At least 50% of people with HF develop pulmonary hypertension (HF-PH). The fact that all people with HF have elevated left ventricular filling pressures suggests that there must be additional factors that drive the development of PH. Identifying these factors is important because HF-PH carries a 50% increase in mortality and no treatments exist to improve outcomes or prevent PH development. Epidemiologic data on the risk factors and natural history of HF-PH are lacking. Similarly, the biological mechanisms underlying HF-PH are unknown because no molecular studies have been performed in this population. In addition to establishing incidence rates and clinical risk factors, an important goal of this application is to identify molecular features associated with HF-PH and right ventricular (RV) compensation (i.e. preserved RV function in the setting of PH). We hypothesize that (1) HF-PH incidence rates using echocardiographic data are higher than previously reported rates based on medical codes (2) HF-PH and RV compensation are genetically influenced, and (3) protein biomarkers will be associated with prevalent HF-PH and RV function. These hypotheses are based on our preliminary showing: 1) higher rates of incident HF-PH using echo data than rates based on medical codes alone; 2) association of poor metabolic health with HF-PH and RV dysfunction; 3) high genetic heritability of pulmonary pressure; 4) shared genetic risk between obesity and pulmonary pressure; 5) a genetic association between insulin resistance and PH; and 6) elevation of inflammatory and vascular tone proteins in HF-PH patients. Developing large, prospective cohorts designed to study the natural history of HF-PH would be prohibitively expensive and inefficient. Leveraging electronic health record (EHR)-based cohorts linked to biobanks presents a scientifically valid, cost-effective, and efficient pathway for studying HF-PH epidemiology and pathophysiology.
In Aim 1, we will establish HF-PH incidence rates and examine the importance of modifiable risk factors for HF-PH (e.g. obesity, insulin resistance) by extracting echocardiographic PASP values on ~425,000 individuals in the Veterans Affairs and Vanderbilt EHRs (64,000 African Americans and 85,000 women). Approximately 100,000 of these individuals have repeat PASP measurements, and 65,000 have gold standard RHC data. Both cohorts are well phenotyped with detailed data on demographics, comorbidities, medication exposure, laboratory, and clinical events.
In Aim 2, we will leverage the VA-funded Million Veterans Program and Vanderbilt's BioVU to analyze genome-wide genotyping data in a total of 25,000 subjects with HF at no cost to this application.
In Aim 3, we will perform proteomic profiling (1129 proteins) in discovery (800 subjects) and replication (600 subjects) HF cohorts collected through BioVU. We have combined existing resources with new phenotypic, genotypic, and proteomic data and assembled a team with the specific expertise to execute our aims. If our hypotheses are correct, the results could improve clinical guidelines for HF-PH and identify new therapeutic targets for HF-PH and RV dysfunction.
Heart failure is the most common cause of pulmonary hypertension in the world. The clinical and molecular features that cause pulmonary pressure to increase in patients with heart failure are poorly understood. Accordingly, we propose epidemiologic, genetic, and proteomic studies of pulmonary pressure in large electronic health record cohorts to better understand the determinants of pulmonary pressure and associated outcomes.
