Heart failure with preserved ejection fraction (HFpEF) can prove a complex problem for clinicians — its prevalence is increasing, it mimics many other conditions, and it can have a devastating impact on quality of life. But while emerging therapies are shifting the treatment landscape, are advances in diagnosis keeping pace?

HFpEF is a complex syndrome in which patients have signs and symptoms of heart failure due to elevated left ventricular filling pressure despite a left ventricular ejection fraction (LVEF) ≥ 50%. Once diagnosed, treatment of HFpEF is relatively straightforward and effective at reducing hospitalizations and improving quality of life.

However, diagnosis remains challenging in certain patient populations, according to experts in the field.

‘Kicking the Tires’

For a significant portion of patients with HFpEF, diagnosis is predominantly objective and based on elevated filling pressures in the heart at rest or during exercise.

“Sometimes it’s clear that patients have congestion based on the physical examination, echocardiography, or an NT-proBNP measurement,” Barry A. Borlaug, MD, the William J. and Sharon A. Schoen Professor of Cardiology at the Mayo Clinic in Rochester, Minnesota, told Medscape Medical News.

But many patients with HFpEF — approximately one third, according to an American College of Cardiology (ACC) scientific statement — have normal filling pressures at rest, so the congestion is only apparent during the stress of exercise. The presence of normal filling pressures can result in significant delays in diagnosis. For example, in a recent survey study, almost 50% of patients with obesity waited an average of 22 months between their initial conversation with their healthcare provider and their eventual HFpEF diagnosis.

“That is not necessarily the fault of those doctors,” said Daniel Silverman, MD, a cardiologist and assistant professor at the Medical University of South Carolina in Charleston, South Carolina. He noted these physicians may have performed proper evaluations but didn’t adequately “kick the tires.”

“These patients become symptomatic when they exert themselves, but that is rarely when we assess their hearts,” Silverman said in an interview.

Indeed, the patients in the survey who experienced delays in diagnosis also waited an average of 11 months from the onset of their symptoms to see their primary care provider, partly because they believed that dyspnea on exertion was a normal part of aging.

To shorten the time between symptom onset and diagnosis, researchers have made considerable effort to identify HFpEF phenotypes.

The obesity-related phenotype is reasonably simple to identify. Once identified, the ACC recommends behavioral and lifestyle interventions to promote weight loss and improve functional status. They also suggest GLP-1 and glucose-dependent insulinotropic polypeptide/GLP-1 receptor agonists as potential therapies, noting that research shows these drugs may reduce symptom severity, improve exercise capacity, and decrease the risk for heart failure hospitalization.

Clinical trials are also underway to further investigate how patients with left heart disease, with and without concomitant pulmonary vascular disease, can be distinguished and targeted with specific therapies, according to Silverman.

“We’re trying to tailor therapies to hemodynamic signatures,” he said. “It’s an imperfect process because it requires careful hemodynamic subtyping, mostly through exercise right heart catheterization. That’s just not readily accessible for most patients or for most catheterization labs.”

But the research is being done. The hope, Silverman said, is that the results of invasive testing can produce a hemodynamic signature that can be identified through noninvasive means, such as imaging or a blood sample. That signature can then be used as a surrogate for more invasive methods such as right heart catheterization.

The Utility of Algorithms

Novel HFpEF scoring tools have also notched a win for noninvasive assessment.

H2FPEF, introduced by Borlaug and colleagues in 2018, is a weighted composite score based on simple clinical characteristics in addition to echocardiography. The tool enables clinicians to predict the likelihood of HFpEF among other noncardiac causes of dyspnea and can also help determine the need for further testing. One of its biggest advantages is ease of use. 

“If you’re a cardiologist or an internist, it’s easy to remember that there’s a tool for trying to figure out whether this patient with unexplained dyspnea might have underlying heart failure,” Silverman said.

In some cases, the patient’s H2FPEF score can indicate a slam-dunk HFpEF diagnosis when combined with a thorough workup. In others, he said, the algorithm can help guide the clinician’s decision to refer the patient to a center for right heart catheterization or to a practice with enough HFpEF expertise to guide the patient’s future care.

The HFA-PEFF algorithm, endorsed by the European Society of Cardiology in 2019, is another diagnostic method that offers the clinician more deliberate guidance. It provides a framework for how to approach a patient with unexplained dyspnea or one at a risk for HFpEF using both noninvasive and invasive methods.

“It shows you how to rule in or rule out HFpEF rather than giving you an integer score that’s readily applicable,” Silverman said. “It’s useful for a physician who sees patients with unexplained dyspnea or for those whose patients are at risk for HFpEF — which, quite frankly, is everyone.”

In 2024, Borlaug and his colleagues published one of the more recent advances in diagnostic algorithms: the HFpEF-ABA score. They developed the tool to provide diagnostic assistance using three easily accessible criteria: patient age, BMI, and a history of atrial fibrillation.

“Just knowing those three criteria can help identify which patients with shortness of breath may have HFpEF,” Borlaug said. “These scores can really help on the front lines in primary care and internal medicine, since that’s where most of these patients show up.”

Machine learning is also paving the way for potential breakthroughs. Borlaug described several interesting tools, including an ECG-AI model that uses neural networks and LVEF values derived from echocardiography to categorize 12-lead ECG results into four categories of heart failure. The model demonstrated moderate accuracy in identifying patients with HFpEF.

Another AI tool uses neural networks and a single four-chamber echocardiogram to calculate the probability that a patient has HFpEF.

“In the real world, many patients with HFpEF don’t have a lot of diastolic dysfunction on echocardiography,” Borlaug said. “These models are promising, but they’re not ready for prime time yet.”

Advances in Diagnostics

While perhaps not as headline-grabbing as a diagnostic algorithm powered by neural networks, the addition of an intravenous bolus to a right heart catheterization is a much more widely available alternative to a true invasive cardiopulmonary exercise test.

“It’s a fluid bolus of 500 mL administered over 5 minutes during right heart catheterization, and you’re looking for the pulmonary capillary wedge pressure to rise above 18 mm Hg,” Silverman said. “The procedure has good fidelity for ruling in or out the impaired diagnostic filling that is pathognomonic for HFpEF.”

Silverman also acknowledged the benefits of stress echocardiography. The modality requires a fair bit of sophistication for the echo lab and the sonographer, but it has been an effective noninvasive surrogate for discriminating between normal and abnormal diastolic filling.

“It’s an area that is promising, but it isn’t widely implemented,” he said. “But it speaks to the goal of moving away from giving everyone an exercise test during right heart catheterization.”

Experts also hope to identify a biomarker that can serve as a surrogate for invasive diagnostics. Silverman said, for instance, that researchers are investigating the soluble ST2 protein and other inflammatory markers. Also, he noted many clinical trials in HFpEF collect blood samples specifically to identify responders vs nonresponders in the search for a predictive biomarker.

That goal, he said, “is somewhere between a daily accessible lab and a pipe dream.”

Shortening the Search for a Diagnosis

In the not-too-distant past, a common path of a patient with HFpEF may have involved a presenting complaint of unexplained dyspnea in primary care, followed by a referral to an internist, a referral to another internist, a referral to a pulmonologist, a referral to a cardiologist to rule out obstructive coronary disease, and, finally, a prescription to lose weight.

In modern practice, Silverman explained, it’s much more common to see internists refer patients for simultaneous pulmonology and cardiology consultations or even for an internist to refer a patient directly to a heart failure center.

“I think there are now internists who are more aware of HFpEF as a differential diagnosis and who know that direct referral to a heart failure specialist is not an inappropriate option,” he said.

Borlaug agreed and explained that overcoming inertia on the front lines would lead to quicker, appropriate referrals for heart failure. If a patient with overweight presented with unexplained dyspnea in the past, the physician may have said something along the lines of, “Well, you’re getting older,” or “You’ve gotten too heavy.”

“Being older and being too heavy are two of the strongest risk factors for HFpEF,” Borlaug said. “You have these [algorithms] now, and you can calculate these scores and refer them to cardiology if the results suggest additional workup.”

Ideally, he added, primary care physicians could obtain echocardiography and an NT-proBNP measurement simultaneously in such patients so that the cardiologist has those results in hand at the time of evaluation.

“Not everyone whom you refer is going to have HFpEF, of course, but you’re going to detect a lot of patients who can then receive treatment and feel substantially better,” Borlaug said.

Borlaug reported having relationships with Amgen; Boehringer Ingelheim; Edwards LifeSciences; Merck; Novo Nordisk; Eli Lilly and Company; Aria; AstraZeneca; Corvia Medical; National Institutes of Health/National Heart, Lung, and Blood Institute; Tenax Therapeutics; Medtronic; Tin Alley Ventures; Imbria Pharmaceuticals; Janssen; NGM Bio; ShouTi; VADovations, Inc.; DoD; Axon; and Rivus Pharmaceuticals.Silverman reported having no relevant financial relationships.