles from oxidative damage (ID 1333, 1638, 1639, 100. Barter PJ, Caulfield M, Eriksson M, Grundy SM, Kastelein JJ, 1696, 2865), maintenance of normal blood HDL cholesterol concentrations (ID Komajda M, et al. Effects of torcetrapib in patients at high risk for coronary 1639), maintenance of normal blood pressure (ID 3781), “anti-inflammatory events. The New England journal of medicine. 2007;357(21):2109-22. properties” (ID 1882), “contributes to the upper respiratory tract health” (ID 101. Schwartz GG, Olsson AG, Abt M, Ballantyne CM, Barter PJ, 3468), “can help to maintain a normal function of gastrointestinal tract” (3779), Brumm J, et al. Effects of dalcetrapib in patients with a recent acute coronary and “contributes to body defences against external agents” (ID 3467) pursuant syndrome. The New England journal of medicine. 2012;367(22):2089-99. to Article 13(1) of Regulation (EC) No 1924/2006. EFSA Journal. 2011;9(4):2033. 102. Lincoff AM, Nicholls SJ, Riesmeyer JS, Barter PJ, Brewer HB, Fox KAA, et al. Evacetrapib and Cardiovascular Outcomes in High-Risk 112. Khandouzi N, Zahedmehr A, Nasrollahzadeh J. Effects of canola Vascular Disease. The New England journal of medicine. 2017;376(20):1933- or olive oil on plasma lipids, lipoprotein-associated phospholipase A(2) and 42. inflammatory cytokines in patients referred for coronary angiography. Lipids Health Dis. 2020;19(1):183. 103. Rosenson RS, Brewer HB, Jr., Ansell BJ, Barter P, Chapman MJ, Heinecke JW, et al. Dysfunctional HDL and atherosclerotic cardiovascular 113. Morvaridi M, Jafarirad S, Seyedian SS, Alavinejad P, disease. Nat Rev Cardiol. 2016;13(1):48-60. Cheraghian B. The effects of extra virgin olive oil and canola oil on inflammatory markers and gastrointestinal symptoms in patients with ulcerative colitis. European journal of clinical nutrition. 2020;74(6):891-9. 75 74 Nutritional impact on health in patients with Rheumatoid Arthritis 114. Amiri M, Raeisi-Dehkordi H, Sarrafzadegan N, Forbes SC, 124. Bland JM, Altman DG. Statistics Notes: Some examples of Salehi-Abargouei A. The effects of Canola oil on cardiovascular risk factors: A regression towards the mean. BMJ. 1994;309(6957):780. systematic review and meta-analysis with dose-response analysis of controlled clinical trials. Nutr Metab Cardiovasc Dis. 2020;30(12):2133-45. 125. Perneger TV. What's wrong with Bonferroni adjustments. BMJ (Clinical research ed). 1998;316(7139):1236-8. 115. Malik M, Suboc TM, Tyagi S, Salzman N, Wang J, Ying R, et al. Lactobacillus plantarum 299v Supplementation Improves Vascular Endothelial 126. Landberg R, Hanhineva K, Tuohy K, Garcia-Aloy M, Biskup I, Function and Reduces Inflammatory Biomarkers in Men With Stable Coronary Llorach R, et al. Biomarkers of cereal food intake. Genes & nutrition. Artery Disease. Circulation research. 2018;123(9):1091-102. 2019;14:28. 116. Hofeld BC, Puppala VK, Tyagi S, Ahn KW, Anger A, Jia S, et al. 127. Arner P, Bernard S, Salehpour M, Possnert G, Liebl J, Steier P, Lactobacillus plantarum 299v probiotic supplementation in men with stable et al. Dynamics of human adipose lipid turnover in health and metabolic disease. coronary artery disease suppresses systemic inflammation. Scientific Reports. Nature. 2011;478(7367):110-3. 2021;11(1):3972. 117. Zamani B, Golkar HR, Farshbaf S, Emadi-Baygi M, Tajabadi- Ebrahimi M, Jafari P, et al. Clinical and metabolic response to probiotic supplementation in patients with rheumatoid arthritis: a randomized, double- blind, placebo-controlled trial. International journal of rheumatic diseases. 2016;19(9):869-79. 118. Zamani B, Farshbaf S, Golkar HR, Bahmani F, Asemi Z. Synbiotic supplementation and the effects on clinical and metabolic responses in patients with rheumatoid arthritis: a randomised, double-blind, placebo- controlled trial. The British journal of nutrition. 2017;117(8):1095-102. 119. Alipour B, Homayouni-Rad A, Vaghef-Mehrabany E, Sharif SK, Vaghef-Mehrabany L, Asghari-Jafarabadi M, et al. Effects of Lactobacillus casei supplementation on disease activity and inflammatory cytokines in rheumatoid arthritis patients: a randomized double-blind clinical trial. International journal of rheumatic diseases. 2014;17(5):519-27. 120. Kiadaliri AA, Petersson IF, Englund M. Educational inequalities in mortality associated with rheumatoid arthritis and other musculoskeletal disorders in Sweden. BMC musculoskeletal disorders. 2019;20(1):83. 121. Sweden S. Utbildningsnivå för befolkningen efter inrikes/utrikes född, kön och åldersgrupp 2020 [updated 2021-05-06. Available from: https://www.scb.se/hitta-statistik/statistik-efter-amne/utbildning-och- forskning/befolkningens-utbildning/befolkningens-utbildning/. 122. Sweden S. Befolkning 16-74 år efter kön, ålder, nationell bakgrund, utbildningsnivå och utbildningsinriktning SUN 2020. År 2019 - 2020: Statistics Sweden; 2020 [updated 2020. Females, 45-74 years of age:[Available from: https://www.statistikdatabasen.scb.se/pxweb/sv/ssd/START__UF__UF0506/UtbS UNBefN/. 123. Personcentrerat och sammanhållet vårdförlopp – Reumatoid artrit - etablerad https://d2flujgsl7escs.cloudfront.net/external/Vardforlopp_Reumatoid_artrit_eta blerad.pdf2021 [updated 2021-10-22. 77 76