New Approaches for Heart Failure with Preserved Ejection Fraction with Sanjiv Shah, MD

Erin Spain, MS: This is Breakthroughs. A podcast from Northwestern University, Feinberg School of Medicine. I'm Erin Spain, host of the show.  Heart failure is the most common cause of hospitalization among older adults in the U.S., and with the aging of our population, the number of cases are increasing. Northwestern Medicine cardiologist Dr. Sanjiv Shah leads the world's first clinical program dedicated to the study of heart failure with preserved ejection fraction or HFpEF. Nearly half of all patients with heart failure have HFpEF. Yet there is much that is unknown about this form of the condition and how to best prevent it and treat it. Dr. Shaw's team was recently awarded an $18 million grant from the National Institutes of Health to lead Heart Share, a program dedicated to understanding the mechanisms of HFpEF and identifying therapeutic targets for it. He joins me today to talk about this grant and important discoveries in heart failure taking place at Northwestern. Welcome to the show, Dr. Shah. 

Sanjiv Shah, MD: Thank you very much, Erin. It's really a pleasure to be with you here today. 

Erin Spain, MS: You are the Neil J. Stone, MD Professor of Medicine in the Division of Cardiology here at Feinberg and Director of Research for Northwestern Medicine's Bluhm Cardiovascular Institute. Explain to me heart failure, specifically HFpEF, and how you explain it to your patients who are diagnosed with this disease and why more focus needed to be put on HFpEF. 

Sanjiv Shah, MD: For the past several decades, doctors, patients, when they thought of heart failure, you think of a muscle, the heart muscle that's failing. So it must be sort of enlarged. It's weak. It's not squeezing well, and that's basically what we thought of heart failure. Ever since we've been able to image the heart and show that it's actually not squeezing well, that was the definition of heart failure. But then in the late 1970s, early 1980s, people started using echocardiography. So heart ultrasound, nuclear imaging of the heart, and they started noticing that approximately 40 to 50% of patients hospitalized with heart failure had a normal squeezing function of the heart, what we call a normal or preserved ejection fraction. So at that time they ascribed it to, well, the heart squeeze is fine. It must be relaxing poorly, it must be stiff. And so if the heart muscle is stiff, not enough blood can get in and not enough blood can get out. So two things happened. The blood is not getting into the heart, so it starts backing up into the lungs and the body causing fluid retention and causing congestion in the lungs. But at the same time, especially during exertion, not enough output is coming to the heart, to the skeletal muscles, and there's exercise intolerance, fatigue, etc. and if it becomes bad enough patients get hospitalized and the mortality rate is quite high.  

Erin Spain, MS: I want to talk a little bit more about these patients. They're typically older women who are more likely to be affected by HFpEF. Why is that? Why are women at a higher risk? 

Sanjiv Shah, MD: That's a great question, Erin. In fact, we had a five year grant from the American Heart Association called the Go Red for Women Study, where we teamed up with scientists at Johns Hopkins to study this. So a couple of interesting things came out of that work and the work of others. Although HFpEF is definitely more common in women, some of the work in epidemiology studies that we have done have shown that actually men and women are both at risk for HFpEF and in fact, as the future goes on and HFpEF treatments get better and prevention of effort gets better, it's really men and women equally are at risk. The reason why women more commonly get HFpEF is that they're less likely to get HFrEF. They tend to have higher ejection fractions, better squeezing function of the heart. And so that's really the answer. That said, there are some specific risk factors in women and there are definitely male/female differences in HFpEF pathogenesis. So one big one is menopause. So after menopause, we lose estrogen. Women lose estrogen, right? And estrogen really stimulates the nitric oxide system. And when women lose that, a couple of things happen. First of all, the vascular, the vasculature starts getting more diseased, the heart muscle starts getting more diseased, more stiff and thick. And what's really interesting is that it actually plays a role in obesity, too. There's a big stimulus for increased adipose tissue after menopause. Many of my patients come to see me and they said they were thin their whole lives. And then after menopause, they really started to gain weight. And that obesity and sedentary lifestyle are two of the biggest risk factors, specifically for this type of heart failure. 

Erin Spain, MS: Tell me about your work as a physician scientist here at Northwestern and how you became focused on heart failure. 

Sanjiv Shah, MD: Most times patients or people who go into heart failure, physicians who go into heart failure, they really focus on end-stage heart failure with reduced ejection fraction. They focus on ventricular assist devices and transplant. I kind of wanted to do something else where I combine heart failure with imaging and research and really found a home here at Northwestern starting in 2007. Now, why did we need to start a program like this? Well, two things happened. First of all, in heart failure clinics, which have been around for a long time, typically in a typical heart failure clinic–you know, when I was in fellowship, we might see 30 patients in a day. Heart failure is one of the most common syndromes out there. And maybe, maybe one or two would have heart failure with a preserved ejection fraction. These would be rare forms of HFpEF like cardiac amyloidosis or hypertrophic cardiomyopathy. They really wouldn't be the type of HFpEF we see in the general population, that is associated with obesity, chronic kidney disease, diabetes, hypertension, etc.. You know, when I was in training in the mid 2000s, I was thinking to myself, you know, no one's really focusing on these patients. Clinical trial enrollment was dismal. And so, you know, that was going on at the same time that the field was changing. I explained that most of the time people and doctors thought of this as a stiff heart muscle, which is true, doesn't relax properly. We began to understand that it was a lot more than that. When these patients exercise, their heart doesn't squeeze as well as it should. And it's affecting a lot of other organs besides the heart. It's affecting the skeletal muscles, the lungs, the kidney, the adipose tissue, even the liver. This is diastolic heart failure, a problem with diastole or relaxing of the heart. And people started to say, wait a minute, this is really heart failure with a preserved ejection fraction, even systolic function. The squeezing of the heart, if we look close enough, is not normal. And so that's why at the same time–this is all happening at the same time–and I said, you know what, I'm going to start a clinic in it. And a lot of people thought this was a dumb idea. To be honest with you, they thought, Oh, it's a really niche thing. It's like having a clinic with your left pinkie. And I said, No, no, no, you don't understand. This is really, really common. And luckily here at Northwestern there was the foresight Dr. Bob Bono, Dr. Mihai Gheorghiade, they were sort of like, this is really important, we want to support you in this. And so I came here in 2007, started the program, and the floodgates opened. I mean, there were so many patients in need and so many doctors here that said we would love to send you these challenging patients. There aren't any treatments for them. We don't know what to do with them. It's hard to diagnose it and figure out that they even have it. And so I think people were really, really accepting and welcoming of a program like this. 

Erin Spain, MS: That was in 2007. What has happened since that time and what is the current therapeutic landscape for these patients? 

Sanjiv Shah, MD: Well, let's talk about 2007. Back in 2007, we were really bad at diagnosing it. We were really bad at identifying it. We were really bad at starting enrolling clinical trials. All the clinical trials had failed. And yet these patients were out there, as I said. And the other thing was that we didn't have a framework of how to take care of these patients because no one was seeing them all in one place. So over the last 15 years, we've really expanded and grown. We've seen thousands of these patients. Our first clinical trial was one of the largest clinical trials done. It was called Top Cat. We enrolled 77 patients. We quickly became the top enroller in the world. You know, I think that was really important in putting our program on the map and saying if you have a dedicated program, you can really identify these patients and see these patients and put them into trials and start to develop a framework of how we even think about these patients. And you know, two things I think we did well. One was the identification. We didn't, you know, at the beginning, especially, we didn't rely on referrals, you know, just word of mouth. We did an electronic search query of the medical record using natural language processing, etc. Back then this was really fancy, but it was really quite simplistic. We were just looking for the words heart failure or two or more doses of I.V. diuretics or elevated biomarkers of heart failure. We'd go out, we find these patients, we pitch it to them and their doctors in the hospital. And we started seeing a lot of these patients and, you know, other people from around the world have come and visited our programs. I'm really proud to say that we've really, you know, been leaders in this and pioneers in this. And now there are programs, HFpEF programs all over the world that are doing what we're doing and even beyond a lot of my mentees out there. So that's been really great. And clinical trial enrollment is much better now. We enroll trials quickly. I think there's a lot of focus on this. The other thing that we did really well was we saw all the patients. We weren't just seeing a sort of rare, sickest patient. We were seeing them across the spectrum. And what that did was let us understand that this is really a heterogeneous syndrome. All of heart failure is heterogeneous because it's sort of the end result of all the bad stuff that happens to the heart. As opposed to heart failure with a reduced ejection fraction where the heart muscle doesn't squeeze very well and it's enlarged, HFpEF is therapeutically heterogeneous. We have four pillars of drug therapy. They work on almost everybody. HFpEF though is a lot more heterogeneous. So we have been pioneers in this area of saying this is a heterogeneous syndrome, we need phenotype specific treatment, we need to do machine learning, advanced analytics. And what I can tell you now is we had our first big success last year. After decades of failed trials, we finally had our first trials of a medication class that clearly and without any equivocation helped these patients. And that was SGLT2 inhibitors, sodium, glucose, cotransport two inhibitors. And these drugs are really amazing, trial after trial in diabetes, in chronic kidney disease, now in heart failure throughout the range of ejection fraction. Even without diabetes, they seem to improve outcomes and they seem to improve exercise tolerance, improve health status in these patients. Really an amazing achievement over the last 15 years. 

Erin Spain, MS: So it sounds like there were three different publications that just came out in the past year that you were involved with. How does this change the way that you're going to be able to treat and diagnose your patients? 

Sanjiv Shah, MD: It's really going to be phenomenal in the sense that we are really going to actually be able to tell patients we have a drug for them that will make them feel better, do more, stay out of the hospital more often. But, you know, I think that our battle is still not over. One of the things when I give talks about HFpEF, I show them that in one of the main trials called Emperor Preserve, which is published in the New England Journal of Medicine last year, even though it was a blockbuster result, you know, highly significant p-value, the absolute risk reduction in heart failure hospitalization was only 3%. And so, you know, we improve things but not a huge amount. And, you know, we have a long way to go. Even in another trial, we did preserved HF where we showed a 20 meter 60 feet improvement in six minute walk test distance and a big large six point improvement in their health status, which is one of the biggest improvements in any heart failure trial ever, where they end up is still pretty dismal. You know, it's still kind of not where you want to be. So there's still a lot more that we need to do.

Erin Spain, MS: What is being done at Feinberg now to detect those who are at risk? I know you're using some machine learning to identify patients. Tell me more about that. 

Sanjiv Shah, MD: Yeah, we are really trying to push the envelope here. Now we have, I would say, low-tech and high-tech ways of diagnosing these patients. A low-tech way that was developed by Sadiya Khan in our division of cardiology, one of the faculty who did her fellowship here and trained under us and now is wildly successful. She developed something called the PCPHF risk score, and that's just a simple score based on comorbidities so risk factors for heart failure and an EKG basically. So things that happen in a primary care office and we can tell pretty accurately who's going to be at risk for heart failure over the next ten years. Sort of how we pick a statin, you know, should we use a statin based on people's risk of coronary artery disease or stroke? This way, we're hoping to use this score to figure out if people are at risk for heart failure, maybe treat them with an SGLT2 inhibitor.  

Now, the high-tech ways, I think are twofold. You mentioned machine learning. We are doing a lot of that. We're using natural language processing, we're using machine learning of ICD codes. So medical diagnostic codes, we're using machine learning of echocardiograms, heart ultrasounds and EKGs to really have a computer learn how to diagnose heart failure and also figure out what type of heart failure it is. You know, is it a rare type of HFpEF like cardiac amyloidosis that is a specific treatment. The other thing is we are doing a lot of advanced analytics in the clinics. So we are doing these exercise studies where we have patients do low levels of exercise. And the one thing we know about HFpEF is that at rest a lot of things look normal and the patients feel fine, but they walk a couple of steps down the hallway and they are completely breathless. And that's really what we want to mimic with our testing. We don't want to just image them at rest and measure their pressures at rest. We really want them to exercise or do something. And that really brings out the abnormalities that we can use to diagnose the syndrome.   

Erin Spain, MS: I want to talk about this $18 million grant from the National Institutes of Health that your team was recently awarded. It's part of the Heart Share program. Tell me about this grant. What is Northwestern's role in the program and why is this such an important grant? 

Sanjiv Shah, MD: Well, Erin, let me tell you how it all started. So back when really when I first started here, as I was saying earlier, I really thought that this syndrome was heterogeneous. And I think anyone who takes care of these patients feels like it's heterogeneous. So my first grant from the American Heart Association was to come up with a way to classify these patients. And so I was trying all these ways. I had developed a very deeply phenotypes or rich phenotyping of a cohort of 400 patients in my clinic. And I was trying to stratify by coronary artery disease versus no coronary artery disease or something, you know, on their echocardiogram, their heart ultrasound. You know, it just felt like it was too simplistic. This is a very complex syndrome. And so I thought, you know, why don't we take all of this phenotypic data, quantitative phenotypic data, about 70 different variables from their laboratory testing, echocardiograms that were expertly quantified, their EKG, etc. And let's try to use machine learning to identify subgroups. And to my surprise, it actually worked. And we found three subgroups in a paper that was published in Circulation in 2015.  

What was interesting about that is that really nothing like that had been done in heart failure and HFpEF really in cardiovascular medicine. So it really took off after that, not just in cardiology, but in a lot of other fields. This sort of thought that you can use this unsupervised machine learning or clustering to let the computer find patterns in the data. And so that's kind of where we were starting from. And the NIH also was sort of thinking HFpEF is a big unmet need. You know, 3 to 4 million people in the U.S., there's not a lot of therapies. So let's call together experts, you know, about ten to twelve experts in the field and in related fields to the NIH in 2017 to talk about this for two days, let's just get deep into it. I chaired that program. It was called A Working Group on HFpEF on Research Priorities. And what we came up with was a network. We wanted to really join together, join forces together of networks, of investigators from the basic science to clinical to epidemiology and really try to figure this out, and sub classify HFpEF. And the NIH agreed to do such a program. They put out a request for applications back in 2020 and luckily enough, we were able to get that grant as the data translation center.   

So I lead the data translation center, sort of the hub, and there's six clinical centers that are sort of the spokes. There wasn't in fact a basic science component, although hopefully that will be added in the future. But really the goal is to try to use existing data. So existing data sets, images, all types of omics like genomics, metabolomics, proteomics that we already have, on thousands, tens of thousands, hundreds of thousands of individuals from studies we've done with the NIH, gather all that data, put it all up in the cloud, and learn from that. Learn about HFpEF from that and how to sub phenotype it. What are the specific types of HFpEF? What are the mechanisms? And then we're also doing a prospective portion of the study at the six clinical centers where we're doing an electronic health record query of all of their heart failure patients over a seven year period, the past seven years, inviting all of them to be part of a what we call a low touch registry where they interface with an app on their phone or computer, they wear wearables. And then of those 1000 patients, about 750 HFpEF, 250 controls without heart failure. We're going to be doing deep phenotyping. And then that means we're going to be doing very sophisticated exercise testing, imaging, not just looking at the heart, but looking at all the organs of the body and doing deep phenotyping of their blood and other tissues, really trying to get to the molecular basis of these subtypes. Now, this is a five year grant and in the future, ultimately, what we hope is that will be able to do precision medicine trials. So for example, if we find that a quarter of these patients have a specific type of HFpEF and there's a specific molecular pathway that's perturbed, we can develop drugs specifically for these patients and do these precision medicine trials. 

Erin Spain, MS: So this grant is happening and all this data collection is happening, but you're still publishing other papers and looking at other ways of treating HFpEF. You just published in the journal The Lancet about an atrial shunt that could benefit HFpEF patients. Tell me about that.   

Sanjiv Shah, MD: Well, that's a great story, Erin. And it's really a journey. And yet, you know, we're trying to do everything we can. My team and I have a wonderful staff of researchers and other faculty and fellows and trainees, post-docs. Our whole group is just dedicated to finding treatments, helping patients get diagnosed, and improving the lives of patients with HFpEF. So besides Heart Share, of course, we have other grants and we also lead a lot of the worldwide clinical trials in HFpEF. So one of the ones that I've been working on over the last 12 years, as I said, was a journey started with an idea. A pediatric cardiologist in Australia said, you know, in mitral stenosis, when the mitral valve doesn't open properly, the left atrium or upper left chamber of the heart gets very high in pressure. Same thing happens in HFpEF. But in that syndrome, it's been noticed over the last hundred years that if there's an atrial septal defect, a hole between the two atria, those patients actually do better. Why is it? Because their overloaded left atrial chamber has its sort of escape route. You know, the high pressure can decompress through that hole to the rest of the body, to the great veins, the right atrium, etc.. And so he had an idea. His name is David Selmayr. He said, you know, what if we put in a shunt device, so we basically create a hole in the inner atrial septum in patients with HFpEF. So he approached me with the idea, a few others. It started with animal studies, then first in human pilot studies and it has progressed. We led the first randomized controlled trial, which was just to see if the shunt in a sham controlled, you know, sort of placebo controlled fashion, did it actually lower pressures in the heart with exercise. And it did.  

And so then we embarked on a pivotal trial, and this was an immense amount of work by a ton of people here at Northwestern and across the world. Ninety centers across the world, and 626 patients were randomized. This was no simple study. This was every single patient going to the cardiac catheterization laboratory, undergoing invasive exercise, hemodynamic testing. And that's the gold standard for this syndrome. And I have to give it to the patients. I mean, they were amazing to really, you know, do a public service by being in this trial. So we randomized 1 to 1. It was a sham controlled trial. And then we followed them for a few years to see if we improved their health status. Did we reduce heart failure, hospitalizations? You know, did it have any adverse effects, which is an important thing. You know, when the results came out, it was a bit disappointing. It was a neutral trial. And, you know, it was like, sigh, again, is it neutral? We couldn't believe it because we had so much good evidence to begin with and we were so careful. We really enriched the trial with the patients we thought would benefit the most. But, you know, when a neutral trial happens, Erin, it's really one of two possibilities. Either everyone kind of just didn't respond. I mean, it really is neutral. It's just kind of had no effect or some patients benefited and some patients didn't. You know, and there may be a dichotomy there. And that's what we found here. Two-thirds of the patients seemed to clearly benefit. One-third didn't. And the two-thirds that benefited, we figured out what we didn't see. You know, we didn't have the crystal ball. You know, we didn't think of everything. And the thing we found out was normally in humans, the blood vessels in the lungs relax during exercise. And in a third of the patients in this trial, that didn't happen. And when that doesn't happen, when they start to constrict or they don't relax with exercise, that extra blood that's getting shunted across the heart does no good and actually could do harm. So in two-thirds of the patients, and we can identify this with that exercise testing, measuring the pressures, we can tell if their blood vessels in the lungs are relaxing. And so we hope to do a new trial to confirm this. Of course, we won't know until we do the confirmatory trial. It's really exciting to say we're getting more and more precise. The work we're doing is really seeming to make a benefit and we're really getting close, hopefully to not just one class of drugs for this patient population, but a whole host of drugs and devices to improve this disorder. 

Erin Spain, MS: If you could leave the listeners with something today that you want them to know about HFpEF and the future of, you know, therapies and treatments, what would it be? 

Sanjiv Shah, MD: That's a great question. I'd say two things. Number one, and I like to say this, heart failure is inevitable. You know, everyone's heart will fail. And I say that a little tongue in cheek, but I mean it in the sense that if we don't take care of ourselves, if we don't lead active, healthy lifestyles, this syndrome HFpEF, which is bad, is coming to get you. So, you know, we can cure all the cancer in the world. We can do a lot for other diseases, but this thing will get you if we're not careful. The second thing I think for the scientists listening in, I think that as we think about the future, I think we have to move away from trials of thousands and thousands of patients for really complex clinical syndromes that are poorly differentiated, like HFpEF. I think it does work for things like hypercholesterolemia, for hypertension, but even those may have more nuanced approaches. They already do. And I think that that's really the goal of precision medicine. It's not perfect and we have a long way to go, but I think that we really have to think about leveraging all the data we get now, the tons of data to really come to a future where we're more precisely identifying, defining these clinical syndromes. 

Erin Spain, MS: Well, I think that is a great place to end this episode today. Thank you so much, Dr. Sanjiv Shah, for sharing this great story about how HFpEF has really started so much research here at Feinberg and the exciting news ahead. We look forward to seeing what happens.   

Sanjiv Shah, MD: Well, thank you, Erin, very much into the listeners for listening. It was a great pleasure talking to you. 

Erin Spain, MS: Thanks for listening and be sure to subscribe to this show on Apple Podcasts or wherever you listen to podcasts and rate and reviews. Also for medical professionals, this episode of Breakthroughs is available for CME Credit. Go to our website, Feinberg.Northwestern.edu and search CME.