Assessing Heart Failure Risk Factors and Prediction Models in People with HIV

About This Grant

PROJECT SUMMARY/ABSTRACT Cardiovascular disease is the leading cause of death in the United States, and heart failure (HF) rates in particular are escalating at alarming rates in the general population, with approximately one in four adults in the United States anticipated to develop HF in their lifetime. Today, HF affects an estimated 6.7 million American adults and is associated with a high risk for premature death, hospitalizations, and poorer quality of life. For the aging cohort of people with HIV (PWH)—who are living longer lives thanks to the advent of antiretroviral therapy (ART)—there is a well-established higher risk of atherosclerotic cardiovascular disease (ASCVD) driven by a higher prevalence of traditional and HIV-specific risk factors. Unfortunately, there is very limited research for HF in PWH, largely due to the scarcity of large longitudinal cohorts and the difficulties involved in reliably identifying HF cases. In one of the few studies on the topic, a large multicenter analysis from Kaiser Permanente (KP) demonstrated a 73% higher risk of developing HF in PWH compared with demographically matched people without HIV (PWoH). Thus, there is now an urgent need for a clear understanding of the underlying drivers of risk in PWH. To address critical uncertainties in the field—and ultimately to inform strategies for HF risk prediction and prevention among aging PWH—we propose a cohort study in three integrated health systems with an overarching goal to improve our understanding of contributors to the elevated HF risk in PWH. Our specific aims include: (1) comparing the impact of traditional and HIV-specific risk factors on the risk of incident HF among PWH and PWoH; (2) evaluating and refining the recently introduced American Heart Association Predicting Risk of CVD EVENTs in HF (PREVENT-HF) equations specifically for PWH; and (3) evaluating the impact of medications—including statins and ART classes—on the risk of HF among PWH. We plan to evaluate these aims using advanced biostatistical methods for the development, calibration, and validation of HF prediction models and applying causal inference techniques to evaluate medication effectiveness using observational data. The study will be conducted in three US healthcare delivery systems—KP Northern and Southern California and Mass General Brigham—where we anticipate a cohort of over 50,000 PWH and 10:1 matched PWoH from 2008–2025, with a highly sensitive approach for ascertainment of HF cases using a validated natural language processing algorithm. Study innovations include: (1) exceptional cohort of PWH and matched PWoH with comprehensive electronic health record data and validated HF outcomes; (2) first-of-its-kind analysis of the PREVENT-HF equations in PWH; and (3) use of advanced analytics to evaluate medication effectiveness for preventing HF. This study has great potential to generate critical evidence to guide HIV-specific HF risk prediction and prevention strategies, ultimately reducing HF incidence in PWH.