HFpEF Is Missed at Referral, But Not Untreatable: Mark Belkin, MD

Heart failure with preserved ejection fraction (HFpEF) isn’t missed because there’s nothing left to treat it with anymore—it’s missed because clinicians stop looking once an echocardiogram shows a normal ejection fraction, according to Mark Belkin, MD, director of the hemodynamics laboratory and an advanced heart failure cardiologist at the University of Chicago Pritzker School of Medicine.

In an interview with The American Journal of Managed Care^® (AJMC^®), Belkin, who moderated a recent Population Health Roundtable on the urgency gap between heart failure with reduced ejection fraction (HFrEF) and HFpEF¹, said the bigger obstacle now is awareness and referral, not a lack of therapy. He argued that HFpEF’s sheer patient volume means care cannot rest on advanced heart failure specialists alone and needs general cardiology, nephrology, endocrinology, and primary care involved. He also walked through how health systems and payers should weigh the mixed trial evidence behind nonsteroidal versus steroidal mineralocorticoid receptor antagonists (MRAs) as a newly updated American College of Cardiology (ACC) pathway² pushes finerenone toward first-line status and where AI-driven identification and incretin-based therapies fit into HFpEF’s future.

This interview was edited for clarity.

AJMC: The roundtable discussed the urgency gap in HFpEF and why some patients get missed. Can you speak to some of the reasoning for this and what you think causes the biggest gaps in the referral process?

Belkin: That was a big focus of the first part of the AJMC roundtable, and it really comes down to awareness, diagnosis, and, in some cases, almost a perceived futility of the treatment options that we have. On awareness, heart failure with preserved ejection fraction, or HFpEF, as a diagnosis is often missed. Their shortness of breath or dyspnea is often blamed on other reasons, whether it be obesity, deconditioning, or diagnoses they carry of asthma or COPD [chronic obstructive pulmonary disease] that often aren’t confirmed with testing via PFTs [pulmonary function tests] with a pulmonologist, and that carries forward in the chart.

From a diagnostic standpoint, echocardiograms are often one of the first things people do in anyone who’s short of breath, and when the ejection fraction is noted to be preserved versus reduced, there’s definitely less urgency there. People see that, okay, it’s not the heart; the EF [ejection fraction] is preserved, and don’t look beyond that to some of the other cardiac structural changes we see with HFpEF, things like left atrial enlargement, changes in the right ventricle, and diastolic dysfunction.

And lastly, particularly as there has been a significant increase in the number of available treatments for these patients being new, there is always a lag from data and science to guidelines to practice, so many providers and practitioners may still find that there isn’t much to treat HFpEF with outside of maybe diuretics. So they see it as an effort of futility—do you need to go through the extra steps to diagnose this properly, to refer appropriately, if there really aren’t treatment options? Which is no longer the case.

AJMC: When it comes to care ownership, who do you think should be responsible for patients with HFpEF?

Belkin: I personally am an advanced heart failure provider who has an HFpEF subspecialty clinic. Realistically, there are a lot of patients with heart failure in America—over 6.7 million, and that’s supposed to be about 8 million by 2030 based on the most recent heart failure stats, and roughly half of those have heart failure with preserved ejection fraction. There are only, I believe, somewhere around 1500 or so advanced heart failure providers in the country, and the fellowship itself is not filling as much.

Advanced heart failure providers take care of a lot of patients with heart failure with reduced ejection fraction, transplant, and LVADs [left ventricular assist devices], as well as patients with HFpEF and other infiltrative diseases such as amyloidosis or hypertrophic cardiomyopathy. That being said, while there is a renewed focus in the advanced heart failure space on HFpEF, such as I have, and a lot of ongoing exciting research there, we can’t, as a society, rely on just advanced heart failure physicians and practitioners to take care of these patients. At minimum, where the dissemination needs to happen is we need general cardiologists to be aware of the problem, diagnose it, and be up to date on the updated therapies for it. And even beyond that, with the advent of many of these therapies that fall into this cardiometabolic space that drive the pathophysiology of not only HFpEF but chronic kidney disease, diabetes, and MASLD [metabolic dysfunction-associated steatotic liver disease], a lot of these therapies are almost interchangeable. So that’s where we want not only advanced heart failure cardiology and general cardiology but also nephrologists, endocrinologists, hepatologists, and even primary care doctors.

As we look at CKM syndrome—cardio-kidney-metabolic syndrome—that essentially drives all these processes, and they tend to perpetuate each other. That is where, even amongst all the things primary care doctors are doing in the management of diabetes and CKD [chronic kidney disease], even without the multiple steps you need to diagnose HFpEF appropriately, they can use a lot of the similar treatments to treat those underlying diseases that not only predispose to HFpEF but actually are treating the HFpEF in the patients who have it.

AJMC: During the panel, there was both skepticism and support when it came to factors such as cost and evidence for nonsteroidal MRAs. In the real world, how should health systems and payers weigh that disagreement when guidelines haven’t yet caught up to the newest trial data?

Belkin: A lot of the skepticism is around essentially 2 big trials. There was a big trial for steroidal MRAs, or spironolactone, in HFpEF called TOPCAT (NCT00094302) back in 2014, and that was a neutral trial that had some signals of controversy related to some of the geographic variations—that’s really going beyond the scope of this discussion. On the other hand, there is a positive trial with the nonsteroidal mineralocorticoid receptor antagonist, finerenone, that was positive in patients with heart failure with ejection fraction over 40% for the combined primary outcome, which includes cardiovascular events, cardiovascular death, and heart failure hospitalizations combined.

With the controversy over the prior trial, there are questions among many in cardiology regarding whether we need to use a nonsteroidal MRA vs the older steroidal MRA. The reality is, while one is a negative trial and one is a positive trial, and that on its own often carries weight, you can also look a little deeper: we’re talking about the multiple comorbidities that lead to HFpEF, and many of these patients have chronic kidney disease and diabetes. The data for the nonsteroidal MRA finerenone in the chronic kidney disease and type 2 diabetes space has been positive on repeated large-scale trials, while the data looking at steroidal MRAs, specifically spironolactone, in those same populations has been neutral. So even though there’s controversy if you’re treating just specifically HFpEF, if you’re treating holistically the patient and their other comorbidities, the clear answer ends up being that there is optimal benefit with using the nonsteroidal MRAs.

Now that being said, of course, coverage and insurance issues can always be problems in our health care system, and at some point some MRA is better than no MRA—some patients aren’t able to be covered. There was generally widespread agreement on using spironolactone, particularly when finerenone is not able to be covered. In terms of guidelines, while the last heart failure guidelines were in 2022, they take close to 10 years to repeatedly come out because there’s a lot of work that goes into them, which was detailed during the roundtable, particularly as we had Dr Yancy [Clyde Yancy, MD, professor and chief of cardiology at Northwestern Medicine] there, who is one of the writers of the guidelines. They’re very specific about the way that evidence is interpreted and disseminated, and there will no doubt be additions in the next heart failure guidelines, particularly as it relates to HFpEF.

However, in the meantime, often the way forward to inform the rest of the cardiology society and beyond about appropriate therapy in the fast-paced, changing world of these novel therapies is through things such as the American College of Cardiology Expert Consensus Statement on HFpEF. They put one out in 2023 because there were new data at that point from SGLT2 [sodium-glucose cotransporter 2] inhibitors, and now, just 3 years later, they’ve put out a second one because now we have the addition of the nonsteroidal MRA data, 2 SGLT2 inhibitor studies, and what we’re starting to see with incretin-based therapies and GLP-1s [glucagon-like peptide-1 receptor agonists], and how those should all be incorporated into current diagnosis and management of HFpEF.

While, again, these aren’t the formal heart failure guidelines yet, they are official documents that point you in the direction with the appropriate evidence, carefully scrutinized before they’re published. A lot of that ends up turning into what’s needed: additional education, starting at the trainee level for our residents and fellows, but also for our advanced practice providers, and ongoing education that we all do with continuing medical education as practicing cardiologists and practicing physicians outside of cardiology.

AJMC: Registries such as Northwestern’s, largely being HFpEF, may not be replicable. How can health systems build a realistic and scalable version for heart failure with mildly reduced ejection fraction?

Belkin: The issues raised with the Northwestern registry were less about recreating that registry and more about noting the volume. They have this registry of all of their heart failure patients who are seen in clinic and/or admitted over the years—I’m not privy to it specifically, but there was a discussion around it in terms of volume. This was a few tens of thousands of patients with heart failure, and one of the members of the roundtable who has done research there was talking about how well over half of these patients, appropriately, as we would be expected statistically, have heart failure with preserved ejection fraction.

So the question became: in a large health system, which presumably has similar numbers of HFpEF as we see in the general population, we’re talking about tens of thousands of patients who need appropriate diagnosis and optimization of their therapy. How do we take care of all these patients and help optimize their care? One, it goes back to improving awareness and identification of HFpEF as a syndrome that is morbid, that is mortal, that is urgent to treat; the fact that we have tools to better diagnose it, and the fact that we have treatment options now to help change the trajectory of the disease and help keep patients out of the hospital and help them feel better. And starting there, awareness needs to spread beyond heart failure cardiology to general cardiology and to our other subspecialists such as endocrinology, nephrology, and hepatology, and to our primary care doctors, who are really the foundation of our health care system and are managing a lot of these chronic comorbidities such as hypertension, hyperlipidemia, and chronic kidney disease—and, in the end, they are treating HFpEF whether or not it is fully diagnosed.

The other piece there’s a lot of discussion around is expanding what have been novel utilizations of things like pharmacy-led clinics that we’ve used for heart failure with reduced ejection fraction into this HFpEF and CKM space, similarly with many heart failure discharge clinics, where they’re seeing patients who were discharged recently for a heart failure hospitalization—our highest-risk heart failure patients—who often see an advanced practice provider first before they see a physician. Having those practitioners well-versed in awareness, diagnosis, and optimal treatment for these patients is really the best way to at least begin by catching these patients in these bottlenecks posthospitalization. But really, we want to go upstream from that and prevent those hospitalizations and prevent the development of HFpEF through optimal risk reduction of these comorbidities.

AJMC: The discussion also touched on incretins as a potential fourth pillar and also AI-driven patient identification. When do you expect to see these move into more standard practice, and what needs to happen first?

Belkin: First, about the incretin-based therapies, our GLP-1s. These have largely entered the practice for a lot of patients with HFpEF. The overlap of HFpEF with obesity, particularly in that cardio-kidney-metabolic space, is really 2 Venn diagrams that are essentially overlapping. What we’ve seen is there was a lot of promise, or signal, that we’ve seen benefit in heart failure in the first place. In the large-scale diabetes trials that were not focused on heart failure, such as SELECT, there was actually a reduction in heart failure hospitalization regardless of whether patients had a history of heart failure.

This is very similar to what we saw in the early SGLT2 studies such as EMPA-REG OUTCOME (NCT01131676), CANVAS (NCT01032629), and DECLARE-TIMI 58 (NCT01730534) that showed us a reduction in heart failure hospitalization with SGLT2 inhibitor use regardless of history of heart failure. Those were later validated in direct trials for heart failure with reduced ejection fraction—there were 2 of them—and heart failure with preserved ejection fraction in phase 3 studies that then did show that reduction in the primary end point of cardiovascular death and heart failure hospitalization. We want to see that similarly with GLP-1s and HFpEF.

There have been earlier studies, such as STEP-HFpEF (NCT04788511) with semaglutide and SUMMIT (NCT04847557) with tirzepatide, that did show us reductions in heart failure hospitalizations, although the event rate is a little lower, improvement in quality of life, and reduction in some of the changes in cardiac biomarkers and cardiac structure that we anticipate seeing improvement in with HFpEF, such as a reduction in BNP [B-type natriuretic peptide] despite there being weight loss. For instance, with semaglutide, normally we see an increase in BNP with weight loss—there was actually a reduction, meaning there’s likely some underlying cardiac benefit. There are ongoing trials there, and I think if those turn out to be positive, as is currently hypothesized, those would very likely enter guideline-directed practice for patients with HFpEF.

As it stands now, there’s not a direct indication for using GLP-1s specifically for HFpEF, but there are very proven indications for using them for type 2 diabetes, with CKD, and for use with sleep apnea, and these are, again, many comorbidities that our patients have. So when it comes to managing, holistically, the patient with HFpEF and their associated comorbidities that may perpetuate HFpEF on its own, GLP-1s are often used. That being said, we always have to be stringent with the way we do clinical trials and interpret evidence, so for now those phase 3 trials aren’t there, but I would anticipate—they’ve already been entered into the ACC [American College of Cardiology] expert statement for use for the comorbidities that revolve around HFpEF—and I would anticipate, eventually, pending the trials, if they were to be positive, which again we don’t know yet, that they would enter the guidelines.

The second part of your question, about AI-driven patient identification, is obviously a very hot-button issue right now—how, now that we have the technological advances of AI, we can use that to leverage our large databases, such as Northwestern’s example, to identify our patients, especially the highest-risk patients, see who’s on optimal therapy, and encourage practitioners who have limited time to go to CME [continuing medical education] events and things like that to really optimize care in this fast-paced world. I don’t have the answer to what the best system currently is. I know there is a lot of very important research going on here in the implementation science realm—both how to identify patients, let’s say, from the EMR [electronic medical record] to provide them therapy, and a lot of AI research into using echocardiography, which, as we talked about, is really the bread and butter of an assessment of dyspnea, with multiple different technologies out there really assessing how to identify HFpEF based on the echo images, beyond what is being seen by humans.

There is a lot there that still needs to be assessed and validated before it really is broadly applied across the system, saying this is a diagnosis of HFpEF based on something AI-driven. And then, on the other hand, what is the optimal way to use AI for implementation? We don’t want all the alert fatigue and things we see with current electronic medical records that people are just going to ignore anyway. So how can we leverage that to truly get to the patients and ensure they’re on optimal medical therapy? I think that’ll be coming. I think we’re far from being in guidelines, for sure. Moving into clinical practice is really going to depend on what is happening with ongoing research in this realm. But overall, I think it’s likely coming regardless, and if done appropriately, in a way that helps our patients and improves morbidity and mortality, it would, of course, be very welcome.

References

1. Steinzor P. Confronting the “urgency gap” between HFrEF and HFpEF care. AJMC. August 18, 2026. Accessed September 15, 2026. https://www.ajmc.com/view/confronting-the-urgency-gap-between-hfref-and-hfpef-care
2. Steinzor P. Cleveland panel weighs new HFpEF pathway, access to finerenone. AJMC. September 8, 2026. Accessed September 15, 2026. https://www.ajmc.com/view/cleveland-panel-weighs-new-hfpef-pathway-access-to-finerenone