Ongoing Projects

Background: Overweight and obesity are major risk factors for and contributors to increase morbidity and mortality from cardiovascular disease, cancer and other chronic diseases. Conversely, maintenance and loss of weight and fat mass are crucial to ameliorate health issues related to overweight and obesity. Several OMICS techniques have emerged as powerful tools to investigate the molecular signatures associated to changes in weight and body composition, namely gain or loss of weight and fat mass. However, there is still a need for more studies to identify molecular profiles associated to changes in weight and body composition and their relationship with different metabolic health traits. Defining OMICS-based signatures that reflect differential metabotypes associated with weight and body composition change over time could help to better understand the biological processes underlying weight and fat mass gain and loss, as well as their repercussion on cardiovascular health.

Hypothesis: The proteome and metabolome profiles are important determinants of weight and fat mass trajectories and cardiometabolic risk factors, in response, or not, to different lifestyle behaviors.

Aim: To characterize molecular phenotype signatures reflecting changes of weight and fat mass, and the association of these with cardiometabolic traits, using data from the Swedish Mammography Cohort Clinical (SMC-C) and the Cohort of Swedish Men Clinical (COSM-C).

Method: We will use available data on diet, sociodemographic, lifestyle, clinical history, blood biochemistry, and OMICS from men and women participating in the SMC-C and COSM-C. Changes in weight and body composition will be using data from the previous and/or next time-point to the OMICS analyses. OMICS data, including plasma metabolome and proteome will be used as predictor variables for changes in weight and body composition in multivariate models, as well as for their association with specific cardiometabolic health parameters. Importantly, further analysis including microbiome data will be conducted when these are available. Diagnoses of chronic diseases, as the aforementioned in the CD10 codes will be used as exclusion criteria or controlling variables in multivariate models, as appropiate.

Background: Carbohydrates are one of the main sources of energy in the diet and have important health implications depending on their quality and the quantity in which they are consumed. Recent studies have identified OMICS such as metabolomics and proteomics as powerful tools to investigate the molecular signatures called metabotypes associated to the intake of specific foods and food patterns. To date, however, there is a scarcity of studies trying to identify metabotypes associated dietary carbohydrate quality patterns and to their impact on metabolic health. Identifying such metabotypes could help to better understand the biological processes that link carbohydrate quality to cardiovascular disease and health.

Hypothesis: Prolonged intake of simple and complex carbohydrates will differentially modulate plasma metabolome and proteome producing OMIC signatures that can be further used to elucidate underlying mechanisms linking carbohydrate quality and cardiometabolic risk.

Aim: To characterize molecular phenotype signatures reflecting the intake of different carbohydrate qualities, and subsequently to associate them with cardiometabolic risk factors, using data from the Swedish Mammography Cohort Clinical (SMC-C) and the Cohort of Swedish Men Clinical (COSM-C)

Method: We will use available dietary, sociodemographic, lifestyle, clinical, biochemistry, and OMICS (metabolome and proteome) data from adult individuals belonging to SMC-C and COSM-C. A carbohydrate quality index will be calculated considering fiber intake, glycemic index, whole grains/total grains ratio, and solid carbohydrates/total carbohydrates ratio intake among participants. OMICS data will be used as predictor variables in multivariate models to assess their association with the carbohydrate quality index and specific cardiometabolic health traits. Importantly, further analysis including microbiome data will be conducted when these are available. Diagnoses of chronic diseases, as the aforement

Aims and hypothesis

Sub-project 1. The aims are to:
1a. Explore the association of novel metabolites and proteins with physical inactivity, obesity, and T2D
1b. Determine causality of these associations using Mendelian randomisation
2a. Identify causal pathways linking physical inactivity, obesity, and T2D with breast, colorectal, and endometrial cancer and investigate extent of mediation
2b. Explore the extent of interaction between physical inactivity, obesity, and T2D for the risk of breast, colorectal and endometrial cancer development, within causal models
3. Evaluate the relative contribution of novel metabolites and proteins to each causal pathway identified

I hypothesise that a number of novel metabolites and proteins will be associated with physical inactivity, obesity, and T2D. The direction of effect will vary depending on the mechanistic pathway. I anticipate that a small proportion of these molecules will have a causal effect, providing insights on the molecular aetiology of important risk factors for cancer. Extending the Mendelian randomization paradigm to investigate a broader causal network, I hypothesise that physical inactivity and obesity cause T2D and breast, colorectal and endometrial cancer and are also upstream factors for a causal pathway between T2D and cancer. I hypothesise that physical inactivity will interact with both obesity and T2D to influence risk of cancer. Finally, I hypothesise that a number of previously unanticipated metabolites and proteins will mediate the causal relationships between physical inactivity, obesity, T2D, and cancer subtypes providing vital information about mechanistic pathways.

Sub-project 2. The aims are to:
1. Determine which blood plasma metabolites, identified by untargeted liquid chromatography-quadrupole time of flight-mass spectrometry (LC-qTOF-MS) and nuclear magnetic resonance (NMR) based approaches, are associated with incidence of 10 common types of cancer
2. Assess if these biomarkers can improve non-invasive prediction models for incident cancer
3. Examine if there are causal effects of the identified blood plasma metabolites on cancer development

I hypothesise that a number of novel metabolites will be associated with cancer incidence. The strength of association will vary between men and women, and there may be some metabolites that have sex-specific associations. The direction of effect will vary depending on the mechanistic pathway, and some metabolites will have non-linear associations with clear thresholds at which risk of cancer incidence markedly increases or decreases. I anticipate that hazard ratios for some metabolites will vary by time since biological sampling and cancer stage at diagnosis. Metabolites associated with a short-term increased risk of cancer incidence will represent potential biomarkers for early detection of disease. I hypothesise that the novel biomarkers we identify will improve the sensitivity and specificity of models to predict the occurrence of cancer, compared to previously developed non-invasive prognostic models. I anticipate that a small proportion of molecules will have a causal effect, providing insights into the molecular aetiology of cancer incidence.

Sub-project 3. The aims are to:
1. Examine if there is evidence of a causal association between physical activity and cancerspecific survival, using cutting edge Mendelian randomisation approaches accounting for diabetes status and other pleiotropic traits
2. Explore if (change in) physical activity is associated with novel biomarkers (proteins, metabolites, and gut microbiota) in individuals with and without a previous cancer diagnosis
3. Determine if associations between (change in) physical activity and novel biomarkers in individuals with and without a previous cancer diagnosis differ by diabetes status

Background: Overall, total joint replacement (TJR) of the hip or knee joint has proven to be a successful treatment of primary and secondary osteoarthritis (OA) as well as other degenerative joint diseases and fractures. Yet, a considerable portion of all patients treated with TJR are subject to revision surgery for aseptic loosening, infection or dislocation amongst other reasons. The risk for revision surgery of the hip or knee joint subsequent to primary arthroplasty seems to be influenced by the individuals’ genetic profile to some extent. Several candidate-driven studies have identified certain single nucleotide polymorphisms (SNP) associated with the risk for revision surgery. Most notably, both MacInnes et al. and our own research group have performed genome-wide association studies (GWAS) and found associations between SNPs and the risk for revision surgery for aseptic loosening following TJR. We now seek to replicate and thus verify the results obtained in our preliminary study.

Aims and hypothesis: The aim of the proposed study is to identify SNPs associated with a higher risk for revision arthroplasty subsequent to primary TJR. In particular, we aim to verify and validate the results reported in our previous study. Hence, our hypothesis is that certain SNPs increase the risk for revision arthroplasty.

Methodology: We plan to perform a GWAS on a subgroup of patients within the SIMPLER-cohort. All patients treated with a TJR of the hip or knee joint as indicated by the surgical procedure codes NFBXX and NGBXX will be included in the study. Data on diagnosis, age and BMI at primary TJR, along with the specific procedural code will be collected. For all patients undergoing revision arthroplasty, information on reason for revision surgery, age, BMI and procedural codes at revision surgery will be gathered. Furthermore, for all deceased patients, date of death will be checked for to enable time-to-event analysis. Lastly, the cohort will be cross-matched against the Swedish Arthroplasty Register to ensure that no secondary procedures went unnoticed.

After data collection, we aim to perform a GWAS with the primary endpoint revision surgery for aseptic loosening in the above defined study population. Our secondary endpoint is revision surgery of the affected hip or knee joint for any cause.

Background: We examined whether the inverse association between adherence to a Mediterranean diet and hip fracture risk is mediated by incident type 2 diabetes mellitus (T2DM) and body mass index (BMI).

Methods: We included 50 755 men and women from the Cohort of Swedish Men and the Swedish Mammography Cohort who answered lifestyle and medical questionnaires in 1997 and 2008 (used for calculation of the Mediterranean diet score 9mMED; low, medium, high) and BMI in 1997, and incident T2DM in 1997-2008). The cumulative incidence of hip fracture from the National Patient Register (2009-14) was considered as outcome.

Results: We present conditional odds ratios (OR) 9[95% confidence interval, CI) of hip fracture for medium and high adherence to mMED, compared with low adherence. The total effect ORs were 0.82 (0.71, 0.95) and 0.75 (0.62, 0.91), respectively. The controlled direct effect of mMED on hip fracture (not mediated by T2DM, considering BMI as an exposure-induced confounder), calculated using inverse probability weighting of marginal structural models, rendered ORs of 0.82 (0.72, 0.95) and 0.73 (0.60, 0.88), respectively. The natural direct effect ORs (not mediated by BMI or T2DM, calculated using flexible mediation analysis) were 0.82 (0.71, 0.95) and 0.74(0.61, 0.89), respectively. The path-specific indirect and partial indirect natural effects ORs (through BMI or T2DM) were close to 1.

Conclusions: Mediterranean diet has a direct effect on hip fracture risk via pathways other than through T2DM and BMI. We cannot exclude mediating effects of T2DM or BMI, or that their effects cancel each other out.

The role of milk and fermented milk consumption in stroke risk is unclear. We investigated associations of time-updated information on milk and fermented milk consumption (1997 and 2009) with total stroke, cerebral infarction, and hemorrhagic stroke risk among 79,618 Swedish women and men (mean age 61.3 years). During a mean follow-up of 17.7 years, we identified 9735 incident cases of total stroke, of which 7573 were cerebral infarctions, 1470 hemorrhagic strokes, and 692 unspecified strokes. Compared with an intake of 100 g/day of milk, the multivariable-adjusted hazard ratios (95% confidence interval) of cerebral infarction were 1.05 (1.02-1.08) for 0 g/day, 0.97 (0.95-0.99) for 200 g/day, 0.96 (0.92-1.00) for 400 g/day, 0.98 (0.94-1.03) for 600 g/day, and 1.01 (0.94-1.07) for 800 g/day. Corresponding estimates for hemorrhagic stroke were 0.98 (0.91-1.05) for 0 g/day, 1.02 (0.97-1.07) for 200 g/day, 1.07 (0.98-1.17) for 400 g/day, 1.13 (1.02-1.25) for 600 g/day, and 1.19 (1.03-1.36) for 800 g/day. No associations were observed between milk consumption and total stroke or for fermented milk consumption and any of the stroke outcomes. Higher long-term milk consumption based on repeated measures of intake was weakly and non-linearly associated with cerebral infarction, and was directly associated with hemorrhagic stroke.

Background: Milk and dairy products are recommended as a part of a healthy diet. In contrast to the traditional view that milk consumption protects against fractures, we recently showed that a high intake of milk instead increased the risk of hip fracture, especially among women. The risk of mortality was also increased. In contrast, a high intake of fermented milk products including sour milk and yogurt or cheese were associated with lower risks of fracture and mortality. A potential explanation for these discrepant results is different effects on inflammation and oxidative stress, suggested underlying mechanisms for age-related diseases including cardiovascular disease, cancer, fracture, and diabetes. We could also demonstrate that a higher intake of milk was also associated with higher levels of oxidative stress and inflammation markers whereas a higher intake of sour milk and yogurt were associated with lower levels.

Hypothesis: That milk and fermented milk products display different associations with protein biomarkers selected to be involved in cardiovascular disease and meatabolism.

Aim: We aim to study the associations of milk, fermented milk and cheese intake with a large number of proteins in the OLINK CVDII, CVDIII, and metabolism panels. We further aim to compare if the different dairy products are associated with the same or different proteins and to assess and compare the strength and direction of the associations.

Method: Information on milk, fermented milk, and cheese intake from the questionnaires administered at the clinical examinations of SMCC and COSM-C will be primarily used. To study long-term intakes, information will also be collected from the 1987 (for women only), 1997, and 2008/2009 questionnaires. We will use linear regression analyses with the dairy products as independent variables and the protein levels as dependent variables, adjusting for potential confounders. We will apply methods that take multiple testing into account. The analyses will be performed among both men and women.

The aim is to study milk and dairy product intake and cardiovascular disease.

Background: With aging, changes in body composition is characterized by an increase of fat mass and a decrease of lean mass. Loss of lean mass can result in deteriorated muscular function that may cause a limitation of mobility, physical performance and increased dependence of assistance from others. The term sarcopenia, often defined by using a combination of muscle mass, muscle strength and physical performance, has been proposed to define the age-related loss of these. Result from our own research suggest that a healthy dietary pattern may be associated with a lower prevalence of sarcopenia in old men. As gender-differences for the impact of diet on both muscle strength and physical performance might be present, it is of great importance to study both men and women.

Hypothesis: Healthy diet pattern is associated with lower prevalence of sarcopenia in old women.

Aim: To assess the associations between diet, nutrients and dietary patterns and both sarcopenia, the comprised parameters and the change in muscle mass in old women.

Method: All women who participated in the first clinical examination with dietary information and muscle mass measurement and those who have returned to the second examination will be included. We will analyze the associations between adherence to dietary patterns, both a posteriori (i.e. data driven patterns) and a priori (i.e. Mediterranean diet), at baseline and sarcopenia. Where it is not possible to define sarcopenia, the comprised parameters in the definition will be analyzed separately. We further plan to estimate the associations between adherence to dietary patterns and the change in muscle mass between baseline and the repeat examination

Low-grade inflammation, i.e. a small chronic increases in inflammation markers in the circulation, is involved in the pathogenesis of cardiovascular disease (CVD). How nutritional composition affects low-grade inflammation is not well understood. The overall aim of this project is to study the effects of macronutrient intake on markers of low-grade inflammation.

Specific aims: 1) To assess the association between markers of inflammation and intake of fat, defined as total fat, groups of fatty acids and individual fatty acids. 2) To assess the association between markers of inflammation and intake of protein and carbohydrate (total carbohydrate, sucrose, whole grain and fiber).

Data from the Swedish Infrastructure for Medical Population-Based Life-Course and Environmental Research (SIMPLER) will be utilized for this project. This infrastructure includes data from ca 100 000 individuals, from the Swedish Mammogram Cohort and the Cohort of Swedish men. In these cohorts, data on inflammation markers and dietary intake are available for ca 13 000 participants. The exposure of interest is firstly intake of fat, including total fat, SFA (total, medium and long chain), MUFA, PUFA (total, omega 3 and omega 6) and individual fatty acids. Additional exposures will be protein (including total protein and protein from animal and plant-based sources) and carbohydrate (including total carbohydrate, sucrose, whole grain and fiber). Outcome variables are inflammation markers assessed by a proteomic panel analyzed by Olink, Uppsala.

Methods: We will perform genome-wide association studies of microbiota characteristics (GWAS) in the SCAPIS-Uppsala and SCAPIS-Malmö cohorts with available data (n=9,847 with metagenomics), and SIMPLER (n=7,375 with metagenomics).

To be able to meta-analyze the two cohorts, we want to send the raw sequencing data to Clinical Microbiomics in Copenhagen for them to rerun the same processing and mapping of species and functions as in SCAPIS.

We have extensive experience of running and publishing GWAS. Both cohorts are already genotyped with SNP arrays and imputed genotype data that passes initial quality control will be used. We will further meta-analyze results (sharing no individual level data) with results from other cohorts. We will use state-of-the-art methods, including mixed models to properly account for population stratification.

Expected outcome and future directions: The anticipated outcome of this project is that we will discover genetic variation linked to microbiome variation. This will improve our biological understanding and might provide tools for Mendelian Randomization.

The overall purpose of this project is to investigate the effect of specific host factor on the composition of the gut microbiota with potential implication for cardiometabolic health. We focus on two factors: antibiotic use and physical activity. To achieve this aim, we plan to combine data from large population-based studies that include fecal metagenomics data and physical activity data with information from the Swedish national register of prescribed drugs.

Specific aims:

1. Investigate how previous antibiotic exposure is associated with the gut microbiota alpha and beta diversity
2. Investigate how the number of previous antibiotic courses is associated with the abundance of gut microbiota species
3. Investigate how the number of previous courses of each antibiotic class is associated with the abundance of gut microbiota species
4. Investigate how leisure time physical activity and sedentary behavior are associated with the abundance of gut microbiota species
5. Investigate a causal association of physical activity and sedentary behavior with gut microbiota species using a Mendelian randomization

Aim and hypothesis: The overarching aim of the project is to understand the role of gut microbiota composition and functionality on the risk of depression, anxiety, stress-related disorders, and cognitive impairment, as well as on the treatment response to antidepressants. Specifically, we aim to assess:

1. the association of gut microbiota composition and functionality with the risk of depression, anxiety, stress-related disorders, and cognitive impairment;
2. the effect modification of genetic predisposition to psychiatric disorders and cognitive impairment (i.e., risk genes or polygenic risk score) in the associations of gut microbiota with common psychiatric disorders and cognitive impairment;
3. the role of gut microbiota composition and functionality on modulating the link between infections (including COVID-19) and depression/anxiety/stress-related disorders and cognitive impairment;
4. the associations of gut microbiota composition or functionality with the treatment response to antidepressants.

Methods: We will address the four aims using SIMPLER (Swedish Infrastructure for Medical Population-Based Life-Course and Environmental Research). We will define a cohort including ≈7300 individuals with gut microbiome data from SIMPLER. Fecal samples were collected during 2010-2013 in SIMPLER. These samples were sent for deep metagenomics sequencing according to standard processes. We will assess gut microbiota composition and functionality using the metagenomics data. Gut microbiota functionality will be defined by relevant metabolic pathways and proteins which will be identified by microbial proteins, coding genes and other genomic features (such as tRNAs, noncoding RNAs, etc.) using assembly-based profiling approach. Questionnaires on sociodemographic, lifestyle, dietary, health-related factors, as well as depression, anxiety, and stress-related disorders have been conducted at baseline and repeated several times. Cognitive function of the participants was assessed by a validated cognitive phone interview when fecal samples were collected.

SIMPLER have been linked, through the unique personal identity number, to the Swedish Patient Register, Cause of Death Register, Prescribed Drug Register, Longitudinal integrated database for health insurance and labour market studies, and Total Population Register. With these linkages, we will retrieve information on sociodemographic factors, comorbidities (including cancer, cardiovascular disease, diabetes, infections including COVID-19, etc.), as well as use of medications for all cohort participants. We will identify clinical diagnoses of depression, anxiety, and stress-related disorders according to inpatient and outpatient care records from the Patient Register using the International Classification of Diseases codes. We will identify information on the use of antidepressants and other medications known to impact gut microbiome such as antimicrobial agents (e.g., antibiotics, antivirals, antifungals, antiparasitics, etc.), antineoplastic agents, and proton pump inhibitors from the Prescribed Drug Register using anatomical therapeutic chemical codes.

Project 1: To investigate the association of gut microbiota composition and functionality with the risk of depression, anxiety, stress-related disorders, and cognitive impairment

We will conduct a case-control study among all participants with gut microbiome data in SIMPLER, and use as cases all participants with a diagnosis of depression, anxiety, or stress-related disorders during follow-up of SIMPLER whereas as controls all other participants. Similarly, we will conduct another case-control study among participants with gut microbiome data in SIMPLER, and use as cases all participants with cognitive impairment whereas as controls all participants without such.

Project 2: To test whether the genetic predisposition to psychiatric disorders and cognitive impairment modifies the associations of gut microbiome with the risks of depression, anxiety, stress-related disorders, and cognitive impairment

We will analyze effect modification by genetic predisposition among individuals with data on genotypes in the three case-control studies described in Project 1.

Project 3: To test whether gut microbiome modulate the associations of infection with the risks of depression, anxiety, stress-related disorders, and cognitive impairment

We will analyze the interaction of gut microbiome with infection among individuals from SIMPLER in the case-control studies described in Project 1.

Project 4: To study if gut microbiota composition and functionality is associated with the treatment response to antidepressants

We will conduct a cohort study including participants that used antidepressants and with available data on gut microbiome in SIMPLER (n≈1200). Information on åkessonantidepressant use at baseline and during follow-up will be retrieved from the Prescribed Drug Register. We will assess the treatment response to antidepressants by 1) changes of depressive symptoms acquired from questionnaires in 2008 (baseline) and 2019 (during follow-up) in SIMPLER and 2) changes of use pattern of antidepressants (i.e., switching of drugs or changes in dosage) in both SIMPLER.

Background: Non-alcoholic fatty liver disease (NAFLD) is characterized by excess amount of fat deposited in the liver and is associated with increased risk of cardiometabolic diseases. Substituting saturated fat (SFA) with polyunsaturated fat (PUFA) has been shown to reduce liver fat content in randomized controlled trials, findings that have been corroborated by both crosssectional studies and cohort studies. However, few studies have been conducted to investigate associations between isocaloric substitutions of foods differing in fat- and carbohydrate quality on more advanced stages of NAFLD, i.e. cirrhosis and liver cancer.

Hypothesis: We hypothesize that substituting SFA-rich foods with PUFA-rich foods as well as refined carbohydrate-rich foods with whole-grain/fiber-rich foods is associated with a lower risk of incident NAFLD cirrhosis and liver cancer.

Aim: To investigate associations between isocaloric substitutions of foods differing in fat and carbohydrate quality and NAFLD cirrhosis and liver cancer in middle aged men and women.

Method: Statistical substitution models (leave-one-out method) will be used to examine associations between food (and nutrient) replacements and NAFLD cirrhosis and liver cancer, with foods (and nutrients) expressed in grams. Multivariable Cox regression models will be employed to estimate hazard ratios and 95% CI. Time-updated data on both the exposure and confounders from the 2008/2009 investigation will be used. Potential confounders will be identified using directed acyclic graphs.

Background: This is a cohort study of metabolic factors and inflammation in relation to chronic diseases of major public health importance, including cardiovascular disease, cancer, dementia, renal, auto-immune and psychiatric disorders. The project is based on the AMORIS (Apolipoprotein MORtality RISk) cohort. The project started in 2010 and has generated more than 40 international scientific publications. The current project represents a continuation with an extended follow-up time and a more in-depth analysis of certain research questions and disease areas as well as a greater focus on older age groups and women’s health.

Hypothesis: Metabolic disturbances and inflammation are related to long term risk of chronic diseases of great public health importance.

Aim: The aim of this project is to evaluate long term effects of metabolic and inflammatory disturbances for the risk of developing chronic diseases of major public health importance.

Method: Altogether the AMORIS-cohort includes more than 800.000 subjects taking part in health examinations during 1985-96 (baseline) where blood and urine samples were analysed at one and the same clinical laboratory (the CALAB laboratory). These baseline examinations yielded extensive information on a great number of biomarkers. The cohort is followed regarding morbidity, mortality and emigration using Swedish national registers on hospitalizations and specialized outpatient care, cancer incidence, causes of death, prescribed drugs, national quality registers and population registers. Information from the SMC and COSM surveys, together with data from other sources, are used to characterize the cohort regarding primarily life style factors, self-reported health and factors associated with women’s health.

Specifika frågeställningar:

1. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla hjärt- kärlsjukdom respektive prognos vid hjärt- kärlsjukdom, främst akut hjärtinfarkt (inklusive typ av hjärtinfarkt), hjärtoperationer, stroke, förmaksflimmer och hjärtsvikt liksom tidiga manifestationer av hjärt- kärlsjukdom?
2. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på sjukfrånvaro och hälso- och sjukvårdskostnader?
3. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla olika cancersjukdomar respektive prognosen vid olika cancersjukdomar?
4. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla njur- respektive leversjukdom, bl a NAFLD, samt prognos vid dessa sjukdomar?
5. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla neurologiska sjukdomar inklusive ALS, Parkinson’s sjukdom, ME/CFS, PEM, (Post Excertional Malaise) respektive fibromyalgi?
6. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla autoimmuna sjukdomar som bl a vissa hormonsjukdomar (hypothyreos) samt reumatoid artrit?
7. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla psykisk ohälsa?
8. I vilken utsträckning inverkar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes respektive inflammation på risken att utveckla demenssjukdom (bl a Alzheimer)?
9. Vilken roll spelar metabola rubbningar i form av bl.a. dyslipidemi, förhöjt fasteglukos, diabetes, inflammation och biomarkörer för åldrande?

För ovanstående frågeställningar kommer även frågor om interaktionseffekter mellan olika indikatorer på metabola rubbningar samt modifiering av samband av demografiska faktorer (ålder, kön, utbildning, civilstånd, yrke, socioekonomi respektive födelseland) att belysas.

Research objective: I aim to increase understanding of the biological mechanisms behind cognitive and physical impairment in dementia-free general and clinical populations (i.e., cancer survivors/diabetes patients) by focusing on novel molecular pathways and the role of physical activity.

Specific Aims:

Part 1. Novel molecular pathways behind cognitive and physical impairment

• Examine the association of metabolites and proteins in blood plasma and gut microbiota with measures of cognitive and physical impairment.
• Explore effect modification of these associations by age, sex, education, and body composition.
• Determine causality of the relationship of important metabolites, proteins, and gut microbiota with cognitive and physical impairment.

Part 2. The role of physical activity in cognitive and physical impairment

• Examine the association of physical activity with measures of cognitive and physical impairment.
• Explore effect modification of these associations by age, sex, education, and body composition.
• Determine causality of the relationship of physical activity with measures of cognitive and physical impairment.

Part 3. Can novel molecular pathways explain the role of physical activity in cognitive and physical impairment?

• Investigate if putative causal metabolites, proteins, and gut microbiota identified in part 1 mediate the association of physical activity with measures of cognitive and physical impairment.

General background: Our project builds upon the recent advancements in methodology for causal inference to conduct comparative effectiveness research in the field of nutritional epidemiology. We will use the target trail conceptual framework and causal analytical methods, the parametric g-formula (1,2) to evaluate the effectiveness of hypothetical dietary interventions on the risk of cardiovascular diseases (CVD) by using data from the Swedish Mammography Cohort and the Cohort of Swedish Men.

Nutritional epidemiology uses observational data extensively, as randomized controlled experiments are expensive and often difficult to conduct. However, previous research has been struggling with methodological shortcomings, often as the consequences of ill-defined interventions, i.e., limiting the research question of examining the associations between different levels of a nutrient or different dietary patterns and the risk of CVD, but not explicitly formulating the causal question. In nutritional epidemiology, the effect of interventions usually involves the replacement of one exposure by another and estimating the effect of sustained dietary interventions requires careful specification of the replacement diet. For example, evaluating a target intervention of "eat 2 servings of fish per week." is too vaguely stated for practical implementation. It also requires to specify the intake of which foods (e.g., red meat, poultry) will be reduced in order to maintain a constant intake of proteins and energy. Otherwise, the interventions, which have been evaluated in previous research, i.e., "Keep diet score in a range corresponding to the top 2 quintiles of the observed data" or "eat 2 servings of fish per week" are essentially meaningless. However, conventional methods and measurement of dietary information one point at a time, assuming that the given dietary pattern and the measured confounders are sustained in time and do not interact with each other do not allow to ask the relevant research questions. Diet and other lifestyle factors may interact with each other, i.e., changes in diet may influence other risk factors, i.e., smoking, physical activity, which may also have consequences on diet itself. Therefore, evaluating the effectiveness of sustained interventions requires both repeated longitudinal measures and causal analytical methods which can handle the feedback loop between the examined dietary factors and other risk factors. Only a few previous studies in US data from our group has attempted to implement the g-formula to compare the effectiveness of sustained lifestyle and dietary interventions and the risk of CVD (2-4). There have been no studies examined these research questions in a Swedish context.

Therefore, in this project, we will implement new analytical methods on the Swedish Mammography Cohort and the Cohort of Swedish Men to compare the effectiveness of different, sustained dietary interventions for the prevention of cardiovascular diseases.

Aims: We quantify the effect of dietary intervention on the intake of specific foods, such as

• fish
• red meat/pork
• carbonated beverages (soda)
• coffee
• on the risk of CVD.

For each intervention we will consider (i) changes in the dietary exposure of interest (e.g., fish) and (ii) changes in other (replacement) dietary exposures (e.g., read meat, chicken, cheese, etc) to maintain a constant caloric intake.

Methodology: We will use the parametric g-formula to quantify the effect of the different hypothetical dietary interventions on the risk of CVD. So far, there has only been limited use of the g-formula in causal inference research due to the consequence of the impossibility to apply its non-parametric version to high dimensional data, and from the widespread misbelief of its sensitivity of model misclassification. However, our previous research has shown (2-5), that compared to other causal inference methods, the parametric g-formula has several advantages: (i) The estimates that have been derived from the model tested to be quite robust to model specification, even more stable than estimates from alternative methods. (ii) It can be applied in complex settings, for example, the comparison of joint interventions on two or more exposures. Therefore, it will be the ideal choice to calculate the effect of hypothetical dietary interventions when one nutrient is substituted with another. When estimating the g-formula under a joint intervention, we will need to adjust for time-varying confounders as different lifestyle factors, change in lipid status or blood pressure, which possibly affected by either of the exposures. We will use longitudinal measures on dietary factors, health and lifestyle factors from the SMC and COSM cohort and information from the inpatient register to identify incident CVD cases. Furthermore, we also will estimate risk ratios, differences, attributable risks, and survival curves in both cohorts.

Circulating protein biomarkers are widely used in medical research and clinical practice, but the understanding of their regulation, inter-individual differences due to DNA variation and potential causal roles in disease is incomplete. In this project, we will map protein quantitative trait loci (pQTL) for circulating proteins, measured using the Olink proximity extension assay. Data will be meta-analyzed with other studies having both genetic and proteomic data within the SCALLOP consortium to maximize statistical power. For the first project, the SMC-C data were used for replication of a discovery meta-analysis encompassing 21,758 participants. In this discovery effort, a total of 545 pQTLs at p<5x10-8 for 85 proteins were identified. By integrating pQTLs with curated protein-protein interaction networks and text-mining, demonstrated examples of molecular processes involved with the regulation of circulating proteins. Using Mendelian randomization, several proteins were found to play a causal role in disease, including IL6RA and PGF with coronary heart disease; IL1RA and CD40 in rheumatoid arthritis; and MMP-12 and CD40 in stroke. This manuscript is currently under review with Nature Medicine.

Moving forward, the SMC-C and COSM-C data will be included in the discovery meta-analyses of the CVD-II, CVD-III and metabolism GWAS analyses. The implications of this line of research include biological understanding of cardiometabolic disease, and importantly, causal inferences of potential novel drug targets or indeed, adverse effects of launched therapies.

Background: Dental diseases (caries and periodontal disease) develop from complex genetic, biological, behavioural and environmental interactions. Biological variation (including genetic variation) leads to substantial individuality in susceptibility and development of these diseases, which calls for personalized approaches for their prevention and treatment.

Hypothesis: The hypothesis is that data-driven refined clinical measures of caries and periodontitis perform better in genome-wide association studies than previously used dichotomized and often self-reported classifications which have been disappointing in identifying gene loci.

Aim: The aim is to use nouvelle, clinically assessed data-driven measures of caries and periodontitis in hypothesis-free genome-wide association studies in the GLIDE consortium. This will help identify a genetic basis for caries and periodontal disease to search genetic and lifestyle factor interactions and identify which interventions might be effective for reducing risk for people with high genetic/molecular susceptibility.

Method: Participant ids will be linked to the national quality register SKaPa. SKaPa compiles and ensures quality control over data from electronic records from all public dental health and some private clinics in Sweden. Dental data (caries and periodontitis signs) are retrieved at the tooth surface level and aggregated as identified in our recent large-scale twin analyses on heritability of dental disease and oral microbiota (PMID: 31905308 , 32726935, and one under review). GWAS summary statistics are modelled as described in our previous publication (PMID: 31235808) and metadata are meta-analysed for those GLIDE centres where clinically assessed data are available.

abstract

abstract

Type 2 diabetes (T2D) is a major cause of morbidity, mortality, and high healthcare costs worldwide. Diet has been identified as an important modifiable risk factor of T2D. However, the biological mechanisms underlying pathophysiology of T2D and how diet might affect molecular processes that may be involved in T2D remain largely unclear.

In SMCC cohort, we aim to (1) identify protein biomarkers that could predict incident T2D; (2) investigate whether predictive proteins may provide new insights to the pathways related to disease development and may be used for improving risk prediction beyond established risk factors; (3) investigate the impact of diet in modifying proteins associated with T2D risk.

To achieve these aims, multivariate-adjusted cox proportional hazards regression will be used to estimate hazard ratios of T2D risk for proteins. We will apply machine learning based variable selection techniques to identify most relevant protein predictors and evaluate their incremental value in risk prediction compared to established risk factors. Dietary patterns will be derived using a data-driven approach and thereafter, associations of proteins with food items and/or dietary patterns will be assessed using Partial Spearman correlation analysis, accounted for covariates.

The present study constitutes the hitherto largest study where a high number of proteomic biomarkers targeted for cardio-metabolic risk factors are assessed from a prospective cohort to find protein biomarkers of T2D and their relation to diet. This will provide excellent possibilities to discover novel biomarkers of T2D, as well as provide potential targets for nutritional prevention and therapy for T2D, with large potential implications for public health.

abstract

Background: Cardiometabolic health, reflecting cardiovascular, renal, metabolic, prothrombotic and inflammatory status is an emerging global issue . Cardiometabolic abnormalities increase the risk of CMDs, including cardiovascular diseases (CVD) and type-2 diabetes (T2D), which are the global leading causes of death and under the targets of the UN Sustainable Developmental Goal (SDG) 3 . The risk of developing CMDs has been associated with genetics, as well as exposure to environmental factors, such as diet, gut microbiota and pollutants. With a focus on identifying actionable targets for CMD prevention or intervention, it is clear that investigations into exposures are highly warranted. T2D, MI and stroke were selected as our focus due to the high prevalence and their associations to modifiable exposures, thereby with potential for reducing incidence, morbidity and mortality, once mechanisms and molecular targets are known. Although many prospective cohorts and casecontrol epidemiological studies have investigated exposure-outcome associations, these studies have typically investigated exposures one-by-one but not holistically. Moreover, there is still much unknown about molecular mechanisms linking exposures to outcomes via metabolic regulation. Metabolomics, as a reflection of physiology and biological states at molecular phenotype, is influenced by the cumulative effects of endogenous and exogenous exposures over the life courses. Identifying metabolites that are associated retrospectively (or cross-sectionally) with exposures and prospectively with outcomes may provide insights into linking pathways from exposure to outcome. Metabolomics studies have typically assessed the plasma metabolome due to its availability from convenience in sampling. Thus, identifying the plasma metabolite profiles of individuals under certain environmental exposures may assist us to understand, causally, how combined environmental exposures (e.g. diet and pollutants) regulate metabolic pathways and how that regulation, in turn, relates to CMDs and intermediate risk markers. In this study, we aim for a more holistic understanding of exposure-related metabolic regulations by exploring independent and combined effects of diet and POPs on CMDs through metabolomics. To enhance the robustness of the findings, exposure-metabolite-outcome associations will be validated across three Nordic populations from SIMPLER, the Northern Sweden Health and Disease Study - Västerbotten Intervention Programme - BioDiVa subcohort and the Danish Cancer and Health - Next Generations - MAX validation subcohort.

Hypothesis: Our hypothesis is that data-driven analysis of metabolomics data, through a combination of machine learning with a socalled “meet-in-the-middle” approach (Chadeau-Hyam et al. 2011) and mediation analysis (VanderWeele 2016), will allow identifying biological features and mechanisms relating environmental exposures to CMD outcomes.

Aim: The overarching aim is to explore metabolic regulations represented by untargeted metabolomics that link environmental exposures in SIMPLER (diet, POPs and microbiota) to CMD outcomes (i.e. T2D, MI, stroke and intermediate risk markers).

Our specific objectives are:

1. To discover metabolite features that associate with environmental exposures.
2. To further investigate whether such exposure-related features also associate with the CMD outcomes.
3. To replicate exposure and outcome-related metabolite features in two other Nordic cohorts (the Northern Sweden Health and Disease Study - Västerbotten Intervention Programme - BioDiVa subcohort and the Danish Cancer and Health - Next Generations - MAX validation subcohort).
4. To replicate exposure and outcome-related metabolite features from the Northern Sweden Health and Disease Study - Västerbotten Intervention Programme - BioDiVa subcohort and the Danish Cancer and Health - Next Generations - MAX validation subcohort in SIMPLER.
5. To further explore potential mechanisms from the metabolite features selected to link environmental exposures to CMD outcomes

The outcome of the present project will fill an important knowledge gap in understanding the underlying mechanism linking environmental exposures to CMD outcomes, through the metabolite features identified. Considering these widespread exposures and the large burden of CMDs, even a modest influence of these exposures on CMD risk can have a substantial effect on population health. Hence, the potential role of these exposures as preventable risk determinants markers and the underlying mechanisms linking these exposures to CMD risks could prove of high importance and relevance for society at large. On the other hand, results potentially indicating low likelihood of effects of exposures on CMD would be similarly very important and of high relevance to the authorities and the public.

Background: Recent advances in mass spectrometry have enabled high-resolution analysis of thousands of molecules in a single sample. However, the high resolution and sensitivity that make these techniques attractive also entail drift-related disturbances that are detrimental for data quality. Traditional analytical chemistry procedures using internal standards and manual peak curation are largely unfeasible and data quality issues are therefore frequently not detected until after batches or experiments, comprising hundreds to thousands of samples, are completed. This can lead to severe consequences, such as wasted sample material and increased instrument and operator burden.

Aims & hypothesis: To this purpose we are developing “QualiMon – LaMa”, a novel open-source software in R/Shiny, for real-time quality monitoring in LC-MS metabolomics. By evaluating quality at the time of injection completion, the instrument operator can intervene as soon as the instrument is showing signs of signal deterioration, LC leakage and other problems. Utilizing threshold values for well-known quality parameters, the operator may also get a notification sent to their phone/computer via the free-to-use webservice ‘Slack’ when a sample is showing signs of especially serious deterioration. After completion of an injection sequence where the quality is deemed good, the quality data can also be used to identify samples which are problematic due to failure in sample collection or preparation. We hypothesize that QualiMon will drastically help improve sample quality in many ways: alerting operator of declining quality in real time allows for reinjections while sample quality is still maintained (as compared to doing it up to a year after initial injection), help avoid human error in quality evaluation, evaluate more quality metrics than operators would have time to do and alert scientists working with the data on samples which have low quality due to reasons unrelated to instrument proficiency. We also expect this software to save a substantial amount of quality-checking time for the instrument operator, which can be spent on other work and to lower wastage of precious sample material as a consequence of fewer reinjections.

Methodology: QualiMon assesses data quality on an injection-by-injection basis by extracting general sample information (e.g. total ion current and total peak number) and specific peak characteristics (e.g. intensity, retention time, tailing factor and signal-to-noise ratio) from LandMark features (LaMas). These are easily distinguishable matrix-specific features (100-200 per analytical mode) that are present at high intensity in virtually all samples, well-distributed in retention time and mass-to-charge ratio but distant from other features. By monitoring a folder structure, injections in multiple experiments and multiple analytical modes can be tracked simultaneously. Once an injection is completed, QualiMon automatically copies the original file in vendor format to a back-up site and converts it to mzML. This is followed by XCMS peak picking to extract sample- and LaMa-characteristics into an SQLite database. Characteristics are then compared to previous injections to give a score reflecting estimated sample quality. Transgressing samples activate real-time warnings through web services, so that operators can immediately take appropriate measures. QualiMon comes with several tools for visualization available via the Shiny interface, including all quality parameters, peak shape of each landmark in each sample and a general overview of the scoring system. QualiMon will be freely available under MIT licence at https://github.com/MetaboComp/QualiMon.

To ascertain that QualiMon is able to assess data quality ,three already finished datasets of real human plasma samples will be analyzed using the software and the output evaluated. Since QualiMon is developed for large sample experiments, the SMC-C dataset is well suited for this task. The two other datasets which will be evaluated are MedGICarb and a set of in-house quality control samples, both acquired on the CMSI platform with Chalmers as data owners.

Data applied for from SIMPLER

We will use pseudonimized LC-MS instrument data (both study samples and pooled QC samples) from the SMCC metabolomics data generation. Since the purpose for this application is to assess data quality, no associations to personal traits are required. Consequently, no personal information is applied for.

Background: Peripheral artery disease (PAD) is caused by atherosclerosis and is associated with increased risk of coronary heart disease, stroke, leg amputation, and death. Smoking is the main modifiable risk factor for PAD. Diet and other lifestyle factors can influence atherosclerosis progression and therefore may play a role in the development of PAD. However, studies of diet and lifestyle factors in relation to risk of PAD are limited.

Venous thromboembolism (VTE) refers to a blood clot that starts in a vein. It is the third most common cardiovascular disease after myocardial infarction and stroke. There are two types of VTE: deep vein thrombosis (about 70% of all VTEs), which is a clot in the deep vein, usually in the leg; and pulmonary embolism (about 30% of all VTEs), which occurs when a deep vein thrombosis clot breaks free from a vein wall and travels to the lungs where it blocks some or all the blood supply. VTE is a potentially serious disease with the risk of fatal pulmonary embolism. Yet, VTE is often overlooked as a major public health problem. Treatment with anticoagulant drugs is effective but confers an increased risk of complications in the form of bleeding and hemorrhagic stroke.

VTE is a multifactorial disease caused by a combination of environmental and genetic risk factors. Among environmental risk factors, some are provoking (e.g., cancer, surgery, trauma or fracture, immobilization, long-distance travel, and hospitalization) and others are nonprovoking (e.g., age, sex, ethnicity, oral contraceptives, hormone therapy use, statin use, and air pollution). Diet, lifestyle factors (e.g., physical activity, smoking, and alcohol and coffee consumption), obesity, hypertension, and dyslipidemia may also play a role in the etiology of VTE. However, findings from observational studies on potential modifiable risk factors in relation to VTE are inconclusive. Through this project, we aim to extent the knowledge regarding the role of modifiable risk factors for VTE and other cardiovascular diseases by using data from the SIMPER infrastructure. We would like to conduct stratified analyses based on provoked (hospitalized individuals) and nonprovoked VTE. We except that diet and lifestyle are primarily related to nonprovoked VTE.

Aim and hypothesis: The aim is to assess the associations of modifiable risk factors, such as diet (overall diet, specific foods and nutrients) and lifestyle factors, with the incidence of PAD and VTE in SMC and COSM. We hypothesize that individuals with unhealthy diet and lifestyle habits exhibit increased risk for PAD and VTE compared with those with health lifestyle behaviors.

Methodology and data sources: The hypotheses will be addressed using data from the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM) obtained via the 1997 and 2008-2009 questionnaires. We would like to receive follow-up data for PAD and VTE until the latest linkage to the Swedish Patient Register. We would also like to obtain data for date of death for those who died during follow-up. Cox proportional hazards regression models will be used to estimate hazard ratios of PAD and VTE associated with the modifiable risk factors.

Aim and background: The associations of different protein biomarkers with risk of various cardiometabolic diseases remain unclear. The aim of this project is to assess the associations of protein biomarkers with cardiometabolic diseases, with focus on ischemic heart disease, ischemic stroke, hemorrhagic stroke, venous thromboembolism, heart failure and other heart and valvular diseases (atrial fibrillation, abdominal aortic aneurysm, aortic valve stenosis) as type 2 diabetes.

Ischemic heart disease and ischemic stroke are to a large extent driven by factors that promote atherosclerosis and hypertension. However, the protein biomarkers potentially mediating the effect of the risk factors with ischemic heart disease and stroke have not been fully elucidated.

Venous thromboembolism (VTE) refers to a blood clot that starts in a vein. It is the third most common cardiovascular disease after ischemic heart disease and stroke. VTE is a potentially serious disease with the risk of fatal pulmonary embolism. Yet, VTE is often overlooked as a major public health problem. Treatment with anticoagulant drugs are effective but confers an increased risk of complications in the form of bleeding and hemorrhagic stroke. VTE is a multifactorial disease caused by a combination of environmental and genetic risk factors. Among environmental risk factors, some are provoking (e.g., cancer, surgery, immobilization, long-distance travel and hospitalization) and others are nonprovoking (e.g., age, sex, oral contraceptives, statin use, and air pollution). We have found a strong association of obesity and a suggestive association of smoking with increased risk of VTE. Furthermore, we have observed that genetically higher levels of arachidonic acid (which generates TXA2 which has prothrombotic properties) and inflammatory markers are associated with higher risk of VTE.

A previous study found that protein biomarkers involved in apoptosis, inflammation, matrix remodeling, and fibrinolysis were associated with incident heart failure in elderly Swedish individuals. Whether those associations can be replicated warrants investigation. The associations of protein biomarkers with other heart diseases and valvular diseases is unclear.

Hypotheses: We hypothesize that proteins related to atherosclerosis and inflammation will be associated with ischemic heart disease, ischemic stroke, and other atherosclerotic vascular diseases (e.g., abdominal aortic aneurysm and aortic valve stenosis) and that proteins related to coagulation, thrombosis and inflammation will be associated with VTE.

Data sources and methodology: We propose to use proteomics data (CVD II and III and metabolism panels) collected in the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM) clinical sub-cohorts. We would like to obtain data on date of diagnosis for each cardiometabolic disease (listed in the application form) and date of death for those who died during follow-up. Cox proportional hazards regression models will be used to estimate hazard ratios for the associations between circulating proteins and cardiovascular diseases. To be able to adjust for potential confounders, we would also like to obtain data on risk factors for cardiometabolic disease from the questionnaire completed by participants at the time blood sampling.

Cardiovascular disease (CVD) is a leading cause of death and disability worldwide. A growing evidence suggests a multifactorial origin but traditional risk factors such as hypertension and a sedentary lifestyle do not explain the entire risk for CVD. Thus, studies of novel risk factors as well as the combination of risk factors are needed to better understand the causes of CVD. Several studies have demonstrated that modifiable risk factors (e.g., diabetes, sleep-related problems, shift work, unhealthy dietary patterns, and stress) are associated with increased risk for CVD. However, results are inconclusive and the mechanisms underlying the relationship between these factors and CVD are not well understood. Proteomic and metabolomic data analyses may help us to identify new biomarkers for CVDs and suggest potential mechanisms underlying associations between modifiable risk factors and CVDs.

The overall aim is to examine the role of modifiable risk factors (e.g., lifestyle-related risk factors) in the development of CVDs in Swedish males and females. In addition, we aim to investigate if circulating proteins or metabolites are associated with modifiable risk factors for CVD.

Cox proportional hazards regression models will be used to estimate relative risks (95% CI) of CVD associated with modifiable risk factors. We believe that our results will contribute with important scientific knowledge regarding the role of modifiable risk factors in prevention of CVD and thus are of importance from a public health point of view and for societal costs.

Background: Several genetic variants have been identified to be associated with cardiovascular disease but those variants generally explain only a small proportion of estimated heritability. To identify more genetic variants for cardiovascular diseases, meta-analyses of many large cohorts are necessary. In addition, genetic variants have not been found for several potential risk factors for cardiovascular disease.

Aim: The aim of this project, as specified in our original application, is to conduct GWAS analyses to identify genetic variants associated with:

1. cardiovascular diseases
2. potential risk factors, including dietary salt intake, heavy metals, proteins, and anger, for cardiovascular diseases.

Methods: GWAS analyses will be conducted to obtain regression coefficients (betas), standard errors (SEs), and p-values for the associations of all genetic variants with cardiovascular disease and its potential risk factors. These analyses will be done at Uppsala University. To increase statistical power, our summary statistics results (betas and SEs) will be meta-analyzed with the corresponding results from other cohort studies participating in the AAAgen (consortium to identify the genetic variants for abdominal aortic aneurysm), PADgen (peripheral artery disease), and ISGC (ischemic stroke and intracerebral and subarachnoid hemorrhage). The plan is also to conduct GWAS analyses of aortic valve stenosis. In addition, we will perform GWAS analyses to identify genetic variants for potential cardiovascular disease risk factors. Those variants will then be used as instrumental variables in Mendelian randomization analyses to determine whether the risk factors are causally associated with cardiovascular disease and other aging-related diseases and mortality.

Background: Cancer is a complex disease caused by both genetic factors and non-genetic factors, such as diet, obesity, and lifestyle factors. To date, several meta-analyses of genome-wide association studies (GWAS) have been carried out and these studies have uncovered many genetic variants associated with different cancers. Large GWAS consortia are available on common cancers, including prostate cancer (Prostate Cancer Association Group to Investigate Cancer Associated Alterations in the Genome consortium) (1), breast cancer (Breast Cancer Association Consortium) (2), ovarian cancer (Ovarian Cancer Association Consortium) (3), endometrial cancer (Endometrial Cancer Association Consortium) (4), and colorectal cancer (e.g., the Colorectal Transdisciplinary Study) (5). The Swedish Mammography Cohort and Cohort of Swedish Men are included in these consortia. GWASs on other cancers, such as pancreatic, bladder, and blood cancers (i.e., multiple myeloma, non-Hodgkin’s lymphoma, Hodgkin’s lymphoma, and leukemia) as well as overall cancer are either lacking or based on small sample sizes. Therefore, large-scale GWASs are warranted to explore the genetic architecture of these less common cancers.

Aim: The aim of this project is to perform GWAS analyses on cancer, including 1) pancreatic cancer; 2) bladder cancer; 3) endometrial cancer; 4) blood cancers; and 5) overall cancer.

Methodology: Data from the Swedish Mammography Cohort and Cohort of Swedish Men will be used in this project. GWAS analyses will be conducted to obtain regression coefficients and standard errors for the associations between genetic variants and cancer outcomes. To increase statistical power, our summary statistics results will be meta-analyzed with the corresponding results from published consortia (e.g., PanCan Study for pancreatic cancer) and studies, including the UK Biobank, FinnGen study, and BioBank Japan. Our GWAS analyses will be followed by subsequent multiomics analyses of risk factors for the specific cancer and overall cancer using Mendelian randomization designs and other applications using publicly available datasets and UK Biobank.

Background: We and other researchers have shown that modifiable factors, including diet, certain beverages (e.g., coffee, sweetened beverages, and milk), excess adiposity, physical activity, and smoking play an important role in the development of cardiometabolic outcomes, like type 2 diabetes, hypertension, hypercholesterolemia, and cardiovascular diseases. With regard to diet, the Dietary Approaches to Stop Hypertension (DASH) diet has been found to reduce blood pressure and the risk of cardiovascular diseases. Likewise, it has been shown that a plant-based diet rich in healthier plant foods is associated with lower risks CVD.

Emerging evidence also indicates that modifiable factors are associated with gut microbiota composition, and that gut microbiota and associated metabolites are implicated in the development of cardiometabolic outcomes. Nevertheless, few associations have been replicated and few studies have examined whether the modifiable factors-outcome associations may be mediated, at least in part, by gut microbiota and related metabolites.

Aims: Using a comprehensive approach and data from the SIMPLER cohorts, we aim to: (1) examine the associations of modifiable factors and genetics with gut microbiota composition and blood metabolites; and (2) assess whether the identified microbial taxa and associated metabolites mediate the links between modifiable factors and cardiometabolic outcomes and mortality.

The specific objectives are to:

• Examine the influence of adiposity, diet, lifestyle factors, sleep, stress, bowel habits, medication use, pet ownership, gastrointestinal diseases, health status, and genetic factors on variation of gut microbiota and metabolites.
• Assess the associations of adherence to specific dietary patterns, including the DASH, plant-based, and animal-based diets and the included food items with gut microbiota composition, the associations of identified gut microbial taxa with metabolites, and the possible mediating role of gut microbiota and related metabolites in the links between these diets and cardiometabolic outcomes and mortality.
• Examine the associations of beverage (coffee, tea, sweetened beverages, and alcoholic beverages) consumption with gut microbiota composition, the associations of identified gut microbial taxa with metabolites, and the possible mediating role of gut microbiota and related metabolites in the links between beverage consumption and cardiometabolic outcomes and mortality.
• Evaluate the associations of body composition, particularly DXA-derived fat mass and lean mass at various body sites, with gut microbiota composition, the associations of identified gut microbial taxa with metabolites, and the possible mediating role of gut microbiota and related metabolites in the links between body composition and cardiometabolic outcomes and mortality.
• Investigate the associations of lifestyle factors, including physical activity, sedentary behavior (e.g., TV watching/reading and sleep), and tobacco use (smoking and snuff-dipping) with gut microbiota composition, the associations of identified gut microbial taxa with metabolites, and the possible mediating role of gut microbiota and related metabolites in the links between these lifestyle factors and cardiometabolic outcomes and mortality.

Methodology: Data from the SIMPLER cohorts will be used in these studies. Specifically, we will use self-reported, measured, and registry data from participants in the sub-cohorts who provided fecal samples. To define the modifiable factors and metabolite levels, we will use both self-reported data and genetic data, whenever suitable instruments are available. We will randomly group participants into a discovery and replication sample in the analysis of the modifiable factors and gut microbiota composition. Gut microbiota taxa that are replicated will be taken forward for the subsequent analyses in the total sample to increase power in analyses of the outcomes. The cardiometabolic outcomes of interest include blood pressure, blood lipids, glycemic traits, markers of inflammation, metabolic syndrome, type 2 diabetes, and cardiovascular disease, overall and divided into atherosclerotic and thrombotic diseases.

Aim: The aim of this project is to decipher the influence of alcohol and caffeine intake on blood proteins, metabolites, gut microbiome, and major CVDs, type 2 diabetes and breast cancer. The specific aims are:

• assess the associations of self-reported alcohol consumption (total alcohol and specific alcoholic beverages), genetically predicted alcohol consumption (using a variant in the ADH1B gene and other alcohol-associated variants (19)), and blood biomarkers for alcohol consumption with circulating proteins, metabolites, and gut microbiome as well as compare the differences in associations for self-reported, genetically predicted, and biomarker predicted alcohol intake;
• assess the associations of different alcoholic beverages (wine, beer, spirits) with omics patterns;
• investigate whether the identified proteins, metabolites, and gut microbiota species, in aims 1 and 2 are associated with CVD (atrial fibrillation, ischemic heart disease, ischemic stroke, and hemorrhagic stroke) and breast cancer incidence, as well as the potential mediating effect.
• assess the associations of self-reported coffee and caffeine intake, genetically predicted caffeine intake, and blood biomarkers for coffee consumption with circulating proteins, metabolites, and gut microbiome;
• investigate whether the identified proteins, metabolites, and gut microbiota species, in aims 4 are associated with type 2 diabetes incidence, as well as the potential mediating effect

Methodology: Data from the SIMPLER cohorts (Swedish Mammography Cohort and Cohort of Swedish Men) will be used in this project. Cross-sectional analysis will be used to assess the associations of alcohol consumption (total and specific alcoholic beverages) with proteins and metabolites. For aim 3, we will use Cox proportional hazards regression analysis to investigate whether the identified alcohol-associated proteins and metabolites are associated with incidence of atrial fibrillation, ischemic heart disease, ischemic and hemorrhagic stroke, and breast cancer in the SIMPLER cohorts. In the analysis based on self-reported alcohol consumption, the model will be adjusted for potential confounders, such age, sex, Charlson comorbidity index, education level, income (from LISA register), smoking, body mass index, physical activity, a healthy diet (e.g., modified MED or DASH diet score), energy intake, and reproductive factors (in analyses of breast cancer).

In the observational analyses, self-reported alcohol consumption will be analyzed both as a continuous exposure (per drink; 1 drink is defined as 12 g alcohol) in alcohol drinkers and as a categorical exposure (no intake or occasional drinkers [defined as <2 drinks/month]); light consumption [³2 drinks/month to 1 drink/day]; moderate consumption [>1 to <3 drinks/day]; and high to heavy consumption [³3 drinks/day]). In the MR analyses, genetically predicted alcohol consumption will be analyzed both as a continuous exposure in self-reported alcohol drinkers as well as having none, one, or two alcohol-increasing variants of the ADH1B gene. Genetically predicted alcohol consumption will also be analyzed as a continuous variable using all identified alcohol-associated variants from the GWAS & Sequencing Consortium of Alcohol and Nicotine use. Former drinkers will be excluded from all analyses.

Bakgrund: Menopausal hormonbehandling (MHT) är idag den mest effektiva behandlingen mot klimakteribesvär som kan vara svåra, inskränka stort på det dagliga livet samt i ett längre perspektiv innebära ökad risk för ohälsa. I rådande behandlingsriktlinjer framhålls att nyttan med MHT dock ska ska vägas mot risker för framtida sjukdom som behandlingen kan medföra. Trots att en stor mängd forskning gjorts inom detta område är kunskapen om hur MHT ökar eller minskar risk för framtida sjukdomar och för tidig död fortfarande begränsad. En av anledningarna till detta är att de flesta tidigare studier inte beaktat tidpunkt för insättning av behandling i relation till tidpunkt för menopaus - något som under senare år i forskningen visat sig ha stor betydelse. Det finns ett behov inom detta forskningsfält att systematiskt studera hur MHT kopplar till risker och möjliga nyttoeffekter vad gäller framtida sjukdomar - under beaktande av tidpunkt för insättning.

Mot bakgrund av en hypotes om att östrogen motverkar låg-gradig inflammation och därmed skulle kunna skydda kvinnor mot kroniska sjukdomar och minska risken att avlida i samband med ett akut insjuknande/trauma, är det intressant att studera om MHT associerar med överlevnad i samband med ett eventuellt hjärtstopp senare i livet. Detta har hitintills inte studerats. Däremot har ett fåtal epidemiologiska studier undersökt om pågående eller tidigare MHT associerar med minskad 28-dagarsdödlighet i samband med akut hjärtinfarkt hos menopausala kvinnor, och visat att så är fallet. I en studie av överlevnad i samband med hjärtstopp baserad på svenska register kunde dock ingen könsskillnad noteras efter justeringar för möjliga störfaktorer såsom ålder, vilket talar emot att kvinnor skulle ha ett skydd kopplat till östrogennivåer.

Huruvida kvinnliga könshormoner bidrar till att skydda mot ett allvarligt sjukdomsförlopp eller död i samband med insjuknande i Covid-19 har diskuterats, eftersom kvinnor visat sig ha lägre dödlighet i Covid-19 än män. Baserat på hypotesen beskriven ovan om att östrogen motverkar låg-gradig inflammation och därmed skulle kunna skydda kvinnor mot kroniska sjukdomar respektive död i samband med akut insjuknande/trauma, skulle MHT kunna skydda mot att drabbas allvarligt eller att avlida i samband ett eventuellt insjuknande i Covid-19 senare i livet.

Syfte och vetenskapliga frågeställningar: Forskningsprojektets övergripande syfte är att öka kunskapen kring hur MHT inverkar på kvinnors risk att drabbas av framtida kronisk sjukdom och förtida död. Projektet har både en bred ansats, där ett stort antal olika kroniska sjukdomar ska beaktas samtidigt, och några smalare ansatser där specifika sjukdomsdiagnoser ska beaktas separat vad avser risk för insjuknande och/eller svårighetsgrad/letalitet. Specifika sjukdomar för analys inkluderar demenssjukdomar respektive hjärntumör. En specifik frågeställning är också om huruvida nuvarande eller tidigare MHT är associerad med risk att avlida inom 60 dagar eller att behöva intensivvård i samband med eventuellt insjuknande i Covid-19. En motsvarande frågeställning föreligger avseende risk att avlida eller behöva intensivvård i samband med ett eventuellt hjärtstopp. Vid besvarande av forskningsfrågorna ska projektet beakta tidpunkt för insättning av MHT samt dess typ, duration och administreringsväg. I subanalyser ska genetisk bakgrund beaktas. Material och metod Projektet baseras på en sammanslagning av data från sex populationsbaserade befintliga svenska kohortstudier samt kontrollindividerna i en svensk fall-kontrollstudie. Dessa kohortmaterial listas i tabell 1 nedan. Totalt kommer data från cirka 110 000 menopausala kvinnor att bearbetas varav cirka 60% rapporterat användning av MHT. Av dessa har ca 40% rapporterat tidig initiering av MHT i relation till tidpunkt för menopaus. Data om användning av MHT i dessa studiematerial samlades in med hjälp av frågeformulär eller intervjuer.

Background: It is well known that cardiovascular risk factors often cluster in obese individuals, a phenomenon that has been termed the metabolic syndrome (MetS). However, a proportion of the obese do not show other risk factors, at least not at an initial investigation. This condition has been termed metabolically healthy obesity (MHO).

The first longitudinal studies suggested that MHO was a rather harmless condition in terms of risk of future cardiovascular disease (CVD). However, later studies with a longer follow-up periods disclosed that this was just the case during the first years of follow-up and that the risk increased after 10 years. Thus, two different meta-analyses published in 2013 concluded that MHO is not a harmless condition, and a recent large population-based study in approximately 500,000 individuals (Epic-CVD) with more than 7,000 incident cases of myocardial infarction, and thereby a very good power to investigate the impact of the interplay between obesity and other risk factors, concluded that “irrespective of metabolic health, overweight and obese people had higher coronary heart disease risk than lean people”, challenging the use of the term MHO.

However, there is no widely accepted gold standard definition of MHO. Some investigators suggest that MHO should be defined as obesity with the absence of any other risk factors. A more commonly used and less strict definition is that the MHO is present if obese individuals do not have the MetS, i.e at least two other MetS components than abdominal obesity.

Aim of the research plan: The fist overall aim is to identify factors (in this case proteins) that differ between metabolically healthy and unhealthy obesity (as well as overweight) that might explain why certain obese (or overweight individuals) develops metabolic derangements and others do not.

The second overall aim is to study which proteins that are common for the five components of the metabolic syndrome; overweight/abdominal obesity, high blood pressure, high triglycerides, low HDL, and an impaired glucose tolerance.

The third overall aim is to evaluate how different definitions of metabolically healthy and unhealthy obesity (as well as overweight), as well as different combinations of the five components of the metabolic syndrome are associated with later development of cardiovascular diseases (CVD) and mortality.

The fourth overall aim is to evaluate how proteins linked to the Metabolic syndrome and different combinations of the components are related to future CVD.

The aims in this research program will be answered by the use of data from four large Swedish epidemiological samples (EpiHealth, SCAPIS, COSM, SMC), and we strive to use as many samples as possible in meta-analyses of the different samples depending on the research question. All 4 samples do not have data needed for each of the research questions.

The materials will be divided into three different group regarding BMI: Normal-weight; BMI<25 kg/m2, overweight; BMI 25-30, obesity; BMI>30 kg/m2.

Metabolically healthy vs unhealthy will be defined by the occurrence of the metabolic syndrome (MetS) or not.

The different MetS criteria were defined according to the harmonized criteria (1); 1) Blood pressure ≥ 130/85 mmHg or antihypertensive treatment, 2) fasting plasma glucose ≥ 5.5 mmol/l or use of glucose lowering drugs, 3) plasma triglycerides ≥ 1.7 mmol/l, 4) high-density lipoprotein (HDL) cholesterol < 1.0 mmol/l.

In COSM and SMC, the examination cycle which first collected data needed to define MetS and the three OLINK chips (CVD II&III and metabolism) will be used also as baseline for AIMS 3 and 4.

Background: Critical illness leading to intensive care means suffering and high risk of death for the individual, and costs to society. There are several reasons why patients end up in intensive care units (IVA), but common to them is that they develop failure in vital organs. Sepsis, COVID-19 infection, and surgical trauma are common causes of failure of vital organs. During the past year, these conditions have accounted for a very large proportion of all admissions to IVA in Sweden. Apart from causing significant hospitalization, these three causes of illness are also responsible for great suffering and significant mortality. One in three patients admitted to intensive care unit (ICU) in Sweden with sepsis dies 30 days after admission. Similarly, one in four patients admitted to a Swedish ICU in the first wave of COVID-19 (Corona virus disease19) died. 15% of patients die who are admitted unplanned to IVA after surgery according to the Swedish Intensive Care Registry's output portal (30-04-2021). These mortality figures are comparable to figures seen for 5-year mortality from cancer, with the difference that they are measured one month after the illness. Furthermore, these admissions to IVA are associated with substantial burden on costs and human resources. Sepsis alone caused 10,000 care admissions (11% of all ICU care admissions) and 49,000 days (20% of all ICU days) in Swedish intensive care units during 2018-2019 with an estimated cost of SEK 2.5 billion for intensive care alone. The cost of COVID-19 is significantly higher. Unplanned admission to IVA after surgery means 12,000 ICU days each year and entails an additional cost of half a billion kronor. The reason why patients need organ-supported treatment for IVA varies. We have a special interest in acute renal failure. Acute kidney failure, which according to the latest definition is called acute kidney injury, had a 30-day mortality of 35% during the years 2018-2019 according to the Swedish Intensive Care Register's out-data portal (25-03-2021). Given the high mortality for these hospitalization causes and acute kidney injury, a small reduction in mortality can be of great importance to the individual, healthcare and society. We have previously shown that genetic factors are important in disease severity and outcome in COVID-19. With an increased understanding of disease mechanisms, key factors that affect disease severity and outcome after illness can be identified, as well as risk groups of individuals where healthcare can make preventive measures or long-term targeted treatment measures.

Hypothesis: We have hypothesized that specific genetic factors are of great importance for the severity of disease in in critical illness such as sepsis, infection with Sars-CoV-2 and infectious or thromboembolic complications after surgery.

Aim: The project will identify whether specific genetic factors contribute to increased severity of disease, increased level of hospital care and increased mortality in severe infections or after surgery.

Method: We will merge pseudo-anonymised data from Epihealth, Simpler, the Swedish Twin Registry, The PRONMED study, the Swedish ICU registry, the Swedish perioperative registry, Patientregistret on comorbidities, Läkemedelsregistret for comorbidities, socioeconomic factors from SCB, microbiologic data from Smitnet and regional microbiological labs. SCB and Socialstyrelsen will handle personal ID data. Please see the attached Data acquisition flowchart document. The data delivered to us will be stored and analyzed on an UPPMAX server for analyzing sensitive personal data. currently called the Bianca Cluster.

We perform quality control and filtering of genotyped data at individual and SNP level as appropriate. Genotype data will be imputated to overcome missing data and chip differences. In the main analysis, we will control for age, sex and other identified confounders such as population structure. For phenotype data will be imputated if data missing at random. We will perform genome-wide association studies and use identified SNPs in regression models. Dichotomous data for main outcomes will be assessed with logistic regression. Continuous data will be assessed with linear models. We adjust for multiple comparisons as appropriate. Fixed effect meta-analysis will be used in the meta-analyses with other studies. SAIGE R library and Regenie will be used.

Aim 1: Characterize gut microbiota variation, asymptomatic pathogenic carriage and associated antibiotic resistome

Research questions:

1. Investigate the diversity, composition, and prevalence of the gut microbiome and pathogenic bacteria in the human population, and links with various known covariates (medication prescription history, biometrics, host genetic effects);
2. Understand individual variability of the gut microbiome and asymptomatic pathogenic carriage based on demographics, genetics, dietary habits, and other relevant variables.

Methodology: Metagenomic classification using custom databases and pipelines: we will use custom Kraken2/Bracken indices built on GTDB definitions (release RS08-214; April 2023), which have been used in our lab in previous studies. Pathogen detection and validation: we will use a combination of custom mapping pipelines along with validation using low-abundance strain-level profiling tools such as InStrain or StrainGE or similar. Additional validation and resolution will be achieved from virulence factor profiling and RPKM normalization accounting for sequencing coverage similarities. Resistome profiling: we will use accepted pipelines with custom approaches developed in our group, such as RGI and based on the Comprehensive Antibiotic Resistance Database (CARD). Datasets and covariates: we aim to investigate the role of antibiotic and non-antibiotic medication prescription and history (a very important confounder of gut variation) and other covariates on gut microbiome and pathogenic carriage variation, to properly account for them in subsequent comparisons and associations.

Aim 2: Develop predictive models for health outcomes based on gut microbiota variation (including pathogenic carriage) combined with conventional risk factors and other phenotypes, including high-dimensional data.

Research questions: Implement machine learning algorithms to perform predictive analysis for clinical outcomes using gut microbiome, conventional risk factors, clinical metadata and multi-omic features (including genomics, proteomics and metabolomics), and integrate multi-modal characterization of common diseases

Methodology: Modelling and machine learning approaches have been described in 2 publications from our group. We will combine PRS, gut microbial features, circulating metabolomes, proteomes and conventional risk factors for each disease of interest into a prediction model to assess performance increases to predict disease risk and outcome, if available. We aim to reproduce and complement preliminary results obtained using other cohorts.

Aim 3: Linking disease risk with gut microbiota variation and function.

Research questions: Using PGS to stratify cohorts and associate microbiome with predicted genetic disease risk rather diagnosed outcomes could highlight particular microbial contribution to disease. An underlying ecological rationale is that as variations in microbial taxa in the gut can be genetically associated with human genotypes (REF Owen), genetic variation could create varying micro-ecological niches that could promote or not different microbial features. Disease risk may provide an early indication of these microniches.

Methodology: We will derive PRS for non-communicable diseases of interest from genotype data (in a first stage, coronary artery disease, T2D, Alzheimer’s disease and prostate cancer) to segregate individuals according to their disease risk scores.

Aim 4: Evaluate the metabolomic and proteomic profiles associated with gut microbiota compositions in SIMPLER, and associate with various health variables and outcomes.

Research questions:

1. Investigate the genetic association and causal links between circulating metabolites and gut microbiome variation;
2. A large number of molecular genetic scores can be used to cluster populations, which could be then used to associate with microbiome and disease. It might help us to better understand disease via this multi-level approach. We aim to develop genetic scores to predict multi-omic traits, including gut microbiome.

Methodology: We have had a particular interest in linking features measured using untargeted plasma metabolomics using LC-MS at Chalmers as a new possible resource for multi-omics genetic score development, as part of ongoing projects. In addition, O-Link characterization of proteomes could be used to validate existing protein genetic scores that we have, as part of the OmicsPred project, and link the gut microbiota to disease-associated metabolites and circulating proteins. Additionally, Mendelian Randomization will be used as described in Qin et al. 2020 & 2022: briefly, we will also plan to use a combination of various MR approaches after having performed GWAS using common approaches to infer causal links between gut microbial abundances and metabolite/protein concentrations.

Aim 5: Genome-scale analyses of gut microbiome and whole blood metabolism.

Research questions: Investigate the association between gut microbiome and host metabolism through personalized genome-scale models of gut microbiome and whole blood metabolism. From personalized genome-scale models of gut microbiome metabolism we will predict the metabolic functions and capacities of the microbiome of each individual (e.g., capacity to metabolize a drug or produce metabolite). Such measures will be integrated as features for risk prediction (Aim 2) and used to help establish a functional link between gut microbiome and metabolite concentrations in blood (Aim 4). Microbiome composition and personalized metabolic functions will also be cross-referenced with personalized metabolic models of whole-blood metabolism to unveil their association with the activity of major metabolic pathways in blood cells.

Methodology: Genome-scale reconstruction of microbiome metabolism will be achieved by mapping gut microbiome strains and their abundance to strain-specific metabolic models. Personalized models of whole blood metabolism will be built by mapping the proteome and metabolome measures to a human genome-scale metabolic using the quadratic metabolic transformation algorithm.

Aims/objectives: This project will investigate

1. The associations between social support and the risk of chronic diseases;
2. The associations between social support and the all-cause and cause-specific mortality

Analysis methods: This project will be prospective cohort studies based on linking Swedish Mammography Cohort (SMC), the cohort of Swedish men (COSM) to the Swedish national registers. The main exposure variables will be social support measured in SMS questionnaire 2008 and COSM questionnaire 2008; family status; number of people in household; living alone for how long; are you active in any club/organization. The outcome variables will be the first occurrence of cardiovascular diseases (CVD), diabetes, and cancer. The all-cause and cause-specific mortality. Participants will be followed from the date of interview to the first occurrence of the diabetes/CVD/cancer, or death, or the end of the follow-up, whichever occurred first.

COX proportional hazards models to evaluate the associations (hazard ratios and 95% confidence intervals) between each of the exposures and the occurrence of CVD, or diabetes or cancer, and mortality respectively. The survival time was defined as the duration from the start of the follow-up to the end of follow-up as described above.

Potential implications of study: This project is highly relevant from public health perspective. Social support is modifiable factors and could be provided from the society or the health care personals in health promotion programs. The finding from this project would project scientific basis for design effective health promotion and disease prevention.

Aims

1. To estimate the impact of season for determination of the lower insufficiency level and upper safe threshold of total S-25OHD considering the disease burden in women
2. To determine the vitamin D content in fat tissue (from fat biopsies) and the total body content of vitamin D (in fat mass and in circulation) by season in relation to serum concentrations of vitamin D, bone mineral density (BMD) and parathyroid hormone levels
3. To estimate the impact of season for determination of the lower insufficiency level of free S-25OHD in relation to bone mineral density and parathyroid hormone levels, also taking into account vitamin D intake.
4. To investigate the impact of a genetic risk score in women who maintain low S-25OHD levels in summer, in those who have average values and in those who reach high summer values
5. To examine the relation between baseline vitamin D status, change in vitamin D status and change in bone mineral density in samples taken at 10 years apart, by season and vitamin D intake

Background: Every year millions of people throughout the world suffer from osteoporotic fracture and these fractures are associated with decreased quality of life, increased mortality and high health care costs. One major determinant of fractures in elderly is low bone mineral density (BMD). Chronic inflammation (CI) can be induced by aging, low sex hormones, obesity, diet and smoking and is typically featured with low level of inflammation throughout the body. Evidence showed that T cells persistently produced by CI can directly activate osteoclastogenesis and together with pro-inflammatory cytokines accelerate bone resorption and thus CI is considered to increase risk of osteoporosis and fragility fractures in the elderly. Thus, it is important to explore preventive strategies that may impact inflammatory level and can be modified by intervention. One such modifiable lifestyle factor is diet which has been shown to modify inflammation through both pro-inflammatory and anti-inflammatory mechanisms. However, the influence of individual food components on BMD and osteoporosis is weak to moderate and evidence is inconclusive. In fact, all nutrients and other bioactive compounds are consumed as foods in the context of a diet and the food matrix may deliver synergistic effects beyond the sum of its components. Consequently, the outcome usually occurs as a result of natural interactions and patterns of diet. Dietary patterns in a population can be formed either using a posteriori data driven methods or a priori forming diet indices (scores), both with strengths and pitfalls.

Previous studies have shown that a priori formed diet scores, such as Mediterranean Diet characterised by high intakes of antioxidants are associated with higher BMD in the elderly populations. Currently, there are three diet-based inflammation indices, including DII, E-DII and AIDI. Of those, DII and its modified versions are literature review-based score in different populations and E-DII was developed in the U.S. population. AIDI was developed using food questionnaire data from Swedish Mammography Cohort Clinical (SMCC) and showed a robust association with circulating inflammatory marker hsCRP and thus can be used to predict the low-grade systematic inflammation in the Nordic populations.

In addition, a study showed that the high intake of “vitamin C and β-cryptoxanthin” pattern formed by principal component factor analysis is associated with higher radial BMD in post-menopausal females in Japan. By using the same method, a higher prevalence of low BDM was found in individuals with dietary pattern dominated by processed red meat and snacks in an Australian population. Compared to factor analysis which focus on the patterns explaining the variation in food intake, another a posteriori data driven method, reduced rank regression (RRR) was shown to be able to form the patterns that more closely related to specific disease by using disease-related response variables. For example, the dietary pattern best explaining intake of protein, calcium, potassium identified by RRR was associated with BMD in a British female population. Such results however, need validation in other European populations but there are few studies applying RRR to study association of diet intakes and BMD change.

It is worth noting that the dietary patterns used to associate with BMD is established only by subjective recalls of food intake in most current studies, which might be inclined to systematic and random errors and such errors could hamper scientific findings and interpretations. Thus more reliable dietary assessment methods are needed to understand the link between diet and CI-related disease profiles. Metabolomics is a field of OMICs science using cutting-edge analytical chemistry techniques and advanced algorithms to characterize complex biochemical systems. In untargeted metabolomics, as many metabolites as possible are analyzed and differences in profiles can be assessed to find specific markers that are related to processes of disease and exposures of dietary patterns. How dietary intake modulates bone health and the underlying metabolic mechanisms may be captured by analysis of metabolomics. At present, there are no large-scale cohort studies examining the relation between dietary patterns and bone phenotypes using metabolomics.

Aims and hypothesis: The general aim of this project is to identify biomarkers of aforementioned dietary patterns and associations with likelihood of developing osteoporosis in a case-control study nested in SMCC:

1. To identify the biomarkers of dietary patterns to objectively reflect dietary intake;
2. Using biomarkers of dietary patterns as mediators to investigate how dietary patterns associate with the risk of osteoporosis.

Methodology:

1. Three different strategies will be performed to establish the dietary patterns using FFQ data of SMCC, including one a priori method (AIDI) and two a posteriori methods (factor analysis and RRR);
2. Metabolomics data of SMCC will be subjected to multivariate modelling in a well-established R package “MUVR” to select most important metabolite features to best predict the specific dietary patterns;
3. The selected biomarkers of a dietary pattern will be reconstructed into two principle components (PC) by PCA;
4. Correlation analysis of PCs (scores) and dietary patterns as well as individual food groups will be performed;
5. PCs (scores) will be subjected to (unconditional) logistic regression to investigate the odd ratio (OR) of osteoporosis with each PC;
6. A triplot that superimposes metabolites, dietary variables and ORs will be generated to visualize the association of diet and the likelihood Body mass index and Mediterranean diet in relation to mortality rates in women and menof developing of osteoporosis.

Cardiovascular diseases, stroke, peripheral and aortic calcifications predispose to future higher risk of both hip fracture and other types of frailty fractures, independent of previously recognized clinical risk factors. Conversely, increased bone resorption, osteoporosis, and prior frailty fractures are associated with higher rates of cardiovascular events in both women and men. Identical co-twin control analyses indicate that common genes may predispose to the development of both CVD and fractures. Interestingly, the age-adjusted incidence of both myocardial infarction and stroke declined in Sweden by more than 40% during 1998-2016, and hip fracture rates also declined in parallel. The lower fracture rates are not explained by bisphosphonate use, at an ecological level. Since there exists strong indications of common disease pathways for the occurrence of frailty fractures and CVD, we need to better understand these mechanisms; from genes and lifestyle to disease occurrence. Proteomics offer new possibilities for investigating the molecular underpinnings of these chronic diseases.

While rigorously conducted genomic discovery and replication studies have been accomplished, similar proteomic replication analyses are hitherto non-existent. Only one recent cohort project, of 1800 men, has used proteomics and linkage to BMD. Twenty annotated proteins were identified as related to BMD and five of these were further associated to incident hip fractures. No replication cohort was used, fracture related proteins independent of BMD were not identified, the concentration of proteins was estimated from peptide compositions and not directly measured, and the 14 most abundant serum proteins were deleted from the samples. Comprehensive studies, including simultaneous information about ‘omics’ with a discovery and replication design are therefore needed. Such information, together with longitudinal lifestyle data, can discover common pathogenic pathways between CVD and fractures, including via BMD, fat mass or sarcopenia. Causal inference of identified risk proteins can be accomplished by a complementary Mendelian randomized design. We therefore aimed to identify novel protein biomarkers of risk for frailty fractures ̶ shared or non-shared with BMD, fat or lean muscle mass ̶ using proteomics, genomics and lifestyle information, also aiming to discover common pathogenic mechanisms for cardiovascular diseases and fractures.

Aim: Harmonize untargeted metabolomics data and data processing pipelines between the CMSI metabolomics platform at Chalmers University of Technology and the Metabolomics Platform at the Broad Institute of MIT and Harvard.

Samples: Li/Hep blood plasma samples from SMCC, 247 samples.

Method: The CMSI utilizes a combination of two identical QTOFs operating with orthogonal chromatographic techniques (reverse phase [RP] and hydrophilic interaction liquid chromatography [HILIC]). Each sample is injected twice, one time for each electrospray ionization mode (ESI+ and ESI-), maximizing compound coverage and method sensitivity due to iFunnel technology.

Stress and anger are two commonly experienced emotional states that can have profound effects on an individual's physical and mental well-being. In recent years, there has been a growing recognition of the detrimental health consequences associated with chronic stress and anger. This has prompted extensive research efforts aimed at elucidating the underlying mechanisms and understanding the specific health risks posed by these emotional states.

Chronic stress, defined as the prolonged activation of the stress response system, has been implicated as a significant contributor to various adverse health outcomes. It is well-established that stress can have detrimental effects on the immune system, cardiovascular system, endocrine system, and neurological functions. Several studies have reported associations between chronic stress and increased risk of cardiovascular diseases, such as hypertension, coronary heart disease, and stroke. Moreover, chronic stress has been linked to dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, leading to alterations in cortisol levels and potentially contributing to the development of mood disorders, such as depression and anxiety.

In addition to stress, anger is another emotional state that has garnered significant attention due to its potential health risks. Anger, characterized by feelings of displeasure, hostility, and an inclination toward aggressive behavior, has been associated with adverse physiological and psychological consequences. While evidence indicates that strong anger may trigger acute cardiovascular events, the long-term influence of an angry behavior in general on CVD events has been scarcely investigated and the results are inconsistent. Research suggests that chronic anger may contribute to the development of cardiovascular diseases, similar to the effects of chronic stress. Several studies have demonstrated a positive association between anger and hypertension, increased arterial stiffness, and a higher risk of myocardial infarction. Moreover, anger has been linked to alterations in inflammatory markers, oxidative stress, and immune system dysregulation, all of which may contribute to the development of various health conditions, including metabolic disorders, autoimmune diseases, and even cancer.

While the health risks associated with stress and anger have been studied individually, there is a need for further investigation into the causal cumulative impact of these emotional states on health outcomes. By examining stress and anger together, it is possible to gain a more comprehensive understanding of the mechanisms involved and their combined effects on various physiological systems. The present study aims to address this gap in knowledge by investigating the conceivable true impact of chronic stress, anger, and their potential synergistic effects on health outcomes.

Mendelian randomization (MR) is an increasingly exploited study design that can improve causal inference in observational studies by leveraging genetic variants that strongly associated with the exposure (e.g., stress or anger) as instruments to decipher the causal effect of the exposure on the outcome (e.g., ischemic heart disease or stroke). Owing to the unchangeable nature of genetic variants and the random allocation of alleles at conception, MR studies diminish reverse causation as well as confounding from self-selected behaviors and environmental factors. With the use of ordinary cohort designs, it will be difficult to decipher the true independent effect of stress or anger on the outcome. Since established and validated genetic instruments for stress and anger have not been developed, our primary aim was to perform a GWAS for those exposures with the use of SIMPLER, and a development and validation design. Aggression and perceived stress in humans are heritable, but their genetic basis remains to be uncovered. No sufficiently large GWASs have been carried out yet.

In conclusion, chronic stress and anger have emerged as significant risk factors for various adverse health outcomes. Understanding the true evidence-based impact through which these emotional states contribute to disease development is crucial for developing effective prevention and intervention strategies. The current study seeks to contribute to the existing body of knowledge by investigating the combined impact of stress and anger on health and exploring a potential causal mechanism of disease development. For such purpose, we need to develop genetic instruments – a first aim of our study. If the development is successful, we will use these instruments, both independently and jointly, on different health outcomes.

Methods: We will use GWAS data within SIMPLER and maybe also EpiHealth for replication to identify strong genetic markers for self-reported anger and perceived stress (health questionnaire questions 36 and 37). If such markers are found, we will widen the study to also an examination of the genetic instruments in relation to different health outcomes.

Sweden has one of the highest incidences of frailty fractures in the world, with a lifetime cumulative incidence of 50% in women and 25% in men. The risk of hip fracture, the most devastating fragility fracture, increases 44-fold in Swedish women from 55 to 85 years of age, so the lifetime risk of hip fracture is 25% in women and 12% in men. Although osteoporosis, low bone mineral density (BMD) is a strong risk factor for fractures, less than one in three hip fractures is attributable to osteoporosis, and for other types of frailty fractures, the proportion is even lower. Depending on the country, the average residual lifetime risk of hip fracture for women at 50 years of age is 10-25%, with the highest risk found in northern Europe. It is a significant cause of disability and mortality in older people, for whom it is associated with low quality of life and long-term need for institutional care. Because of population aging, the total health loss due to hip fractures is estimated to double in the next two decades. There is thus an urgent need to understand and mitigate the causes of hip fracture.

The principal causal components of frailty fractures include not only weakened bone but also falls, possibly due to age-related factors such as sarcopenia (muscle shrinking and impaired muscle strength), comorbidity, or side effects of medication influencing balance. Muscle size, associated with comorbidity and physical function, is an essential determinant of hip fracture risk. During adult aging, skeletal muscle mass and muscle strength fall on average by 40%.

While rigorously conducted genomic discovery and replication studies have been accomplished, similar metabolomic and proteomic replication analyses for identifying determinants for fragility fractures are hitherto non-existent. Metabolites are small molecules that are intermediate or end products of metabolic reactions. Many factors, including genetics, disease status, intestinal microbiota composition, and lifestyle influence their levels. Metabolic processes can affect disease risk and provide therapeutic targets. Metabolites causally associated with bone mass have been identified in a Mendelian randomization approach. Although the heritability of a proportion of all metabolite levels is high, enabling Mendelian randomization for those metabolites, many other metabolites are mainly determined by lifestyle factors. Recently, direct large-scale measurement of circulating metabolites or metabolites in adipose tissue has become a reality, but no previous study has linked truly untargeted metabolomics to future fracture risk. Still, experimental and small human cross-sectional studies are promising.

Circulating proteins act as essential keys in disease-developing paths, and proteomics offers new possibilities to discover molecular underpinnings of fragility fractures, including intrinsic or extrinsic mechanisms. While rigorously conducted genomic discovery and replication studies with bone phenotypes have been accomplished, no such designs using proteomics and with fragility fracture rates as a primary outcome exist. Only one recent cohort project, of 1800 men, has used proteomics and linkage to BMD. Twenty annotated proteins were identified as related to BMD, and five were further associated with incident hip fractures. No replication cohort was used, fracture-related proteins independent of BMD were not identified, the concentration of proteins was estimated from peptide compositions and not directly measured, and the 14 most abundant serum proteins were deleted from the samples

Genomic data combined with Mendelian randomization (MR) techniques opens the possibility of directly determining the effect of circulating metabolites and proteins on complex outcomes, such as fragility fractures. More data are needed to elucidate the causal role of circulating metabolites and proteins in fragility and hip fracture risk.

Our study aims to uncover important unidentified metabolite-regulated mechanisms of fragility fractures, particularly hip fractures, by applying an untargeted exploratory study design using two cohorts of older Swedish women and men. We first aim to discover novel metabolic and proteomic fingerprints reflecting a higher risk of fragility fractures and hip fractures specifically. We will then examine common pathways for these identified metabolites and proteins and their relation with bone mineral density and frailty parameters (gait speed, hand grip strength, and chair stand test). Finally, we aim to determine the causal effect of these metabolites and proteins on bone mineral density and femoral fracture risk using two-sample MR.

Men and women with type 2 diabetes mellitus (T2DM) have higher risk of hip fracture, but the mechanisms are not fully understood. One such mechanism may be lower bone area. We have shown in a cross sectional study including SMCC and ULSAM men that those with T2DM had lower bone area compared to those without T2DM. We identified men and women with normal fasting glucose (NFG), impaired fasting plasma glucose (IFG), and T2DM. Bone mineral density and bone area at the total hip and femoral shaft were measured using dual energy X-ray absorptiometry (DXA). Men and women showed a progressively higher BMD following the clinical cut-offs of fasting glucose from NFG to IFG to T2DM. In contrast, there was a progressively lower bone area. We now aim to further explore this question with longitudinal data and assess changes in BMD and bone area over time in those with T2DM compared to those without. T2DM will be defined using a combination of fasting plasma glucose, diabetes medicine, and self-reported diagnoses from questionnaires and BMD and bone area will be determined from dual energy x-ray absorptiometry (DXA). This study will hopefully show that bone area decreases over time in those with T2DM compared to those without. This may highlight a mechanistic reason for why those with T2DM have an increased risk of hip fracture due to decreasing bone area which may highlight an area for intervention whether it be pharmacologically or via another form of intervention to reduce this decrease in bone area.

Men and women with type 2 diabetes mellitus (T2DM) have higher risk of hip fracture, but the mechanisms are not fully understood. One such mechanism may be lower bone area. We have shown in a cross sectional study that men and women had a progressively lower bone area following the clinical cut-offs of fasting glucose, from normal fasting glucose (NFG) to impaired fasting glucose (IFG) to T2DM.

There have been two studies using Mendelian Randomization (MR) methods to estimate the genetic effect of both T2DM and glucose levels on bone mineral density (BMD) and the risk of fracture. We intend to replicate the results found in one of these studies by showing the genetic effect of glucose on greater BMD but aim to add analysis on the genetic effect of glucose levels on bone area. We will conduct a two-sample MR with summary data on glucose related single nucleotide polymorphisms (SNP’s) from the MAGIC consortium GWAS in combination with data from the SMCC and UKBB. Glucose concentrations in SMCC will be used for estimating the association between them and the SNP’s of interest.

This study may highlight a mechanistic reason for why those with higher levels of glucose leading to a greater risk for T2DM have an increased risk of hip fracture. Therefore if we are able to determine those individuals with a genetic predisposition for higher glucose levels and T2DM it may be possible to intervene to ensure the expansion of bone area is not diminished.

Background: Consumption of red and processed meat has been associated with increased risk of type 2 diabetes (T2D), but the results remained inconsistent. One of the responsible factors is the risk of misreporting of the dietary intake, which calls for robust biomarkers to represent meat intake.

Aims and hypothesis: The main aim of the study is to mine potential biomarkers for meat intake by comparing self-reported intake of various meat groups with the plasma metabolome, and to validate the findings by replicating the study in independent cohorts. The secondary aim is to investigate the association between the biomarkers and the risk of developing T2D in the main cohort and two independent cohorts. These aims are then broken down into three objectives:

1. To establish intake biomarkers of different groups of meat: total meat, poultry processed, and unprocessed meat in the main cohort
2. To validate the findings by replicating the study in other study populations
3. To investigate the associations between the biomarker candidates and the risk of T2D and cardiometabolic outcomes in the main and in two replication cohorts

We expect to find both previously reported and novel metabolites to reflect meat intake. We also hypothesize that individual factors, such as age, BMI, and total energy intake, may vary the levels of these biomarker candidates in the blood and confound the associations with the cardiometabolic outcomes and T2D risk.

Methods: A nested case-control study from a population-based prospective Västerbotten Intervention Program (VIP) that started in 1985 in Northern Sweden served as the discovery cohort. 421 participants (225 women and 196 men) who developed T2D in the median diagnosis period of 7 years and controls with matching age, sex, ethnic group, and time of baseline visits, were selected. Using untargeted metabolomics approach, we applied both multivariate and univariate analyses to unbiasedly select panels of biomarker candidates to reflect total meat, processed and unprocessed meat, and poultry intake. The findings were then replicated in two independent cohorts: a clinical subcohort of Swedish Mammography Cohort in Uppsala (n=5032 women) and a subset of Kuopio Ischaemic Heart Disease Risk Factor Study in eastern Finland n = 239). Additionally, the associations between these metabolites and the risk of T2D and cardiometabolic outcomes were also assessed.

The basic research problem: Gradual loss of bone and muscle mass with ageing leads to increased prevalence of osteoporosis and sarcopenia, with a subsequent increased risk of fractures in the elderly population. Fractures, especially those of the spine and hip, result in reduced quality of life, increased morbidity and mortality as well as high societal costs. There is a medical need for new fracture preventive therapies with minimal side effects and novel biomarkers that improve the prediction of fracture risk. The aim of the proposed research is to test our hypothesis that the increased fracture risk in elderly people is caused by an agedependent unhealthy change of the gut microbiota (GM) composition, resulting in reduced bone and muscle mass and thereby increased fracture risk. This hypothesis is based on previous findings Old Figure 1 Hypothesis: Increased fracture risk in elderly subjects is caused by an agedependent unhealthy change of the gut microbiota (GM) composition, resulting in reduced bone and muscle mass and thereby increased fracture risk. Young Healthy GM Unhealthy GM Muscle + Bone ↑ Muscle + Bone ↓ No Fracture Fracture 3 that ageing is associated with both reduced GM diversity and a decline in bone and muscle mass. Our vision is to first identify bacteria that are missing (or should not be present) in the GM of an elderly person with high fracture risk and then to treat that specific patient with tailored probiotics to reduce fracture risk via increased bone and muscle mass.

Methodology: Our hypothesis will in project 1 be tested using six unique, well-powered, thoroughly characterized, Nordic population-based cohorts with gut metagenome sequencing and information on incident fractures available in 2024. A major emphasis will be to determine causality of the GM using a variety of recently developed GM-related genetic instruments (for GM composition and function as well as for metabolites) as exposures in 2-sample Mendelian randomization (MR). In addition, as an alternative independent approach, we will use fecal microbiota transplantations (FMT) from humans to mice to determine causality of age-dependent GM alterations for reductions in bone and muscle mass. Based on this information, we will in project 2 design candidate probiotic treatments and test their efficacy in mouse osteoporosis models. Finally, we will in project 3 determine the clinical usefulness of the identified GM signatures for fracture prediction in elderly subjects.

Research progress beyond the state of the art: Previous studies on the association between GM signatures and outcomes in humans have in general not included the important evaluation of causality. The demonstration of causality is central for this proposal. We will employ three complementary independent methods, (i) 2-sample MR, (ii) FMT, and (iii) placebo-controlled treatment studies, to determine causality for human-derived GM on musculoskeletal parameters. For this research, we have established large well-characterized Nordic cohorts with metagenome sequence data. As the metagenome sequence data and information on incident fractures for these cohorts will be available in 2024, it is a unique timely opportunity to perform the proposed studies. One of these cohorts also has frozen fecal samples available from as many as 2,470 elderly women as a source to select from for the proposed FMT studies. If we are successful in identifying a GM-signature with a robust effect on agerelated bone and muscle loss, this finding would open-up completely new avenues to avoid fractures in elderly.

Background: Understanding trends in ADRD (Alzheimer’s disease and related dementias) incidence, prevalence, and associated risk factors, is critical to improving our understanding of the burden of the disease and for public health efforts and individual interventions focusing on hitherto identified and potentially modifiable lifestyle factors. Over the last several years, an increasing number of studies have suggested that the incidence of dementia has been decreasing over the past two decades. However, there are few epidemiological studies evaluating the role of birth cohort on cognitive impairment/trajectories later in life and there are some conflicting results. Moreover, alongside the description of temporal trends in cognitive trajectories, attempts have been made to understand which factors may underlie such changes, but findings are thus far inconclusive. The NEAR and SIMPLER infrastructures are ideally suited to collaboratively tackle questions regarding potential cohort shifts in risk factors, emphasizing modifiable risk factors, and changes in trends in cognitive function over the last several decades. While few single studies permit decomposition of age and cohort differences, integrative data analysis of the NEAR- and SIMPLER will provide a combined range of age and birth cohort samples required for this analysis.

Hypothesis: Differences in environmental exposures, lifestyle factors, and patterns in chronic disease across birth cohort might explain observed changes in patterns of cognitive aging. Cohort effects, defined by birth year, result from differences in exposures and develop through different historical periods. There is potential for interactions among age, cohort, and period effects (e.g., a period effect differentially affecting older individuals) and these too must be considered in understanding changing trends in cognitive dysfunction.

Method: For evaluation of cohort differences in age intercepts and trends, we have been fitting logistic models. However, the logistic function defines the functional form of aging, and not simply the link function between the model and the outcome (as in normal logistic regression). The model has three basic parameters: intercept, rate, and an inflection point/midpoint parameter. The functional form of the model takes the form where M is the midpoint parameter. The intercept is then further defined as (some form of): , where is the grand mean intercept, is an individuals random deviation from the intercept, is a set of cubic splines (and their associated parameters) modeling cohort shifts, and is a given study’s random deviation from the intercept. Study differences in age and birth cohort will produce overall incomplete data that can be considered missing completely at random (MCAR) and ignorable. However, there may be study-level incomplete data that may be considered missing at random (MAR). We will need to evaluate the potential sources of missingness and potentially include variables that are most strongly associated with missingness.

Research questions: Are there differences in educational attainment (e.g. better cognitive reserve), lifestyle and health behaviors (e.g. physical activity), socioeconomic conditions, and cardiovascular burden in later born generations relative to earlier birth cohorts? Do these cohort differences explain the cohort differences observed in cognitive trajectories and incidence of cognitive impairment? Do these risk factors interact within and across cohorts that might further explain cohort differences in cognitive change and impairment? Can we identify differences between women and men?

Background: Evidence is lacking regarding the association between added sugar consumption and risk of cardiovascular diseases.

Hypothesis: Our hypothesis is that there is an association between added sugar intake and increased cardiovascular disease risk. Further, we hypothesize that the associations may vary between the different cardiovascular diseases and sources of added sugar.

Aim: The aim is to investigate the association between intake of added sugar, and different sources of added sugar, and risk of cardiovascular diseases with varying etiologies.

Method: We will estimate intake of added sugar from participants in in the SMC and COSM. We will also study different sources of added sugar including sugar-sweetened beverages, treats and toppings. The associations between added sugar intake and seven different cardiovascular diseases (i.e. myocardial infarction, ischemic stroke, intracerebral hemorrhage, atrial fibrillation and aortic valve stenosis, heart failure, and abdominal aortic aneurysm) will be examined. The associations will be investigated in normal weight, overweight and obese separately.

Background: Sepsis is a condition of life-threatening organ dysfunction due to a dysregulated immune response to infection. Circa 80 000 individuals are affected annually in Sweden. Sepsis requiring intensive care is associated with high mortality: one third of patients die within 30 days, and close to every other patient dies within one year. The risk of sepsis and the risk of dying in sepsis increases with age as well as liver, kidney, and cardiovascular diseases. These factors cannot, however, be modified.

It is not known how modifiable risk factors, such as physical activity, smoking and alcohol consumption is associated with the risk of sepsis or the risk of dying in sepsis. Low levels of exercise (defined as physical activity causing perspiration) has been associated with higher risk of sepsis and death in sepsis, but previous studies have not been able to compare different intensities of physical activity or examine potential dose-response relations. Smoking increases the risk of pneumonia but previous studies have shown that smoking may, or may not, be associated with a lower risk of death in pneumonia. This paradoxical finding needs further investigation, as does the association between smoking and other infections. Alcohol abuse increases the risk of pneumonia, sepsis and death in sepsis, but it is not known whether low to moderate alcohol consumption is associated with the risk of sepsis.

Aims and Hypothesis: Sepsis is a dysregulated host response to infection, where the individuals’ immune response causes a life-threatening overreaction, but this does not occur in all people or in all infections. Previous studies have shown that lifestyle factors may alter the risk of sepsis, but have been too small to study these associations in detail. The aim of this project is to investigate how physical activity, smoking and alcohol consumption is associated with the risk of severe infection and sepsis; the risk of needing in-patient care and intensive care; the risk of needing organ support such as mechanical ventilation or renal replacement therapy; the risk of death during or after sepsis; and the risk of poor health and altered lifestyle after sepsis. Our hypothesis is that the physiological reserve and immune system is affected by lifestyle factors, which in turn affects the risk of severe infection and sepsis.

Specifically, we would like to investigate:

1. Is physical activity associated with severe infection and sepsis, need of in-patient or intensive care, organ support or death in severe infection and sepsis?
2. Is smoking associated with severe infection and sepsis, need of in-patient or intensive care or organ, support or and death in severe infection and sepsis?
3. Is alcohol consumption associated with severe infection and sepsis, need of in-patient or intensive care, organ support or and death in severe infection and sepsis?
4. Is severe infection and sepsis associated with changes in lifestyle or self-rated health after the illness, and does premorbid lifestyle affect these associations?
5. Is severe infection and sepsis associated with new-onset morbidity or mortality after the illness, and does premorbid lifestyle affect these associations?

Vetenskaplig frågeställning: Det övergripande syftet med projektet är att studera fetma och viktförändringar i samband med risk för sjukdom och mortalitet. Mer specifikt avser vi studera body mass index (BMI, kg/m2), midjemått och viktförändringar i samband med risk för:

1. Cancer totalt och specifika cancerformer, och mer detaljerat för prostatacancer
2. Kardiovaskulära sjukdomar (t.ex. hjärtinfarkt, stroke, hjärtsvikt, och blodpropp) och typ 2 diabetes
3. Sjukskrivningar, totalt och separat per diagnos (t.ex. cancer och depressiva-,
   muskuloskeletala-, cirkulatoriska- och mag-tarm-sjukdomar)
4. Covid-19, studerat som sjukhusinläggning, intensivvårdsbehandling och mortalitet
5. Mortalitet totalt och de vanligaste grupperingarna (t.ex. cancer och cirkulatoriska-, respiratoriska- och mag-tarm-sjukdomar samt externa orsaker)

Sambanden mellan fetma och viktförändringar och risk för sjukdom och mortalitet kommer att studeras för hela populationen sammantaget men vi kommer också att studera huruvida sambanden skiljer sig beroende på kön, sociodemografiska faktorer, rökning, samsjuklighet och medicinering (s.k. effektmodifiering av dessa faktorer). Ålders- och tids-trender i sambanden mellan fetma och utfall kommer också att studeras. Om den statistiska styrkan är tillräcklig kommer även rökning att studeras som riskfaktor i samband med covid-19 utfall. För prostatacancer kommer vi även att studera huruvida fetma är relaterat till felklassificering av T stadium innan vs. efter radikal prostatektomi (cT vs. pT), samt till positiva marginaler efter
radikal prostatektomi.

Ett sekundärt syfte är att studera orsaker till viktförändringar, exempelvis huruvida
sociodemografiska faktorer och negativa livshändelser såsom sjukskrivning, arbetslöshet och separation är relaterade till påföljande viktförändring. Vi kommer också att undersöka användbarheten av mediationsanalys för att bättre förstå vägarna till sjukdom och mortalitet med hjälp av insamlade data.

Background: We are working towards establishing a national collaborative infrastructure, the Swedish Cohort Consortium (Cohorts.se). In the pilot study (cycle 1), we identified risk factors for subarachnoid haemorrhage among nearly one million cohort participants in 21 cohorts. The outcome was chosen because it is comparatively rare and most cohorts would be too small to study this condition by themselves. We are now collecting data for cycle 2.

Aim: The primary objective is to establish risks factors for and protective factors against rare diseases, including variations in associations by sex and other subgroups. Secondary objectives include describing changes in population-attributable risk for the diseases over time with changing risk factor distributions.

Method: A major part of the project will be harmonization of exposures. Initially, distributional properties of all variables will be investigated and summarized using standard metrics. Missingness will be investigated and we will use Poisson regression models to estimate associations of exposures with risk of outcomes. Sensitivity to between-cohort bias will be investigated by excluding one cohort at a time. Causality of any findings will be investigated using established Mendelian Randomisation (MR) techniques and we will use genetic variants associated with the exposures to assess the causal effect of exposures on outcomes.

Background and medical significance: Sleep is known to be a crucial factor to maintain health. Sleep deprivation is associated with negative health outcomes, such as stroke, heart disease, obesity, depression, and increased risk of death. The National Sleep Foundation recommends 7-9 hours of sleep for adults and 7-8 hours for older individuals. However, epidemiological evidence suggest that a large proportion of the general population sleeps less than the recommended number of hours or suffer from sleep disorders.

Likewise, the global prevalence of insufficient physical activity (PA), is also high. This may contribute to, for instance, type 2 diabetes, heart-disease, cancer, depression, and increased risk of death. According to The Swedish national bord of health and welfare, adults are recommended 150 minutes of moderate PA per week, alternatively 75 minutes of high intensity exercise. In Sweden, approximately 67% of adults meet this criteria.

A bi-directional relationship between sleep and PA has been reported. Increasing evidence indicates that regular exercise may improve sleep. However, little is known about how different sleep characteristics affect PA.

Main objective/research question: The main objective of this project is to investigate the relationship between sleep and physical activity in Swedish middle aged and elderly women and men. We hypothesize that good sleep may be associated with higher physical activity or vice versa.

Material and method: Data will be obtained from The Swedish Infrastructure for Medical Population-Based Life-Course and Environmental Research (SIMPLER) that consists of the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM). The participants completed validated health and lifestyle questionnaires. Data processing and statistical analysis will be performed on a secure server BIANCA (UPPMAX). The analysis will be conducted using STATA version 15 (StataCorp, TX, USA). Statistical models (e.g., logistic regression) will be adjusted for several potential confounders such as smoking, alcohol and education.

Det övergripande syftet med detta projekt är att undersöka kost- och livsstilsfaktorers betydelse för risk och prognos vid prostatacancer, urinblåsecancer och njurcancer, för att bidra till en god grund för preventivt folkhälsoarbete, samt skapa förutsättnignar för livsstilsråd till patienter som drabbats av cancer. Mer specifikt vill vi i detta projekt undersöka om anti-inflammatorisk kost kan påverka risk och prognos för dessa sjukdomar, samt om upplevd psykologisk stress har betydelse för samma utfall.

Primära frågeställningar:

• Är antiinflammatorisk kost, mätt genom ett validerat index (AIDI), associerat med en minskad risk för prostatacancer, njurcancer eller urinblåsecancer?
• Är självrapporterad psykologisk stress associerad med en ökad risk för prostatacancer, njurcancer eller urinblåsecaner?

Sekundära frågeställningar:

• Är antiinflammatorisk kost, mätt genom ett validerat index (AIDI) associerat med en minskad risk för prostatacancer, njurcancer eller urinblåsecancer, uppdelat efter tumörstadium?
• Är antiinflammatorisk kost, mätt genom ett validerat index (AIDI) associerat med överlevnad hos personer som drabbats av prostatacancer, urinblåsecancer eller njurcancer?
• Är självrapporterad psykologisk stress associerad med en ökad risk för prostatacancer, njurcancer eller urinblåsecaner, uppdelat efter tumörstadium?
• Är självrapporterad psykologisk stress associerad med överlevnad hos personer som drabbats av prostatacancer, urinblåsecancer eller njurcancer?

Vi planerar att undersöka de ovan angivna frågeställningarna i en stor populationsbaserad kohortstudie med lång uppföljningstid och utfallsdata från nationella register.

Every year, millions of people throughout the world suffer from fragility fractures. Hip fractures are the most severe type of fractures and is associated with high burden of morbidity, healthcare costs and mortality. Worldwide about one in five women and one in ten men will suffer from a hip fracture in their lifetime. In the last decade the importance of a healthy diet for the prevention of fractures has been emphasized. Indeed, a healthy diet seems to be a key in preventing osteoporosis and many nutrients and bioactive components, including fatty acids, may play a role. We hypothesize that the estimated intake of specific dietary fatty acids will associate with both the odds of being osteoporotic as well as the risk of hip fractures. In this study we also aim to carry out a relative validation of the fatty acid intake from the FFQ by the use of biomarker fatty acid composition from adipose tissue in a sub sample of the cohort. Biomarker fatty acids in plasma and its different compartments (triglycerides (TG), phospholipids (PL), cholesteryl esters (CE)) as well as of adipose tissue and the erythrocyte and platelet membranes, are reflective of the dietary intake of different fatty acids.

Healthy dietary patterns, consisting of an abundance in plant based foods, are emphasized in current dietary recommendations and food based dietary guidelines because of beneficial effects on chronic disease risk including osteoporosis and fracture risk. Vegetarian diets, excluding all or some foods of animal origin are becoming more popular with the awareness of the environmental impact of foods of animal origin and especially meats of ruminant origin. Recent studies have indicated deleterious effect of vegetarian diets, especially vegan, compared to omnivore diets on fracture risk, but so far there have been few studies on this topic. Overall we aim to investigate the association between adherence to three different dietary patterns, lacto-ovo vegetarian, vegan and modified Mediterranean, and incident hip fracture risk. We will also be able to investigate the effect modification of supplement use of calcium and vitamin D. The analyses will be carried out in the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM) using data from the investigations 1987-90 (SMC) and 1997, 2009 and 2019.

Millions of people throughout the world suffer from fragility fractures. Hip fractures are the most severe type of fractures and is associated with high burden of morbidity, healthcare costs and mortality. Worldwide about one in five women and one in ten men will suffer from a hip fracture in their lifetime. Adequate intakes of calcium and vitamin D are vital for the continuous regeneration of the skeleton in adult life and is promoted for the prevention of osteoporosis and fractures, but at what level of intake has long been debated. Recent systematic reviews and meta-analyses have concluded that neither dietary nor short-term supplemental calcium or vitamin D (alone or in combination) help in the primary prevention of fractures in a general nutrient replete population. In the last decade the importance of a healthy diet also for the prevention of fractures, and not only for the prevention of cardiovascular disease, diabetes and cancer, has been emphasized. Indeed, a healthy diet seems to be a key in preventing osteoporosis and many nutrients and bioactive components may play a role. We hypothesize that a healthy diet, defined as a Mediterranean type diet, herein defined as the modified Mediterranean diet (mMED) score, potentially and partially compensate for a low calcium intake. Thus, both a high intake of calcium and healthy diet would be optimal for the prevention of hip fractures. In the present analysis we therefore extend previous analyses by examining combinations of a low, moderate and high calcium intake and mMED categories in relation to hip fracture in a longitudinal cohort study of men and women. The study population consists of participants from the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM) in Sweden, which are part of the research infrastructure SIMPLER. SMC was established in 1987-1990, while COSM was established in late 1997, which will be our baseline for the present analysis with time-updated information of our exposures and covariates. Information on incident hip fractures will be collected from the National Patient Registry and information on deaths will be collectedfrom the Cause of Death Register and time-to-event analysis will be employed.

Background: The composition of macronutrients in the diet can affect the risk of cardiometabolic diseases. Macronutrient quality is also important, with high dietary carbohydrates quality associated with lower cardiometabolic risk. Metabolic profiling studies and multi-omics profiling in prospective cohorts may help identify optimal macronutrient compositions and biomarkers for predicting long-term health impacts.

Aims and hypothesis: We hypothesize that multi-metabolite signatures (MMS) reflect the metabolic adaptations to dietary macronutrient composition and quality and that these MMS are associated with CMD risk and mortality. We hypothesize that high-quality carbohydrate diets are related to lower CMD risk. Moreover, we hypothesize that the dietmetabolite-disease associations are partially mediated and modified by the gut microbiome’s composition and functional capacity and the host’s genome and proteome.

Aim 1: Derive MMS of dietary MNC and MNQ and examine their association with the risk of incident type 2 diabetes, cardiovascular diseases (myocardial infarction and stroke), and cause-specific mortality.

1. Derive dieteary indices that reflect the macronutrient quality (e.g., GI, GL, CQI) and both quality and composition (e.g., GL, CQI) of the self-reported habitual diet and use machine learning from untargeted metabolomics data to build MMS that predict dietary index adherence. Hypothesis: MMS of beneficial dietary indices are enriched with UFAcontaining lipids and secondary plant metabolites.
2. Associate dietary indices and MMSs of dietary MNC and MNQ with CMD risk (diabetes, coronary heart disease, stroke, and cause-specific mortality) and asses to what extent MMSs mediate potential associations of the dietary indeces with disease 3 incidence. Hypothesis: MMS of beneficial dietary indices (e.g., CQI) are related to low CMD risk, whereas MMS of unfavorable dietary indeces (e.g., GL) are related to higher risk.

Aim 2: Examine if the composition and funcitional capacity of the gut microbiome mediate or modify the potential effect of dietary MNC and MNQ with prospective CMD risk. We hypothesize that the relationship between dietary MNC and MNQ with cardiometabolic health depend on the gut microbiome’s functional capacity.

Method: We will request data on diet, lifestyle, and clinical history from SMC and COSM participants (Aim 1). Blood biochemistry, metabolomics, and microbiota from clinical subcohorts will also be collected (Aim 2). We will derive indices that reflect macronutrient quality and composition (e.g., GI, GL, CQUI, GL) and categorize the study population based on deciles of relative energy from source-specific nutrients. Elastic net regularization will be used to identify optimal metabolites to predict adherence to each dietary index (DI) and construct weighted MMS of DI adherence. We will evaluate the association of the MMS with corresponding DI and replicate the results in independent cohorts. We will use Cox proportional hazards regression to associate DI-MMS with the incidence of T2D, CAD, stroke, and mortality, and test for effect modification by the gut microbiome. Covariables will be adjusted in all analyses. The statistical analyses will be conducted in R and Python.

abstract

Diets link human and environmental health. Unhealthy diets cause most global disabilityadjusted life-years of all risk factors. Today global food production and food supply chain accounts for almost one third of total greenhouse gas emissions (GHGEs) of which agriculture alone contributes 24%. Moreover, food production accounts for the largest share of use of land and water and causes the majority of eutrophying and acidifying emissions.

The purpose of the present interdisciplinary research project is to build new integrated knowledge in the combined field of foods’ impact on human health and on environmental health and sustainability, simultaneously addressing societal challenges and global sustainability target. The proposed project will contribute to vital knowledge to support future evidence-based health-promoting sustainable food consumption and production. It will be valuable for policymakers, authorities, public food sector and food producers, and directly for consumers and for public health sector.

The collaborating team consists of nutritional and environmental epidemiologists, and toxicologists (from Karolinska Institutet, Alicja Wolk (PI), etc.) as well as experts in modelling environmental impact from the food chain (from RISE- Gothenburg, Ulf Sonesson etc.). The investigation of quantitative environmental data for different ingredients will be possible because of preparation of a unique climate and environmental food database based on more than 20 years of experience in life cycle assessment (LCA) of food products. The project relies on dietary data from two population-based prospective cohorts, consisting of over 100 000 men and women from central Sweden. Specifically, we will

1. Based on our previous results on what defines a healthy and unhealthy diet regarding mortality, CVDs and diabetes, characterize the broad environmental impact (GHGEs, cropland and water use, eco-toxicity of pesticides) and human toxicity (exposure to food contaminants e.g. toxic metals, PCBs, pesticide residues) of such diets. This health and environmental impact will also be assessed by using a novel statistical method for food substitution approach to identify alternative dietary patterns with the highest health promoting effect.
2. Identify in the study population dietary patterns with the lowest vs. highest impact on environment and evaluate their impact on health outcomes (e.g. all-cause and cause specific mortality). And then based on the statistical modeling define the most optimal diet with regard to both health and the environment in the studied cohorts.
3. Finally, we will examine whether such climate-adapted improved diets are directly associated with the health status of individuals as assessed by measurement of inflammatory biomarkers in blood in a subgroup of the participants.

The results of the project will be communicated in scientific journals, as well as in popular scientific communication targeting stakeholders from food producers, retail, authorities (Environmental Protection Agency, the Board of Agriculture), public food actors, public health actors and public health authorities and academia, as well as the general public.

As the first “health outcome” we will start with analyses of heart failure and as the first dietary exposure studied, we will use a dietary pattern defined as DASH diet and its individual food components and relevant food substitutions. We will use standard Cox models (Stata) and causal inference models (statistical software R).

abstract

abstract

abstract

Diets link human and environmental health. Unhealthy diets cause most global disabilityadjusted life-years of all risk factors. Today global food production and food supply chain accounts for almost one third of total greenhouse gas emissions (GHGEs) of which agriculture alone contributes 24%. Moreover, food production accounts for the largest share of use of land and water and causes the majority of eutrophying and acidifying emissions.

The purpose of the present interdisciplinary research project is to build new integrated knowledge in the combined field of foods’ impact on human health and on environmental health and sustainability, simultaneously addressing societal challenges and global sustainability target. The proposed project will contribute to vital knowledge to support future evidence-based health-promoting sustainable food consumption and production. It will be valuable for policymakers, authorities, public food sector and food producers, and directly for consumers and for public health sector.

The collaborating team consists of nutritional and environmental epidemiologists, and toxicologists (from Karolinska Institutet, Alicja Wolk (PI), etc.) as well as experts in modelling environmental impact from the food chain (from RISE- Gothenburg, Ulf Sonesson etc.). The investigation of quantitative environmental data for different ingredients will be possible because of preparation of a unique climate and environmental food database based on more than 20 years of experience in life cycle assessment (LCA) of food products. The project relies on dietary data from two population-based prospective cohorts, consisting of over 100 000 men and women from central Sweden.
Specifically, we will

A) Based on our previous results on what defines a healthy and unhealthy diet regarding mortality, CVDs and diabetes, characterize the broad environmental impact (GHGEs, cropland and water use, eco-toxicity of pesticides) and human toxicity (exposure to food contaminants e.g. toxic metals, PCBs, pesticide residues) of such diets. This health and environmental impact will also be assessed by using a novel statistical method for food substitution approach to identify alternative dietary patterns with the highest health promoting effect.

B) Identify in the study population dietary patterns with the lowest vs. highest impact on environment and evaluate their impact on health outcomes (e.g. all-cause and cause specific mortality). And then based on the statistical modeling define the most optimal diet with regard to both health and the environment in the studied cohorts.

C) Finally, we will examine whether such climate-adapted improved diets are directly associated with the health status of individuals as assessed by measurement of inflammatory biomarkers in blood in a subgroup of the participants.

The results of the project will be communicated in scientific journals, as well as in popular scientific communication targeting stakeholders from food producers, retail, authorities (Environmental Protection Agency, the Board of Agriculture), public food actors, public health actors and public health authorities and academia, as well as the general public.

As the first “health outcome” we will start with analyses of heart failure and as the first dietary exposure studied, we will use a dietary pattern defined as DASH diet and its individual food components and relevant food substitutions. We will use standard Cox models (Stata) and causal inference models (statistical software R).

Background: Although the underlying pathophysiology of chronic diseases is often complex and multifactorial, low-grade inflammation has proven to play a role in their development. Furthermore, it has also been proposed that diet can be one of driving factors behind this chronic low-grade inflammation. Inflammatory biomarkers are a way to access the link between inflammation and diet, and in a recent systematic review Aleksandrova et al reviewed their use in nutrition research. C-reactive protein (CRP) is among the most commonly used, but a wide range of biomarkers are currently being studied and applied in the field.

To date, the only dietary inflammatory index to our knowledge developed for a Nordic population is the Anti-Inflammatory Dietary Index (AIDI). It was constructed using the results from a 123-iten food frequency questionnaire in a population of 3502 Swedish women aged 56-74 from the Swedish Mammography Cohort with high sensitivity CRP (hsCRP) <20 mg/L. Twenty food groups were statistically significantly correlated with hsCRP, meaning that the AIDI could predict systematic low-grade inflammation. However, it remains to further develop and validate such index in other populations and to test it against other inflammation-related biomarkers to confirm if it is a useful tool assessing inflammatory potential of the diet.

Hypothesis: We hypothesize that the previously developed AIDI is a valid tool to predict low-grade systemic inflammation in a cohort of middle-aged and elderly men using hsCRP and other low-grade systemic inflammation-related biomarkers.

Aim: The aim of this project is to further develop the AIDI as a tool to assess anti-inflammatory potential of diet by using not only hsCRP but also other low-grade systemic inflammation-related biomarkers (as specified on the attached list of variables).

Research questions:

1. Which specific foods and food groups are significantly correlated with low-grade systemic inflammation measured by hsCRP in middle-aged and elderly men?
2. How strong is the association of the newly developed AIDI, based on the total sum of the above identified foods/food groups, with low-grade inflammation measured by hsCRP and several other selected biomarkers of inflammation available in the COSM-Clinical dataset.

Method: The COSM was established in 1997 and consists of 45 906 Swedish men from Västmanland and Örebro counties. During 2010-2017 a sub-cohort of 4754 men from the COSM participated in a clinical health exam, filled in the same FFQ that was used to develop the AIDI, donated blood and other biological samples for research, etcetera. In the present project we will use data from this COSM-subcohort.

Background: The only established risk factors for amyotrophic lateral sclerosis (ALS) to date are older age, male sex, and a family history of ALS, but a handful of other environmental risk factors have been examined in previous epidemiological studies, although often with conflicting results. Further studies, especially based on prospective cohorts are warranted.

Hypothesis: Taking advantage of several unique longitudinal cohort studies and unique Patient Register in Sweden, the overarching aim of this proposal is to significantly improve our understanding of environmental risk factors for ALS. Specifically, we aim to assess the associations between the following factors and the development of ALS: Smoking and snuff dipping; Diet, in particular nutrients; BMI and diabetes; Physical activity; NSAIDs; Sleeping duration.

Aim: Our proposed project will hopefully help to better understand the environmental risk factors for ALS, which might subsequently contribute to the development of primary prevention measures against this dreadful disease.

Method: Cox regression models will be used to assess the associations between various risk factors and ALS risk. Potential interactions between different factors, e.g., smoking and BMI, will also be assessed.

Research questions: Taking advantage of several unique longitudinal cohort studies and unique Patient Register in Sweden, the overarching aim of this proposal is to significantly improve our understanding of environmental risk factors for ALS. Specifically, we aim to assess the associations between the following factors and the development of ALS:

1. Smoking and snuff dipping;
2. Diet, in particular nutrients;
3. BMI and diabetes;
4. Physical activity;
5. NSAIDs.

Background: Chlorination is a widely used method for disinfection of drinking water and although effective in reducing microbial growth it may introduce undesired reactive chemicals into the water. Several by-products are considered genotoxic and carcinogenic, something that has been clearly verified by animal studies. Calcium and Magnesium are minerals naturally present in drinking water. The level of evidence of an effect of both minerals on the development of cardiovascular disease is high, however, the relevance of the concentrations in drinking water is not known.

Hypothesis: Chlorination by-products pre-dispose individuals to develop bladder and colorectal cancer. Calcium increase risk of cardiovascular events whereas magnesium has a protective effect on disease risk.

Aim: To assess whether chlorination of drinking water predispose individuals to develop bladder or colorectal cancer and to assess if calcium and magnesium affects the risk of cardiovascular disease.

Method: The study population will be the male and female participants of SMC and COSM. Retrospective information on concentrations of trihalomethanes and Mg and Ca in municipal drinking water will be collected from Vattentäktsarkivet and the Swedish Water and Waste water Association, together with historical information on the drinking water production methods used. Localities within the SMC and COSM counties will be separated into categories as follows: high, medium, low or no chlorination and high or low calcium/magnesium. Cancer and cardiovascular outcomes will be ascertained via the linkage to national patient registry. Associations between exposures and outcomes will be assessed by fitting cox proportional hazard regression models, adjusting for potential covariates.

Cardiovascular disease (CVD) is a major public health challenge for the 21st century because of its impact on personal and population health, premature mortality and high societal costs. Identifying novel CVD-risk factors, especially prevalent ones, is therefore important to allow further preventive actions. Polychlorinated (PCBs), other organochlorines and perfluoroalkyl (PFAS) substances, widespread and very persistent environmental pollutants, may be involved in various CVD-related pathophysiological processes. The overall aim is to prospectively assess the association between biomarkers of PCBs and PFAS exposure and first incident myocardial infarction and/or ischemic stroke in two population-based cohorts with biobanks, assuring adequate statistical power and cost-effective handling of expensive chemical analyses. The outcome of this project will fill an important knowledge gap. Because all are exposed to these substances mainly via food – especially fish – a better knowledge of the risks and benefits of fish consumption is important for public health, as the current uncertainty presents a dilemma to consumers and advisory boards.

Research questions: What are the associations between biomarkers of PCBs, other organochlorines and PFAS and the risk of incident myocardial infarction (MI) and/or ischemic stroke in women and men (700 case/control pairs)?

1. To what extent are biomarkers of essential fatty acids, especially the marine omega-3, vitamin D and fish consumption, modifying the associations?
2. Are baseline cross-sectional associations between PCBs and major intermediate risk factors/markers for CVD in line with previous and current findings and informative in providing further mechanistic insight
3. Validate/replicate the previously observed PFAS associated metabolites (metabolomics) in different a population, and assess associations between these metabolites and major intermediate risk markers for CVD and risk of MI and stroke.

abstract

abstract

Planned Studies:

1. Environmental exposures in a cohort of Swedish women – Are there inequalities due to socioeconomic determinants?

In the first project, we will assess the distribution of urban environmental exposures, including air pollution, traffic noise and greenness, in the SMC. In particular, we will assess inequalities in exposure depending on socioeconomic determinants of our study participants.

2. Air pollution and traffic noise as risk factors for myocardial infarction in light of lifestyle, occupational exposure and socioeconomic determinants.

In the second project, will assess the separate and combined associations between air pollution and traffic noise and incidence of MI, taking into account lifestyle, occupational exposure to air pollution and noise and socioeconomic determinants.

3. Air pollution and traffic noise as risk factors for stroke in light of lifestyle, occupational exposure and socioeconomic determinants

In the third project, will assess the separate and combined associations between air pollution and traffic noise and incidence of stroke, taking into account lifestyle, occupational exposure to air pollution and noise and socioeconomic determinants.

4. Can residential greenness reduce the risk of myocardial infarction, taking into account air pollution, traffic noise, lifestyle, occupational exposure and socioeconomic determinants?

In the final project, we will assess the associations between residential greenness and incidence of myocardial infarction. In addition, we will study the interaction between all three environmental exposures and how they contribute to or protect from CVDs in light of other important risk factors, including lifestyle, occupational exposure to air pollution and noise and socioeconomic determinants.

Background:Most studies assessing the associations between environmental, occupational or lifestyle factors and cardiovascular diseases (CVD) are limited to single exposures and rarely considered a more complex exposure panorama. Consequently, there is limited knowledge on their separate and combined contribution to health and disease. This PhD-project is designated to bringing all this information on important exposures together for an assessment of the so called exposome.

Aim: The overall aim of this project is to prospectively assess the impact on CVD by air pollution, traffic noise and greenness, and to explore the interaction between these environmental factors and diet and lifestyle, occupational exposure to air pollution and noise, and by socioeconomic status.

Research questions:

• Are there inequalities in the exposure to air pollution, traffic noise and greenness depending on socioeconomic determinants?
• Does exposure to air pollution and traffic noise increase the risk of MI and stroke?
• Does exposure to residential greenness reduce the risk of MI and stroke?
• To what extent do the environmental exposures contribute to CVDs considering other risk factors, including diet and lifestyle, occupational exposure to air pollution and noise and socioeconomic determinants?

Method: Participants’ residential address history (Uppsala county) will be assigned a geographic coordinate. Yearly averaged concentrations (1990-2015) of air pollutants related to traffic, heating and shipping, including NOx, PM10, PM2,5, and Black Carbon, is modeled using a Gaussian dispersion modeling technique, using the Nordic Prediction Method. Greenness exposure is based on satellite images and the Normalized Difference Vegetation Index (NDVI). Socio-economic factors as well as exposure to occupational noise and air pollution will be based on occupational codes (SSYK and NYK) from SCB. The codes will be linked to the exposure using Job Exposure Matrices for air pollution and noise, respectively. Data will be complemented by questionnaire

Background: Exposure to noise in urban environment is unavoidable, and because it has been repeatedly associated with increased risk of several major chronic diseases, including cardiovascular disease, the metabolic syndrome and cancer, it is a rapidly growing public health concern. Yet, only few studies have assessed the association of the joint exposure to noise and air pollution, with these diseases, or explored interactions with other environmental factors such as urban greenness.

Hypothesis: The main aim of NordSOUND is to comprehensively assess the joint effect of environmental and occupational noise in the development of several diseases and conditions of great public health concern. More specifically, we aim to investigate the following hypotheses:

• Occupational and residential transportation noise increase the risk of cardiovascular disease, obesity, and cancer
• Residential transportation noise results in a high burden of disease
• There are socio-economic inequalities in the exposure to transportation noise and in the relationship between noise and health

Aim: The overall aim of this project is to prospectively assess the association of air pollution, traffic noise and greenness with cancer in breast and colon, CVD and BMI and survival after cancer diagnosis

Method: Residential address history (Uppsala county only) of each study participant will be assigned a geographic coordinate. Yearly averaged concentrations (1990-2015) of air pollutants related to traffic, heating and shipping, including NOx, PM10, PM2,5, and Black Carbon, is modeled using a Gaussian dispersion modeling technique, using the Nordic Prediction Method. Greenness exposure is based on satellite images and the Normalized Difference Vegetation Index (NDVI). Socio-economic factors as well as exposure to occupational noise and air pollution will be based on occupational codes (SSYK and NYK) from SCB. The codes will be linked to the exposure using Job Exposure Matrices for air pollution and noise, respectively. Data will be complemented by questionnaire

abstract

abstract

Background: Persistent Organic Pollutants (POPs), including polychlorobiphenyls (PCBs), dioxins, perfluorinated and polyfluorinated alkyl substances (PFAS), organochlorine pesticides and PAHs, are a large group of chemicals regulated by the Stockholm Convention. Several POPs (dioxins, PCBs, PAHs) are carcinogenic to humans according to the International Agency for Cancer Research and some of them exhibit endocrinal disruptor properties and various toxicity effects. Several studies have reported associations between POPs and breast cancer. Metabolomics is the analysis of all the small molecules (known as metabolites) within cells, biofluid or organisms. An aim of this research field is to depict the molecular events associated with biological processes such as pollutants toxicity. Metabolites associated to POPs exposure can be used to gain insight in the relationship between POPs and breast cancer, as well as shed light on the molecular pathways underlying breast cancer development affected by POPs.

Hypothesis: This is an exploratory study to gain insight in the molecular pathways affected by POPs underlying breast cancer development and investigate POP-related metabolites and breast cancer risk. We hypothesize that some of the POP-related metabolites will also be associated with breast cancer risk.

Aim: The aim of the project is to investigate associations between metabolites related to biomarkers POP exposure and breast cancer risk in the Swedish Mammography Cohort clinical subcohort.

1. Metabolites associated with POP exposure will be identified with untargeted LC-MS metabolomics using an in-house developed random forest algorithm
2. Associations between selected POP-related metabolites and breast cancer will be assessed using Cox regression.

POPs and PFAS levels have already been measured in a cardiovascular disease (i.e. myocardial infarction and stroke) case-control study nested within the SMC-C cohort. Metabolomics data is available in the SMC-C cohort. Objective 1 (to identify POP-related metabolites) is currently being performed within the cardiovascular disease nested case-control study in around 700 women for which the data is already available. For objective 2 (to investigate associations between selected POP-related metabolites and breast cancer), we will use the selected POP-related metabolite features for the full SMC-C cohort (data already available), but we would need additional data on breast cancer and data on potential confounders for the full SMC-C cohort.

Method: Associations between selected POP-related metabolites and incident breast cancer will be assessed using Cox regression.

Background: Nitrite is an anion that is ubiquitous found in the environment and is a natural component of our diet. The highest concentration in food is found in some green leafy vegetables and root vegetables, including spinach, lettuce and beetroots. Nitrate is also present in drinking water, and elevated concentrations can be found in agricultural areas, often due to use of nitrate containing fertilizer at agricultural fields that contaminate nearby surface and groundwater resources. With the aim of protecting against infant methemoglobinemia, nitrate is currently regulated in drinking water and the maximum concentration is set at 50 mg/L (European Drinking water directive 2020/2184). Corresponding recommendations are also provided for dietary nitrate, with the acceptable daily intake being currently 3.7 mg/kg/day.

In addition to methemoglobinemia, nitrate exposure has been proposed to be associated with several chronic disease outcomes, including cardiovascular disease and cancer, both in terms of increasing and reducing risk. With respect to cardiovascular disease, nitrate from diet can through the enterosalivary nitrate-nitrite-NO pathway be converted to nitric oxide, a potent vasodilator that plays a critical role in blood pressure regulation. In agreement with this, intervention studies have shown that increasing dietary nitrate intake reduces blood pressure and improves endothelial function in adults. The epidemiological evidence on the link with hard cardiovascular disease endpoint is, however, limited.

Although possibly beneficial for cardiovascular disease, there are some concerns that nitrate may have a carcinogenic potential. The International Agency for Research on Cancer (IARC) has classified nitrate as a probable human carcinogen (2A), with the rationale that increasing nitrate exposure increases endogenous formation of N-nitroso compounds, some of which are known potent carcinogens. This has been proposed to particularly occur in the absence of inhibitors of nitrosation such antioxidants. Given that many vegetables contain vitamin C, polyphenols or other possible nitrosation inhibitors, nitrate from these sources may be less relevant for cancer than nitrate in the drinking-water. Drinking water nitrate has been studied in relation to occurrence of many different cancers, including cancer of the stomach, overall gastrointestinal tract, urinary tract and brain, but only few cohort studies are available for any given cancer site and in their most recent evaluation, IARC judged the overall evidence for nitrate in drinking water and human cancer to be inadequate evidence.

Hypothesis: We hypothesize that nitrate in diet and drinking water is inversely associated with incidence of major cardiovascular events. In addition, we would like to test if nitrate from drinking water alone, and in combination with nitrate from diet is associated with increased risk of certain cancers.

Method: Study population and exposure assessment For this project, we plan to use data from the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM), applying a prospective cohort design and starting follow-up in 1997.

The project will partially build on an existing SIMPLER project (SIMPLER project no: 2019-15, Bianca project no: simp2019019), in which we have previously assessed the association of a number of different drinking water parameters (disinfection by-products, minerals, fluoride) with cancer, cardiovascular disease and bone health. In this project we have generated a high-quality database on a large number of drinking water related exposures, including nitrate, that contains decades of exposure information, and linked this to historical and present information on participants residences that we have gathered from Tätortsregistret and Registret över totalbefolkningen at Statistics Sweden. This information is available for 52 drinking water utilities, supplying 61 localities in the SIMPLER study area where approximately 58,000 study participants live. Hence, in the drinking water studies on nitrate, we will have yearly information on the nitrate concentration in drinking water at the participants’ home, during 1960- 2018. We plan to separate SIMPLER participants exposure categories according to the nitrate concentration in drinking water at their home at baseline.

To assess dietary nitrate exposure, we will apply a similar methodology as in our previous study on fluoride in diet and drinking water in relation to bone health 7. In brief we have built a database on nitrate content in content of common foods and beverages, and plan combine this information with the intake frequencies of individual foods and beverages that are derived from the food frequency questionnaire (FFQ) answered in 1997. The nitrate content of meals and beverages prepared with tap water will be weighted according to the concentration in tap water at the participants homes.

Aim: The overarching aim of the proposed project is to assess the association of exogenous exposure to nitrate from diet and drinking water in relation to incidence of cardiovascular disease and cancer in middle-aged to elderly men and women.

Background and aim: To promote health via safer food and improved nutrition, particular attention is given to processing contaminants in food. However, there is limited knowledge regarding the potential health risks associated with these contaminants. Among them, acrylamide and glycidol, neo-formed high temperature food-processing substances, abundantly present in many commonly consumed food, are of interest due to their high exposure level in the general population and the increasing evidence of toxicity in animals.

Acrylamide forms in foods rich in starch when cooked/processed at temperature >120°C (e.g., roasting, frying, grilling) in absence of moisture. Acrylamide may also form in foods from acrolein during the degradation of amino acids, carbohydrates, lipids, and organic acids. The most common food sources of acrylamide are coffee, biscuits, breakfast cereals, bread especially crispy bread, potato products and vegetables crisps). Acrylamide is also produced in heated tobacco. Acrylamide is classified as probably carcinogenic to humans (class 2A): strong evidence in animals but not enough in humans. Moreover, mechanisms that mediate carcinogenesis in animals also operate in humans. There is substantial evidence showing that the genotoxicity of acrylamide is mediated by glycidamide, a highly reactive compound and the main metabolite of acrylamide, via the formation of DNA adducts. Also, epigenetic (non-genotoxic) mode of actions, involving production of oxidative stress and formation of adducts with protein (especially alkylation), contribute to acrylamide carcinogenicity. Furthermore, these molecular initiating events are postulated to trigger cumulative neurotoxicity and cardiometabolic alterations Most epidemiological studies, mainly based on imprecise measures of dietary exposure i.e., dietary questionnaires, have mainly failed to show an association with increased risk of cancer in humans so far. Exception is for a weak association found for ovarian and endometrial cancer in relation to higher dietary exposure assessed via food only in non-smokers women and for breast cancer in relation to the hemoglobin adducts only in premenopausal women. Recently, a study conducted in the US population (National Health and Nutrition Examination Survey, NHANES) found an association between high levels of acrylamide hemoglobin adducts in relation cancer mortality (n=118). Since the content of acrylamide in the diet may largely vary across countries, the results need to be generalized in non-US populations. Although acrylamide central and peripheral neurotoxicity is well-documented in animal and occupational studies, no epidemiological studies have investigated whether dietary acrylamide exposure is associated with neurodegenerative diseases. Growing body of evidence also indicate that acrylamide's mode of action might lead to cardiometabolic alteration which in turn, play a critical role in atherosclerosis and cardiovascular diseases (CVD) development. However, epidemiological evidence investigating dietary acrylamide and incidence of CVD are lacking.

Glycidol is another food processing contaminant and probable human carcinogen (IARC group 2A), that is much less studied than acrylamide and glycidamide. It is primarily found in refined edible vegetable oils where it is formed during the deodorization process. While palm oil, which owing to its functional benefits is frequently occurring in a multitude of common consumer food products, is particularly susceptible to glycidol formation and has the highest concentrations, no or only mere trace amounts are found in unrefined oils like virgin olive oil. The carcinogenicity of glycidol has been evaluated in a 2-year gavage study of rodents by the National Toxicology Program (NTP), where it induced dose related increases in incidences of neoplasms in several tissues, including the mammary glands and uterus, in rats and mice. In that same study, glycidol was also evaluated for genetic toxicology, and found to be mutagenic in Salmonella typhimurium, gave positive responses for trifluorothymidin induction already at very low doses is mouse lymphoma cells, induced sister chromatid exchanges and chromosomal aberrations in Chinese hamster ovary cells, and micronucleus in erythrocytes in bone marrow of mice. Although these results provide clear evidence for genotoxicity and carcinogenicity in animals, the association of glycidol exposure and risk of cancer in humans is not well understood, as studies on this are entirely lacking.

The importance of using objective measures of dietary intake in epidemiological studies to improve accuracy and limit biases and acknowledged difficulties related to the estimation of dietary exposures from self-reported questionnaire and food databases has been stressed. Biomarkers of acrylamide and glycidol exposure have been identified and validated in blood and in urines. The blood biomarkers i.e., blood hemoglobin adducts are considered more suitable biomarkers of exposure, as they may reflect long term of exposure (previous four months) and are not influenced by daily fluctuations.

Finally, screening for potential acrylamide-related metabolomics and proteomics may contribute, in epidemiological studies, to strengthen the possible exposure-health outcome association and understand its mode of action. To our knowledge, little effort has been made in this direction in this field. In unpublished data based on the EUROMIX (n=120 healthy subject), an association was found between hemoglobin adducts and 4 proteins (JAM-B, CNTN5, CLEC10A and EPHB6) related to neurological alteration.

Hence, we aim to investigate whether acrylamide and glycidol, measured through validated biomarkers, increase the risk of diseases related to its toxicity i.e., cancers, neurodegenerative disorders, and (CVD) in two population-based Swedish cohorts, The Cohort of 60-year-olds (60YO) and The Swedish Mammography Cohort (SMC), employing a case-cohort design.

In specific, we will assess:

the relationship between dietary acrylamide exposure, (i.e., acrylamide and glycidamide hemoglobin adducts) and the risk of:

1. cancers of colorectum (n=280), breast (n=340) and endometrium (n=100)
2. dementia including Alzheimer´s disease (n=700 cases)
3. Parkinson´s disease (n =190 cases)
4. myocardial infarction (n= 390 cases) and ischemic stroke
5. the relationship between biomarkers of glycidol exposure (DiHOPrVal) and colorectal, breast and endometrial cancer

In addition, exposure-affected OMICS’ signatures (mainly proteins and metabolomics) will be used to explore molecular pathways potentially underlying diseases development and to support biological plausibility. Acrylamide and glycidol- related metabolic and proteins biomarkers will be assessed in relation to risk of the diseases under investigation (only SMC) - Aim 6