Cardiovascular disease epidemiology focuses on distinct subtypes (coronary heart disease, stroke, heart failure) with different etiologies, pathways, and prevention strategies. Risk prediction models integrate multiple risk factors (hypertension, lipids, smoking, diabetes) and population-specific baseline risks. Biomarkers (troponin, natriuretic peptides, C-reactive protein) improve risk stratification. Prevention emphasizes modifiable risk factors with strong dose-response relationships. Surveillance of CVD incidence and mortality tracks temporal trends and disparities to guide population health strategies.
From your study of chronic disease epidemiology, you know that non-communicable diseases are defined by their long latency, multifactorial causation, and preventability — and cardiovascular disease (CVD) is the paradigm case. It is the leading cause of death globally, but it is also the domain where epidemiology has arguably had its greatest public health success: the dramatic decline in CVD mortality in high-income countries over the past five decades tracks almost perfectly with the identification and management of modifiable risk factors that epidemiologists discovered and quantified.
The first conceptual move is to recognize that "cardiovascular disease" is not one disease. Coronary heart disease (CHD) — angina, myocardial infarction — results from atherosclerotic obstruction of the coronary arteries. Stroke comes in two forms: ischemic (a clot blocks cerebral blood flow) and hemorrhagic (a vessel ruptures). Heart failure is a failure of the pump itself, often downstream of prior CHD or hypertension. These subtypes share some risk factors but differ in others — atrial fibrillation is a powerful stroke risk factor but less directly linked to CHD; LDL cholesterol is a strong predictor of CHD but a weaker predictor of hemorrhagic stroke — so lumping them together in an analysis can obscure important subtype-specific patterns.
Risk prediction models are the applied translation of CVD epidemiology. Pooling data from large cohort studies, epidemiologists derived multivariate models — the Framingham Risk Score, the Pooled Cohort Equations — that integrate age, sex, blood pressure, cholesterol, smoking status, and diabetes to estimate 10-year absolute risk of a cardiovascular event. These models embody a key lesson from your earlier study of disease frequency measures: absolute risk, not relative risk, drives clinical decisions. A relative risk of 2.0 for a risk factor means something very different in a 30-year-old (whose baseline 10-year risk might be 1%) versus a 60-year-old (whose baseline might be 15%). The same relative elevation doubles to 2% versus 30% — the treatment calculus differs accordingly.
Biomarkers refine risk stratification beyond traditional factors. Cardiac troponins are released when myocardial cells are damaged — even subclinical elevations below the diagnostic threshold for myocardial infarction predict future events. Natriuretic peptides (BNP, NT-proBNP) rise when cardiac walls are under stretch, flagging early heart failure. C-reactive protein, a marker of systemic inflammation, improves risk prediction in people with intermediate Framingham scores where the treatment decision is otherwise ambiguous. The epidemiological validation of a biomarker requires demonstrating that it adds discrimination (moves people between risk categories) and reclassification improvement beyond existing models — not merely that it correlates with outcomes. This is a more demanding standard than simple association, and it highlights the difference between a biomarker that is statistically significant and one that is clinically useful for guiding decisions.