LAUSR

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): AFA Insurance and the European Commission Seventh Framework Programme. Myocardial infarction is a leading cause of death globally, but is notoriously difficult to predict. While most studies to date have considered long timeframes for etiological and prediction models, we aimed to identify biomarkers of an imminent first myocardial infarction and design relevant prediction models. We used a novel case–cohort consortium of 2018 persons without prior cardiovascular disease (420 cases, 1598 subcohort representatives) from six European cohorts. We investigated associations of 817 proteins, 1025 metabolites, and 16 clinical variables at baseline with subsequent risk of a first myocardial infarction within 6 months after baseline. Associations in weighted, stratified Cox proportional hazards models that passed multiple testing bounds in the discovery sample were verified in an independent validation sample. We also developed a prediction model for clinical use. Forty-eight proteins, 43 metabolites, age, sex, and systolic blood pressure were associated with risk of an imminent first myocardial infarction in models adjusted for technical covariates only (Figure 1). Brain natriuretic peptide (BNP) was most consistently associated with imminent myocardial infarction risk, in models adjusting for additional covariates including age and sex. Using clinically readily available variables, we devised a prediction model (Figure 2) with good ability to discriminate between subsequent cases and non-cases (cross-validated C-index 0.78) to which the addition of BNP did not improve performance. We found 94 biomarkers associated with risk of a first myocardial infarction within 6 months in persons without prior cardiovascular disease, and devised a prediction model with good discriminative ability for an imminent first myocardial infarction in the general population, with potential for motivating primary prevention efforts.