HFpEF: The 'New' Heart Failure
HFpEF: The 'New' Heart Failure
Henry R. Black, MD: I am Dr Henry Black, Adjunct Professor of Medicine at the Langone New York University School of Medicine. I'm here today with Dr Bert Pitt.
Could you tell us about heart failure? How do you define it, what is the epidemiology, and what are the symptoms?
Bertram Pitt, MD: In the past, we have focused on heart failure with a reduced ejection fraction (HFrEF), in which patients have had either an infarct, viral myocarditis, or some other form of damage. Over the past several years, we have begun to focus on heart failure with a preserved ejection fraction (HFpEF), in which patients typically have hypertension, diabetes, and aging (or some combination of those conditions). They have a stiff ventricle that is noncompliant compared with patients with HFrEF.
The analogy that I have always used when I talk about HFrEF is that of an eight-cylinder car. If some fuel lines get clogged, that car can go slowly down the street, but it can't get up the hill. It's the same thing in the heart. Patients have only four cylinders. If you lose part of the heart muscle, you are okay at rest and with minimal exertion, but you can't exert yourself to get up the steps and so forth.
For HFpEF, the analogy that I use is a bellows. The heart expands, but if for some reason that bellows becomes stiff and can't expand, it can't fill, and the blood backs up and the heart has low output. It's a different pathophysiology, but both seem to be associated with a very high event rate.
HFpEF is increasing in frequency because of the aging population, and the rising incidence of diabetes and obesity. This is the problem of the future. We have done fairly well—not fantastic—over the years in reducing mortality in HFrEF, but we haven't done very much in HFpEF.
Dr Black: Is that what we used to call "diastolic dysfunction"? Do you still use that term?
Dr Pitt: Yes, but because we haven't done a good job of defining diastolic dysfunction in many of these trials, the preferred term these days is "HFpEF" or "normal ejection fraction." None of these terms are terribly satisfactory, because they probably envelop several pathophysiologies.
Dr Black: Is it possible, without doing any evaluation or blood tests, to distinguish those clinically, with symptoms?
Dr Pitt: Not really. First, you have to be sure that the patient has signs and symptoms of heart failure. We have used the ejection fraction, but people are increasingly demanding a biomarker—proof—that the patient has heart failure, such as an elevated brain natriuretic peptide (BNP) level. They want proof that there is structural disease, although that is not usually the problem with HFrEF. There is structural disease when the ejection fraction is reduced, but it is a problem with HFpEF because those symptoms can be nonspecific. The BNP level could be raised for several reasons.
We would like to see some evidence of structural disease. We would like to see an enlarged left atrium or left ventricular hypertrophy. Diastolic dysfunction is helpful, but that is also very nonspecific. Many people with hypertension without heart failure can have diastolic dysfunction. It's the combination of symptoms—elevated BNP, a biomarker, and some structural disease—that people are increasingly focusing on.
Dr Black: I recall a time back in my Yale days when this was thought to be a more benign condition. That is not the case anymore, is it?
Dr Pitt: Most of the studies now are showing that the morbidity and mortality rates of HFpEF and HFrEF are similar. In some studies, the morbidity and mortality associated with HFpEF are a little—but not much—lower, compared with HFrEF. We are getting a lot more respect for HFpEF.
Dr Black: What do you do pharmacologically for this condition?
Dr Pitt: That's the difficulty. No strategy has shown a definitive reduction in mortality. Obviously, patients are being treated symptomatically, and despite the fact that nothing has been proved, many people are using angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) and beta-blockers, and sometimes calcium-channel blockers, without any real proof that these drugs are altering mortality. There are some clues, but nothing definitive.
We have recently carried out the TOPCAT trial with the mineralocorticoid receptor antagonist spironolactone, which is also controversial but has shown important insights into the effects of this drug.
Dr Black: You were responsible for the RALES trial. How did that evolve? Where did RALES start? Where are we now, ending in TOPCAT?
Dr Pitt: The RALES trial is an interesting story. Back in the early 1990s, the focus was on ACE inhibitors, and it was beginning to be on beta-blockers. I was approached by Searle, the manufacturer of aldactone (spironolactone) at that time, to see whether I was interested in doing a study with spironolactone. Before they had come to me, they had approached several of the leaders in heart failure and had been rejected, because everyone was focused on ACE inhibitors, ARBs, and beta-blockers.
Spironolactone was considered a diuretic. I had just given a talk at the American Heart Association (AHA) meeting on diuretics and had reviewed some studies from Belgium by John Staessen, in which he had shown that if patients with hypertension were on captopril (an ACE inhibitor), they had suppression of aldosterone, but after about 6 months, aldosterone levels of about one half of the patients rose back to control levels or higher. Knowing that, I said there was a chance that given that escape mechanism of aldosterone, if we use an aldosterone antagonist, it would have some effect on volume.
At that time, we didn't understand a lot of the pathophysiology of the mineralocorticoid receptor antagonists as well as we do now. Nevertheless, we thought there was a chance.
This is a very interesting story. We usually talk about translational medicine where we go from the bench to the bedside. Here, to a large degree, we went from the bedside to the bench, because the RALES results stimulated a tremendous understanding in the basic science of what happens to the mineralocorticoid receptor aldosterone and the aldosterone blockade.
Dr Black: The RALES trial was stopped early for dramatically good results. What happened after that?
Dr Pitt: The RALES trial was stopped prematurely because we had a reduction in total mortality. We were all surprised, because the trial was somewhat underpowered. We always thought it would be good, but we never thought it would be that good.
After the RALES trial, we had the opportunity to study another mineralocorticoid receptor antagonist, eplerenone. We decided to study that in patients with heart failure post-infarction. We did the EPHESUS trial, which was also very powerful. We showed a reduction in total mortality as well as hospitalizations.
Then we went on to the EMPHASIS Heart Failure trial in patients with HFrEF with mild symptoms. Once again, that trial was very positive. There was a reduction in total mortality and total hospitalizations. You couldn't want a better story. This was on top of everything we own—ACE inhibitors, ARBs, beta-blockers, and diuretics. We saw a reduction in mortality.
The downside is when you are using a mineralocorticoid receptor antagonist, it can cause some hyperkalemia. Many people got frightened by the incidence of hyperkalemia, and many people are still not using these drugs even though it is a class I guideline recommendation. In our trials, where we very carefully monitored patients and adjusted the dose, we have had hyperkalemia, but we have not had a single death attributed to hyperkalemia. We now have data showing that although spironolactone and eplerenone precipitate hyperkalemia, they also protect against related mortality. The placebo group, at the same level of potassium, has increased risk, but the risk in patients on spironolactone or eplerenone is much less. Patients are protected even though the drug precipitated it.
Dr Black: The latest entry is TOPCAT, which was finished last year. Some interesting data were just presented at the AHA this year. Could you review those data for us?
Dr Pitt: Yes. The first look at TOPCAT was about 3000 patients. We randomly assigned patients to receive spironolactone at a dose starting at 15 mg, going to 30 mg and then to 45 mg. Most people were somewhere around 20 mg.
The primary endpoint (cardiovascular death, resuscitated cardiovascular arrest, or heart failure hospitalization) trended positively but was statistically negative. We didn't reach a significant reduction. We did have a reduction in hospitalization for heart failure and total hospitalizations, but the primary endpoint, which is the most compelling evidence, was negative. That was very disappointing, because we thought this would work.
However, when we looked more closely, we found that half of our patients, who had come from Russia and Georgia, and half who came from the Americas (Canada, the United States, Argentina, and Brazil) were very different. The placebo event rate (having nothing to do with the drug) in Russia and Georgia was about one fifth of that in the Americas, and not compatible with anything we knew about heart failure, whereas in the Americas, the event rate was markedly increased and compatible with what we have seen in previous trials in HFpEF: I-PRESERVE and CHARM-Added. We found that in the Americas, there was a significant reduction (26%) in the primary endpoint of cardiovascular mortality.
Mark Pfeffer recently presented a further analysis at the AHA meeting [simultaneously published in Circulation]. We looked not only at the differences at baseline—and the baseline characteristics were very different between the Americas and Russia and Georgia—but what was striking is that the patients in the Americas had what you would expect. When they received spironolactone, they had an increase in serum potassium (hyperkalemia) and serum creatinine and a marked drop in blood pressure. Patients in Russia and Georgia had very little change in serum potassium, serum creatinine, or blood pressure. At first we thought that maybe it was because they had better renal function, but we matched people for renal function and we still found this difference.
It would appear that people weren't taking the drug and didn't have heart failure. We believe that in appropriately selected patients, spironolactone really works. I still don't understand what happened. We are investigating this further.
We are going to look at blood levels. People said they were taking the drug. We don't have an explanation at the moment. I can understand why they didn't have heart failure, because the inclusion criteria included hospitalization, the major component of which was heart failure.
There is a different economic situation in Russia and Georgia, and it is possible that some of the physicians may have included patients who had shortness of breath owing to obesity or lung disease. I can understand that, but what I don't understand is patients and investigators checking off the box stating that the patients were taking the drug, when we have good circumstantial evidence at least that they weren't taking the drug. That is the part that we don't yet understand.
Dr Black: This is a very interesting story, one that apparently isn't yet complete.
Dr Pitt: We don't have anything definitive or any guidelines for HFpEF, and until we have something, in appropriately selected patients, spironolactone is good. Clinicians will have to make their own decisions. If we had something that was shown to be beneficial, then fine, use that—but until that time, if it was my patient (or it was me), I would use spironolactone.
Dr Black: I agree. Thank you, Dr Pitt, for this enlightening review of what has become a very important problem.
HFpEF: Like a Stiff Bellows
Henry R. Black, MD: I am Dr Henry Black, Adjunct Professor of Medicine at the Langone New York University School of Medicine. I'm here today with Dr Bert Pitt.
Could you tell us about heart failure? How do you define it, what is the epidemiology, and what are the symptoms?
Bertram Pitt, MD: In the past, we have focused on heart failure with a reduced ejection fraction (HFrEF), in which patients have had either an infarct, viral myocarditis, or some other form of damage. Over the past several years, we have begun to focus on heart failure with a preserved ejection fraction (HFpEF), in which patients typically have hypertension, diabetes, and aging (or some combination of those conditions). They have a stiff ventricle that is noncompliant compared with patients with HFrEF.
The analogy that I have always used when I talk about HFrEF is that of an eight-cylinder car. If some fuel lines get clogged, that car can go slowly down the street, but it can't get up the hill. It's the same thing in the heart. Patients have only four cylinders. If you lose part of the heart muscle, you are okay at rest and with minimal exertion, but you can't exert yourself to get up the steps and so forth.
For HFpEF, the analogy that I use is a bellows. The heart expands, but if for some reason that bellows becomes stiff and can't expand, it can't fill, and the blood backs up and the heart has low output. It's a different pathophysiology, but both seem to be associated with a very high event rate.
HFpEF is increasing in frequency because of the aging population, and the rising incidence of diabetes and obesity. This is the problem of the future. We have done fairly well—not fantastic—over the years in reducing mortality in HFrEF, but we haven't done very much in HFpEF.
Dr Black: Is that what we used to call "diastolic dysfunction"? Do you still use that term?
Dr Pitt: Yes, but because we haven't done a good job of defining diastolic dysfunction in many of these trials, the preferred term these days is "HFpEF" or "normal ejection fraction." None of these terms are terribly satisfactory, because they probably envelop several pathophysiologies.
Telltale Signs and Symptoms
Dr Black: Is it possible, without doing any evaluation or blood tests, to distinguish those clinically, with symptoms?
Dr Pitt: Not really. First, you have to be sure that the patient has signs and symptoms of heart failure. We have used the ejection fraction, but people are increasingly demanding a biomarker—proof—that the patient has heart failure, such as an elevated brain natriuretic peptide (BNP) level. They want proof that there is structural disease, although that is not usually the problem with HFrEF. There is structural disease when the ejection fraction is reduced, but it is a problem with HFpEF because those symptoms can be nonspecific. The BNP level could be raised for several reasons.
We would like to see some evidence of structural disease. We would like to see an enlarged left atrium or left ventricular hypertrophy. Diastolic dysfunction is helpful, but that is also very nonspecific. Many people with hypertension without heart failure can have diastolic dysfunction. It's the combination of symptoms—elevated BNP, a biomarker, and some structural disease—that people are increasingly focusing on.
Dr Black: I recall a time back in my Yale days when this was thought to be a more benign condition. That is not the case anymore, is it?
Dr Pitt: Most of the studies now are showing that the morbidity and mortality rates of HFpEF and HFrEF are similar. In some studies, the morbidity and mortality associated with HFpEF are a little—but not much—lower, compared with HFrEF. We are getting a lot more respect for HFpEF.
Dr Black: What do you do pharmacologically for this condition?
Dr Pitt: That's the difficulty. No strategy has shown a definitive reduction in mortality. Obviously, patients are being treated symptomatically, and despite the fact that nothing has been proved, many people are using angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) and beta-blockers, and sometimes calcium-channel blockers, without any real proof that these drugs are altering mortality. There are some clues, but nothing definitive.
We have recently carried out the TOPCAT trial with the mineralocorticoid receptor antagonist spironolactone, which is also controversial but has shown important insights into the effects of this drug.
RALES: Success With Spironolactone
Dr Black: You were responsible for the RALES trial. How did that evolve? Where did RALES start? Where are we now, ending in TOPCAT?
Dr Pitt: The RALES trial is an interesting story. Back in the early 1990s, the focus was on ACE inhibitors, and it was beginning to be on beta-blockers. I was approached by Searle, the manufacturer of aldactone (spironolactone) at that time, to see whether I was interested in doing a study with spironolactone. Before they had come to me, they had approached several of the leaders in heart failure and had been rejected, because everyone was focused on ACE inhibitors, ARBs, and beta-blockers.
Spironolactone was considered a diuretic. I had just given a talk at the American Heart Association (AHA) meeting on diuretics and had reviewed some studies from Belgium by John Staessen, in which he had shown that if patients with hypertension were on captopril (an ACE inhibitor), they had suppression of aldosterone, but after about 6 months, aldosterone levels of about one half of the patients rose back to control levels or higher. Knowing that, I said there was a chance that given that escape mechanism of aldosterone, if we use an aldosterone antagonist, it would have some effect on volume.
At that time, we didn't understand a lot of the pathophysiology of the mineralocorticoid receptor antagonists as well as we do now. Nevertheless, we thought there was a chance.
This is a very interesting story. We usually talk about translational medicine where we go from the bench to the bedside. Here, to a large degree, we went from the bedside to the bench, because the RALES results stimulated a tremendous understanding in the basic science of what happens to the mineralocorticoid receptor aldosterone and the aldosterone blockade.
Dr Black: The RALES trial was stopped early for dramatically good results. What happened after that?
Dr Pitt: The RALES trial was stopped prematurely because we had a reduction in total mortality. We were all surprised, because the trial was somewhat underpowered. We always thought it would be good, but we never thought it would be that good.
After the RALES trial, we had the opportunity to study another mineralocorticoid receptor antagonist, eplerenone. We decided to study that in patients with heart failure post-infarction. We did the EPHESUS trial, which was also very powerful. We showed a reduction in total mortality as well as hospitalizations.
Then we went on to the EMPHASIS Heart Failure trial in patients with HFrEF with mild symptoms. Once again, that trial was very positive. There was a reduction in total mortality and total hospitalizations. You couldn't want a better story. This was on top of everything we own—ACE inhibitors, ARBs, beta-blockers, and diuretics. We saw a reduction in mortality.
The downside is when you are using a mineralocorticoid receptor antagonist, it can cause some hyperkalemia. Many people got frightened by the incidence of hyperkalemia, and many people are still not using these drugs even though it is a class I guideline recommendation. In our trials, where we very carefully monitored patients and adjusted the dose, we have had hyperkalemia, but we have not had a single death attributed to hyperkalemia. We now have data showing that although spironolactone and eplerenone precipitate hyperkalemia, they also protect against related mortality. The placebo group, at the same level of potassium, has increased risk, but the risk in patients on spironolactone or eplerenone is much less. Patients are protected even though the drug precipitated it.
TOPCAT: Unexplained Differences
Dr Black: The latest entry is TOPCAT, which was finished last year. Some interesting data were just presented at the AHA this year. Could you review those data for us?
Dr Pitt: Yes. The first look at TOPCAT was about 3000 patients. We randomly assigned patients to receive spironolactone at a dose starting at 15 mg, going to 30 mg and then to 45 mg. Most people were somewhere around 20 mg.
The primary endpoint (cardiovascular death, resuscitated cardiovascular arrest, or heart failure hospitalization) trended positively but was statistically negative. We didn't reach a significant reduction. We did have a reduction in hospitalization for heart failure and total hospitalizations, but the primary endpoint, which is the most compelling evidence, was negative. That was very disappointing, because we thought this would work.
However, when we looked more closely, we found that half of our patients, who had come from Russia and Georgia, and half who came from the Americas (Canada, the United States, Argentina, and Brazil) were very different. The placebo event rate (having nothing to do with the drug) in Russia and Georgia was about one fifth of that in the Americas, and not compatible with anything we knew about heart failure, whereas in the Americas, the event rate was markedly increased and compatible with what we have seen in previous trials in HFpEF: I-PRESERVE and CHARM-Added. We found that in the Americas, there was a significant reduction (26%) in the primary endpoint of cardiovascular mortality.
Mark Pfeffer recently presented a further analysis at the AHA meeting [simultaneously published in Circulation]. We looked not only at the differences at baseline—and the baseline characteristics were very different between the Americas and Russia and Georgia—but what was striking is that the patients in the Americas had what you would expect. When they received spironolactone, they had an increase in serum potassium (hyperkalemia) and serum creatinine and a marked drop in blood pressure. Patients in Russia and Georgia had very little change in serum potassium, serum creatinine, or blood pressure. At first we thought that maybe it was because they had better renal function, but we matched people for renal function and we still found this difference.
It would appear that people weren't taking the drug and didn't have heart failure. We believe that in appropriately selected patients, spironolactone really works. I still don't understand what happened. We are investigating this further.
We are going to look at blood levels. People said they were taking the drug. We don't have an explanation at the moment. I can understand why they didn't have heart failure, because the inclusion criteria included hospitalization, the major component of which was heart failure.
There is a different economic situation in Russia and Georgia, and it is possible that some of the physicians may have included patients who had shortness of breath owing to obesity or lung disease. I can understand that, but what I don't understand is patients and investigators checking off the box stating that the patients were taking the drug, when we have good circumstantial evidence at least that they weren't taking the drug. That is the part that we don't yet understand.
Dr Black: This is a very interesting story, one that apparently isn't yet complete.
Dr Pitt: We don't have anything definitive or any guidelines for HFpEF, and until we have something, in appropriately selected patients, spironolactone is good. Clinicians will have to make their own decisions. If we had something that was shown to be beneficial, then fine, use that—but until that time, if it was my patient (or it was me), I would use spironolactone.
Dr Black: I agree. Thank you, Dr Pitt, for this enlightening review of what has become a very important problem.
