Search

Additional file 1. Flow chart of the Study Population selection.

Additional file 6. Sensitivity analyses for the prospective association between vitamin K status and echocardiographic measurements

Additional file 2. Baseline characteristics of 427 participants stratified by quartiles of vitamin K intake.

Additional file 3. Baseline characteristics of participants included and excluded for this study

The research described in this thesis addresses the association of several traditional and novel cardiovascular risk factors and lifestyle behaviours with cardiometabolic diseases. This was further specified into two research objectives: I. What are potential mechanisms of the relationship between the modifiable lifestyle factor alcohol consumption and cardiometabolic diseases? II. What is the causal role of modifiable cardiovascular and lifestyle-related risk factors in the development of cardiometabolic diseases? Potential mechanisms of the relation between alcohol consumption and cardiometabolic diseases In Chapters 2 and 3, cardiac function and structure, insulin sensitivity and adiposity were investigated as potential underlying mechanisms in the relation of alcohol consumption with cardiometabolic diseases. Chapter 2 showed that higher levels of alcohol consumption were linearly associated with a decrease in left ventricular ejection fraction, a measure of systolic function, in a prospective study of 404 participants. No association was observed between alcohol consumption and left atrial volume index, which is a measure of diastolic function, or left ventricular mass index, a measure of cardiac structure. These findings suggest that alcohol consumption might play a role in the pathophysiology of heart failure with reduced ejection fraction. Chapter 3 showed that heavy drinking was associated with higher liver fat in a cross-sectional study on 787 men with prediabetes, and central adiposity seemed to explain part of this association. Moderate amounts of alcohol consumption appeared to be associated with a decrease in liver fat in the prospective analysis, but the definitive shape of this association warrants investigation in larger and more long-term follow-up studies. Altogether, heavy alcohol consumption is harmful for preclinical aspects of cardiometabolic health, but whether there is a safe drinking threshold for moderate amounts of alcohol consumption remains a point of discussion. The causal role of modifiable cardiovascular and lifestyle-related risk factors in the development of cardiometabolic diseases In Chapter 4, a systematic review of the current evidence from Mendelian randomization (MR) studies on the association between alcohol consumption and cardiometabolic diseases, mortality and cardiovascular risk factors was conducted. There was large heterogeneity in the methodological quality of the MR studies on this topic so far. It was not possible to draw conclusions on the causal role of moderate alcohol consumption on cardiometabolic health. Recent developments in MR reporting guidelines and methodology including instrument selection and non-linearity are expected to improve evidence from future MR studies. In Chapters 5, 6 and 7, the two-sample MR method was used to investigate the potential causal association of several cardiovascular risk factors and lifestyle behaviours with risk of hypertension, heart failure and longevity, respectively. These MR studies confirmed the causal role of several traditional risk factors in the development of cardiometabolic diseases, with for example genetically predicted smoking initiation being detrimentally associated with hypertension, heart failure and longevity, and genetically predicted higher body mass index being harmful for hypertension and longevity. A genetically predicted higher educational level was associated with a reduced risk of hypertension and increased life expectancy. The MR studies implied that sleep might be an important novel risk factor for cardiometabolic health, with a genetically predicted longer sleep duration being associated with a lower heart failure risk en a potentially lower hypertension risk, and genetically predicted insomnia increasing the risk of hypertension and possibly reducing life expectancy. For lifestyle factors including physical activity, sedentary behaviour and coffee consumption, the precision was low in the majority of the MR analyses, making it not possible to draw causal conclusions for these factors.

Chapter 2 gives an overview of the available evidence on the diagnostic performance of natriuretic peptides (NPs) to detect left ventricular diastolic dysfunction (DD) and heart failure with preserved ejection fraction (HFpEF) based on a systematic review and meta-analysis of 51 studies. The meta-analysis indicated a reasonable diagnostic performance for both NPs for the detection of DD and HFpEF. As the risk of bias of the included studies is high and sensitivity of NPs for detection of DD or HFpEF is low compared to specificity, use of NPs alone to detect DD or HFpEF should be discouraged as recommended by current guidelines. In Chapter 3, we give an overview of the available evidence on the diagnostic performance and risk of bias of novel biomarkers in the detection of HFpEF. We included a total of 28 studies. Use of case-control/two-gated designs, exclusion of difficult-to-diagnose patients, absence of a pre-specified cutoff value for the index test without external validation, the use of inappropriate reference standards and unclear timing of the index test and/or reference standard were the most important forms of bias. In Chapters 4 and 5, we studied the association between inflammation and heart failure (HF), and HFpEF-associated echocardiographic measures. In Chapter 4, we performed two-sample Mendelian Randomization analyses. Nine inflammatory biomarkers with available genome-wide association studies (GWAS) among individuals of European ancestry were identified and included: C-reactive protein, immunoglobulin E, tumor necrosis factor, interleukin 1 receptor antagonist, interleukin 2 receptor subunit α, interleukin 6 receptor subunit α, interleukin 16, interleukin 17 and interleukin 18. For the associations between the identified SNPs and HF, we used the largest GWAS meta-analysis performed by the Heart Failure Molecular Epidemiology for Therapeutic Targets Consortium with 47,309 participants with HF and 930,014 controls. In the analyses, we included 63 SNPs, of which 51 were for C-reactive protein. C-reactive protein, tumor necrosis factor and interleukin 18 were not associated with risk of incident HF: ORs per 1 unit increase in log-transformed biomarker levels of 1.01, 1.11 and 1.01, respectively. Other inflammatory biomarkers were also not associated with risk of incident HF, but may have suffered from weak instrument bias. In Chapter 5, we investigated within-person and betweenpersons associations of biomarkers of inflammation and endothelial dysfunction with left ventricular ejection fraction (LVEF), left ventricular mass index (LVMI) and left atrial volume index (LAVI). We included 383 participants of the Hoorn Study, a Dutch population-based cohort, and 491 participants of FLEMENGHO, a Flemish population-based cohort. Both cohorts had inflammatory biomarker and echocardiographic measurements at baseline and ~8 and ~5 years of follow-up, respectively. Overall Z-scores for low-grade inflammation or endothelial dysfunction were not associated with echocardiographic measures. Effect modification by sex was apparent for TNF-α in the Hoorn Study and E-selectin in FLEMENGHO with LVEF. In the Hoorn Study, women whose TNF-α levels increased with 1-SD over time had a decrease in LVEF. In FLEMENGHO, men whose E-selectin levels increased with 1-SD over time had a decrease in LVEF. In Chapter 6, we determined the sex-specific prospective associations between body composition measures and cardiac structure and function after 8 years of follow-up in 321 participants of the Hoorn Study. Total body fat, trunk fat and leg fat were associated with increased LVMI, while lean mass was associated with lower LVMI. Body composition measures were not associated with LVEF or LAVI. These associations were not modified by sex or mediated by inflammation. In Chapter 7, we studied the associations between serum and dietary advanced glycation endproducts and cardiac structure and function. Increased pentosidine and overall serum AGEs Z-scores over time were associated with decreased LVEF at follow-up.