Body Mass Index and Initial Presentations of Cardiovascular Diseases

Obesity has many detrimental effects on the cardiovascular system that can manifest in a range of clinical presentations. Many of these are mediated by frequently coexisting conditions such as diabetes, hypertension and dyslipidemia but there are also direct, progressive effects on the arteries and the heart muscle. As evident from autopsies, even from early adulthood the degree of coronary artery disease (CAD) correlates with the BMI and particularly the amount of abdominal fat. The multiple mechanisms by which obesity exerts its effects are still being elucidated. Previous studies have assessed the effect of obesity on cause-specific mortality. Here the investigators analyze association of BMI with the first symptomatic presentation of cardiovascular disease across any phenotypes, to which the investigators refer to as "initial presentation" to distinguish from first presentation within a specific phenotype. For example, an initial presentation with myocardial infarction is an MI which is not preceded by stable angina, ischemic stroke or any other phenotype, rather than the first MI in a possible series of MIs, as is commonly used in other studies. Adjustments The key risk factor of interest is baseline BMI, which the investigators define as the most recent measurement of BMI recorded up to 2 years prior to cohort entry date. BMI will be analyzed based on the following categories: underweight (\<18.5 kg/m2), healthy (18.5 to 24 kg/m2), overweight (25 to 29 kg/m2), moderately obese (30 to 34 kg/m2), morbidly obese (\>35 kg/m2). The healthy BMI category will be used as the reference in regression models. Associations with BMI will be adjusted for age, sex, ethnicity, social deprivation, smoking, diabetes, and systolic blood pressure, total cholesterol (TCHOL) and high-density lipoprotein cholesterol (HDL). All baseline covariates will be obtained from within a 2 year window prior to study entry. Statistical analyses Cause-specific Cox models will be used to measure the association between BMI categories and endpoints of interest. Multiple imputation will be used to replace missing values in prognostic factors. The investigators will estimate lifetime risk for each endpoint over time adjusting for competing risks.