PROJECT SUMMARY The 500,000 patients in the United States with end-stage renal disease (ESRD) on hemodialysis (HD) suffer extraordinarily high rates of mortality at ~18% per year, with 50% of these deaths attributed to cardiovascular disease (CVD): namely, atherosclerotic CVD, heart failure, and sudden cardiac death. Effective approaches for prevention, treatment and risk stratification in the HD population are lacking. Medical therapies that are effective in patients without ESRD, such as statins, are not beneficial for patients with ESRD when tested in randomized clinical trials. Ironically, traditional CVD risk factors, such as hypercholesterolemia, obesity, and hypertension, have `reverse' associations with CVD outcomes among these patients as compared to the general population. Major obstacles to progress in these areas that are addressed in our proposal include: 1) poor understanding of the disturbed biology in patients on HD that leads to poor CVD outcomes; and 2) failure to consider CVD in the HD population as a complex, potentially heterogeneous entity that calls for a personalized approach to CVD risk stratification using CVD models that are individualized and whose risk factors are modifiable. Circulating protein levels can serve as modifiable biomarkers for CVD risk, guide therapy and elucidate causal biological pathways and mechanisms. We plan to take advantage of an advanced modified aptamer assay (SOMAscan) that measures ~ 5000 proteins in just 175µl of plasma with high sensitivity and specificity. To date, no studies have utilized large-scale proteomics to understand outcomes and biological mechanisms in patients with ESRD on HD. In Aim 1, we propose to study proteins associated with clinical CVD outcomes in 649 participants of the Chronic Renal Insufficiency Cohort (CRIC) who have had one or more study visits after initiation of HD, and validate our findings in 408 participants in the Predictors of Arrhythmic and Cardiovascular Risk in ESRD study. In Aim 2, we will investigate changes in proteins that associate with progression or regression of left ventricular mass in the Frequent Hemodialysis Network, and validate our findings in CRIC. Proteomics will be measured at two time-points 1-2 years apart in 2/3 of the 1324 participants so that we may study single time-point proteins as well as protein trajectories. We will employ innovative methods, including stratified Cox analyses that support differing baseline hazards for participants with and without CVD, and time-dependent covariates for risk models, to accommodate intervening events in our analysis of protein changes. Agnostic and targeted pathway analyses will elucidate biological networks among protein predictors. In carrying out these Aims we will create accurate, personalized and mutable risk models for CVD events in HD patients. We will identify heretofore-unknown proteins and biological pathways associated with CVD; some of these may eventually become targets for interven...