Heart failure (HF) is a major public health problem with high mortality (~50% at 5 years) and hospital readmission (~25% at 30 days). Angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) improve both outcomes in patients with HF with reduced ejection fraction (HFrEF). However, these drugs also adversely affect kidney function, and may increase the risk of acute kidney injury (AKI), chronic kidney disease (CKD) progression, and incident kidney failure, leading to end-stage renal disease (ESRD) requiring renal replacement therapy. All these risks are higher in HFrEF patients with CKD and those receiving these drugs in high doses. We have demonstrated that ACEIs or ARBs may reduce mortality in HFrEF with CKD (PMC3324926). Findings from our work also suggest that clinical benefits of ACEIs or ARBs might be similar at both low and high doses. The objectives of the proposed study are to test the hypotheses that low-dose ACEIs and ARBs are safe and beneficial in patients HFrEF with CKD. We will then develop a machine-learning algorithm to identify individual HF patients who might benefit from these drugs given their unique ejection fraction, kidney function, and other baseline characteristics.
These aims will be achieved by using VA's national data (over 1 million HF patients) and the American Heart Association's Get With The Guideline (GWTG) HF data (over 1.5 million HF patients) linked to the United States Renal Data System (USRDS) data. HF will be adjudicated using an automated machine-learning algorithm. An active-comparator new-user design with propensity score matching and sensitivity analysis will be used to compare clinical and renal outcomes in patients receiving low-dose vs. high-dose ACEIs or ARBs. Machine learning will be used to develop a risk prediction model to maximize clinical benefit and minimize renal harm for individual patients. The investigative team consists of national experts in key content areas and has the collective experience and expertise to complete the project in a timely manner. Nearly half of the Class-I recommendations (benefit greater than risk) in national HF guideline are based on Level-C evidence (mostly expert opinion) and there is a need to expand the evidence base from which clinical practice guidelines are derived. Findings from the proposed project will provide evidence that will help clinicians use a personalized approach in the use of ACEIs and ARBs in patients with HFrEF so that potential risks and benefits are optimized.
Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers reduce the risk of death and hospitalization in patients with heart failure. However, clinical benefits and adverse effects of these drugs often depend on left ventricular ejection fraction, kidney function, and other baseline characteristics. We propose to develop a machine-learning algorithm that will help clinicians use a personalized approach to maximize benefit and minimize harm for individual patients.
