Development of Integrated Precision Medicine Models in Adults with Cardiovascular Disease

PROJECT SUMMARY This application aims to use existing data within the All of Us Researcher Workbench to improve our ability to identify cardiovascular disease (CVD) patients treated with statins or antiplatelet therapy (i.e. P2Y12 inhibitors) at increased risk for adverse drug responses and adverse outcomes. In order to manage CVD and prevent

adverse outcomes, CVD patients can be treated with multiple medications including lipid-lowering agents or statins and antiplatelet agents such as clopidogrel. There are well known clinical and genetic factors that impact the risk of statin-associated musculoskeletal symptoms (SAMS) with numerous statins and the risk of

adverse cardiovascular outcomes with the P2Y12 inhibitor clopidogrel. Despite guidance from expert consensus groups, and regulatory agencies, the implementation of pharmacogenetic testing for statins, and P2Y12 inhibitors has been limited, occurring mostly at large, academic medical centers. Additionally, almost all of the

studies conducted to date have been in populations of largely European and Asian ancestry. In order to provide equitable care and fulfill precision medicine for the over 53 million Americans estimated to be treated with statins or P2Y12 inhibitors annually, we need to better understand their usage and associated outcomes in

diverse, real-world populations. Our central hypothesis is that precision medicine models derived from diverse, real-world data for SAMS, and adverse outcomes after treatment with an antiplatelet agent, will be more precise and accurate than existing models To test our central hypothesis we will complete the following

Specific Aims: 1) Evaluate statin prescribing patterns, adverse drug responses, and adverse outcomes in patients with CVD by SLCO1B1, ABCG2, and CYP2C9 genotypes using electronic-health record (EHR)-based data, genomic data, and data from surveys and wearables, and 2) Characterize antiplatelet prescribing

patterns, adverse drug responses, and adverse outcomes in patients with CVD by CYP2C19 genotype using EHR-based data, genomic data, and data from surveys and wearables. To achieve these aims, we will utilize existing data from the All of Us Researcher Workbench. The All of Us Research Program is enrolling a diverse

group of persons in the United States, and including multiple types of real-world data (e.g. EHR, demographic, wearables, patient surveys, genomic). We will deploy validated CVD algorithms and determine observed rates of CVD, statin prescribing, and antiplatelet prescribing. We will identify characteristics of SAMS in patients with

CVD treated with statins and characteristics of adverse outcomes in patients with CVD treated with a P2Y12 inhibitor. We will also use multivariable regression analyses and machine-learning methods to model adverse drug responses and adverse cardiovascular outcomes. We will examine characteristics and models of CVD

patients treated with statins by SLCO1B1, ABCG2, and CYP2C9 genotypes, and of CVD patients treated with a P2Y12 inhibitor by CYP2C19 genotypes. Analyses will be conducted overall, by race and ethnicity, urban versus rural areas, across geographic regions, and by sex as a biological variable.