Controversial A-HeFT Produces Groundbreaking Science

At the 2014 AHA Scientific Sessions in Chicago, Illinois, CardioSource WorldNews: Interventions spoke with Dr. Clyde Yancy, professor of medicine and Chief of Division of Cardiology at Northwestern University Feinberg School of Medicine. It is known from V-HeFT that isosorbide dinitrate hydralazine improves survival for patients with heart failure and that the survival effect is very different depending on race; this interview focuses on Dr. Yancy’s recent paper that suggests that it’s not just race that predicts the therapeutic impact but something called the G protein beta-3 subunit of GNB3 genotype.

CSWN: A new paper in JACC Heart Failure looks at how a specific genotype may predict therapy response in heart failure. Our discussion is actually an analysis of A-HeFT. First, remind us about A-HeFT.

CLYDE YANCY, MD: This really is great information. I’m delighted that we’re bringing it to the community. Remember that in V-HeFT, one of the first positive randomized control trials in all of cardiovascular medicine, remarkably we identified that the positive signal was almost entirely couched in the cohort that was described as African American—only about one fourth of the population studied. Imagine that one cohort drove the whole end point. Well, the retrospective analysis was so provocative that it really prompted the curiosity of all of us and [we] said, “I wonder if we shouldn’t test this prospectively.” Dr. Jay Cohen obviously deserves the credit for the concept and driving it forward.

So as we were putting together A-HeFT understanding that we were going to do something very novel, even revolutionary: a race-only clinical trial to test a retrospective finding to see if it was positive. We thought as we were putting the trial together maybe we should identify something else that would help us understand better what’s going on with this group of patients than just a race-based phenomenon. And so in a very provocative way, and in one of the first trials to have a companion genetic sub-study, we actually did that and called it GRAHF (Genetic Risk of Heart Failure in African Americans).1 It was patterned after another similar type-genetic registry that was constructed at the University of Pittsburgh and led by Dr. Dennis McNamara. We had this great confluence of ideas and resources come together for A-HeFT and, indeed, launch this genetic sub-study.

As I recall, the response was, “So, do you mean we should be treating people differently by their race?”

There were bookends. One bookend said, “Wow. This is the beginning of truly personalized medicine.” If we can understand what this crude indicator race really reflects biologically, then this really is getting us to personalized medicine where you take that which responds well in [a patient] and [the patient] doesn’t take the stuff that doesn’t respond. But the other bookend was “How dare you?” I mean, to actually create more polarization in the community rather than less.

We all navigated those issues and, oddly enough, even after we published the results in 2004,2 literally for a decade—meaning up until now—the arguments continued. But nevertheless, the science has been sound. We have a dramatically positive adjunctive therapy on top of ace inhibitors and beta blockers with a 43% relative risk reduction in mortality. That, by the way, eclipses the relative risk reduction we’ve seen in the PARADIGM-HF study (with neprilysin inhibition) compared to ACE inhibitors. So there is a necessity now to really understand whom it is that might get that benefit.

How did GNB3 genotyping get on your radar in the first place?

So for many years, and courtesy of many investigators, there’s been a portfolio of SNPs, single nucleotide polymorphisms genotypes that have emerged as having an odd localization according to race, something that those who are investigators in this field would call a “linkage disequilibrium.” It’s been seen with angiotensin receptor genotypes. It’s been seen with beta adrenergic receptor genotypes. It’s been seen with nitric oxide synthase genotypes, aldosterone synthase genotypes, and now this g protein couple receptor GNB3 genotype.

Now, what’s interesting about GNB3 is that the background makes it especially intriguing because those with hypertension have already figured out “Hey, when this genotype is a target or a signal those individuals have more vasoconstriction. They have more renal disease. They have more left ventricular hypertrophy.” So things started coming together. Now, the first genetic analysis we did from A-HeFT was on the nitric oxide synthase genotype, and it was a soft signal but it was exciting to discover because it seemed to track with responsiveness. The second one we did was with aldosterone synthase. That seemingly was the strongest signal and all of these were biologically plausible because we really presume that what’s happening in this African American construct is that there is lesser availability of nitric oxide, and there’s more evidence of fibrosis. So the story was building. Now, with this third analysis, we have an even stronger signal with the GNB3 genotype.

In this paper from JACC Heart Failure, what are you reporting?

So we’re reporting that this GNB3 genotype has a prevalence in the African American cohort from A-HeFT of nearly 80%. That’s pretty remarkable. It’s more prevalent than hypertension was. It parallels the prevalence seen in the predicate dataset GRACE, and it has a very tight relationship not only with the composite end point—which was mortality, hospitalization and quality of life—but it even has a tight relationship with components of the composite. So when you start looking at p values of .001 for a genotype and a response, your appetite kind of wells up a little bit. But the caution is that this is still a retrospective analysis. Pre-specified, yes, but a retrospective analysis. And the data set was small: 350 patients out of the 1,050.

What could this mean?

Let’s really be very provocative here and say here was a treatment for heart failure, now the disease du jour with a 43% reduction in the risk of death that is minimally utilized in part because the conversation about race is halting. Most people don’t like to go there. Imagine what could happen if we understand the biological underpinnings of this response and recognize that, irrespective of race, if you have this sort of profile, perhaps this is the kind of regimen that would work best in you.

Now that, too, would have to be tested prospectively. But step one is to confirm what we’ve already discovered in a larger cohort, and so we are initiating GRAHF 2—500 individuals. They will be African American because we will use the compound on label, a fixed-dose combination of isosorbide dinitrate and hydralazine. We will follow those patients prospectively and understand if we replicate the findings in A-HeFT but, more importantly, we will have aggregated these patients according to their genotype. And because we’ve done that, we can then test prospectively whether the GNB3 genotype really is the signal that gives us the most information.

It’s too early for genotyping patients but, because the drug is available, it’s not too early to treat. Is that correct?

Patients should be treated with evidence-based therapy. Anyone [reading this] understands my role in the guidelines and my fervent belief that we should use guideline-directed medical therapy. No doubt about that. But as we’re going forward, remember that the technologies to do genotyping are widely available—ubiquitous if you will—and are no longer expensive. But we have to handle that technology very carefully.

And so the idea of testing this one signal, prospectively understanding that, to a certain extent, it’s a gamble because very few things in medicine boil down to just one signal being most important. But if there is at least some reason to pursue this further, it really would be yet another step closer to personalized medicine, and it would actually help all of us close the loop on this idea. Why have a treatment that potentially is this beneficial that is rarely utilized when we might better understand the biology and then maximally utilize it? That would be a good thing.

With more science behind it, one would think that this would be a good time to reduce the tone of any debate and say “We’re getting the science there. Use this drug if it is effective in your patients.”

It’s the beautiful progression of science, right? You end up with this study in the 1980s that has this positive result and, by just exploratory analyses, you discover a group that seemingly has a great response. You don’t know why. Then the next step is that you test that group prospectively, and you replicate the response—and it’s dramatically positive. And then you go through an implementation process that is minimally to moderate, modestly successful. But then you begin to continue to delve into the data set, and you operate on an idea that you established in the beginning: that maybe the answer would be in these data points, the genotypes. You discover that there’s several answers in those genotypes. You pick your best mark and then you go forward. It’s a progression of science. You get more specificity. You get more understanding. And maybe when this process is jelled, we really will have either a definitive answer that says, “Look. This is a profile that we should test prospectively and a group agnostic of race and identify the benefit.” Or you may say, “Well, that’s not the way this thing works.” And that’s okay. Sometimes negative information is good information.


  1. McNamara DM, Tam S, Sabolinski ML, et al. J Am Coll Cardiol. 2006; 48:1277-82.
  2. Taylor AL, Ziesche S, Yancy C, et al. N Engl J Med. 2004; 351:2049-57.

Dr. Clyde Yancy, professor of medicine and Chief of Division of Cardiology at Northwestern University Feinberg School of Medicine.

Clinical Topics: Heart Failure and Cardiomyopathies, Prevention, Acute Heart Failure, Heart Failure and Cardiac Biomarkers, Hypertension

Keywords: CardioSource WorldNews Interventions, ACC Publications, Angiotensin-Converting Enzyme Inhibitors, Heart Failure, Hydralazine, Hypertension, Hypertrophy, Left Ventricular, Isosorbide Dinitrate, Linkage Disequilibrium, Neprilysin, Nitric Oxide, Nitric Oxide Synthase, Polymorphism, Single Nucleotide, Receptors, Adrenergic, beta, Receptors, Adrenergic, beta-2, Receptors, Angiotensin, Vasoconstriction

< Back to Listings