Atherosclerotic disease including coronary artery disease remains the number one cause of death in developed nations, including the United States. While risk factors can predict the majority of incident CAD, a large proportion of patients present with few or no know risk factors. In part this represents the stochastic nature of acute coronary syndromes, but it also results from uncaptured and unrecognized risk promoting events. These have be dubbed SMURFless events, or events without Standard Modifiable Risk Factors.
Specific Aim 1: Explanation of unexplained risk in SMURFless individuals
We will calculate the predicted risk of ACS for each SMURFless individual at enrollment into the UK Biobank using standard risk factors (including gender and age) and the standard QRISK3 algorithm used in the UK. We will then identify the ~1000 SACS in the UKB. We will use hazard ratios (HRs) derived from proportional hazard (PH) Cox regression models to evaluate the association of each potential novel risk factor (RF) on the risk of ACS over an individual’s enrollment in the UK Biobank, using the entire SMURFless population, with ACS at the outcome.
We will use multivariable PH Cox regression modeling to compute a weighted risk score based on selected risk factors. We will evaluate the benefit of employing the resulting risk score alone (comparing resulting ROC curves, AUCs, and C-statistics with the current QRISK3) and in combination with QRISK3 using different metrics, including C-statistics, AUCs, NRI (net reclassification index) or NB (net benefit), at 1 year, 2 years, and 5 years.