EPA Use in Clinical Practice and Safety Profile

Steven E. Nissen, MD: Maybe we don’t want to get into this, maybe we do, but there’s this whole issue of APOC3. And it’s not a simple story.

Deepak L. Bhatt, MD, MPH: Oh, these are lipid lovers. Of course they want to get into APOC3.

Steven E. Nissen, MD: Well, yeah, I do, but not necessarily for all of our audience. What I’m interested in understanding better, and you may be able to help us here Deepak, is why? What were the drivers? What are we doing with this therapy that’s making risk go down? How do we interpret and understand this?

Deepak L. Bhatt, MD, MPH: It’s a great question, and I don’t know the answer. I think it’ll play out over the next decade or so. Much like even statins, the complete mechanisms of action weren’t fully elucidated in the initial launch of statins with the initial trials like 4F. I think it will take awhile. Within the trial, in terms of correlations with biomarkers—at least to date, and preliminary analyses as presented at the FDA—the only thing that we saw some sort of correlation with, in terms of biomarkers, was the EPA [eicosapentaenoic acid] levels. That is, the EPA levels in general in Western populations. By EPA level, I mean eicosapentaenoic acid level, and the blood is low. We gave a highly purified ethyl ester of EPA, and the EPA levels went up by a lot, by over 350%.

Those changes seemed to at least correlate with subsequent outcome. We didn’t find a clear relationship with any of the other biomarkers that we studied to date. Both of you have given the issue a lot of thought in the context of the REDUCE-IT trial but even for years before. What do you think?

Michael Miller, MD: Some of this hinges on really the very nice work, the basic science work, that Preston Mason over in your neck of the woods at Harvard Medical School, has done. And it shows me that EPA does have differential effects with respect to LDL [low-density lipoprotein] oxidation, for example, and intercalating and stabilizing some membranes compared with longer chain fatty acid like DHA [docosahexaenoic acid]. There’s reduction inflammation and oxidation—inflammatory characteristics. There’s also some literature on endothelial function. I think all those, if you look at a hypertriglyceridemia patient who manifests a lot of these prothrombotic tendencies, maybe EPA does work by not only lowering triglyceride levels per se but the company it keeps.

Christie M. Ballantyne, MD: Deepak, initially we talk about precision medicine. And basically, in terms of designing the trials, Steve, you know the trial was done in Japan. They already had a pretty high level of EPA at baseline, 1.8 g on top of low-dose statin. There was a subgroup that had high triglycerides, low HDL [high-density lipoprotein]. Based on that, this trial was designed to be somewhat enriched. There were no HDL criteria. Now in the STRENGTH study that was more strictly adhered to for it.

Deepak L. Bhatt, MD, MPH: Right.

Christie M. Ballantyne, MD: Here, what was done was to take the population, which based on a hypothesis-generating study would be enriched for risk. And tested in that population, which is the same thing that’s being done, and it would be nice to have a second trial. What happens with statins is that we found out they worked in a very broad range of patients. Which is interesting, Deepak, because it’s really not clear that we, as you pointed out, really have to have these. Who would be the patients that this could work in? It may be a broader range of patients. We know it worked in the population that was studied. But it was selected to be a narrower group than the JELIS study. It will be interesting to see.

Deepak L. Bhatt, MD, MPH: The JELIS study, for the audience: even among subspecialty cardiologists, nobody knows about this. It’s not that they forgot about this, but they didn’t even know about it. They never knew. But it was a large Japanese study in which patients were randomized to 1.8 g a day of EPA, eicosapentaenoic acid—a lower dosage, a different formulation from what we studied, but basically largely the same active ingredients. And there was a significant reduction, a 19% relative-risk reduction.

Steven E. Nissen, MD: It was statistically a little more. It wasn’t like the P value that you had, which was very robust. It was kind of very nominal.

Deepak L. Bhatt, MD, MPH: There were issues with the trial. But overall positive, overall secondary primary prevention cohorts with consistent benefit. But you’re right. As a trialist, you know no placebo control. It was open label. It was a low dose of statin. Again, at that time in Japan, that was not such an unusual practice pattern.

Michael Miller, MD: And this was group of patients who eat a lot of fish.

Deepak L. Bhatt, MD, MPH: Yes.

James A. Underberg, MD, MS: What you haven’t mentioned, though, is the safety of the drug, which is another key component when we’re thinking about adding the medication on to other medications, right? It was really well tolerated. This is 1 of the few medications that people come to us asking to be on. Right. How often do they say, “Put me on a statin”?

Steven E. Nissen, MD: Before we get back to that, I want to make 1 more point in response to Christie. I know you know this, but I want to make sure the audience knows this. If you look at all the other “triglyceride-lowering therapies”—the fibrates and various other studies—and then you look at the subgroup in those studies that have high triglycerides and low HDL, there is a signal. Subgroup analyses are risky. But what was done in REDUCE-IT is based on not just JELIS or anything else but the idea that the prior studies that looked at therapies didn’t study the right population. If you think about it, it would be like starting out with statins studying people whose LDLs are 70 mg/dL. You know, it probably wouldn’t have worked out as well in 4S [Scandinavian Simvastatin Survival Study] with 180 mg/dL or something like that. Maybe it was 160 mg/dL.

Christie M. Ballantyne, MD: It was about 190 mg/dL, I thought.

Steven E. Nissen, MD: It was pretty high. And you made the point—that was the right thing to do with REDUCE-IT. It was clearly the right thing to do. We did the same thing with STRENGTH, and we’ll learn that. And somebody may someday go back and do another set of studies and look at different populations to see if they benefit. That signal was very strong for other things. I might point out that Paul Ridker is doing a study with pemafibrate, which is a new fibrate.

Deepak L. Bhatt, MD, MPH: PROMINENT is the name.

Steven E. Nissen, MD: PROMINENT is the name of the study. It’s a contemporary fibrate study. I haven’t seen the design, but I’ll bet you anything he’s studying high-triglyceride, low-HDL populations.

Deepak L. Bhatt, MD, MPH: It’s very similar to STRENGTH, actually, in any population design: low HDL, high triglycerides, a mixture of secondary and high-risk primary prevention. Another important trial. Still enrolling but very important.

Steven E. Nissen, MD: We’ll know in about 4 or 5 years probably whether that works.

Deepak L. Bhatt, MD, MPH: Yeah. So these are very complementary sets of trials in terms of understanding the science and exactly what mechanisms may provide clinical benefit.

Steven E. Nissen, MD: But we got offtrack because there was an overreach to do studies in broad populations first. You made the very important point that you’ve got to study the narrow population first, and then you think about whether you can broaden the indication.

Deepak L. Bhatt, MD, MPH: Absolutely. Part of what also has hurt this particular field is that some of these studies are the supplements or those low dosages, the 1 g a day.

James A. Underberg, MD, MS: In otherwise-healthy people.

Deepak L. Bhatt, MD, MPH: DHA, right. That really has hurt things. But you were talking about safety, Jamie. That’s an important point to consider in any drug. Just so the audience is clear, we did see an increase in hospitalizations for atrial fibrillation [AFib] or flutter with icosapent ethyl versus placebo. And also a significant increase in minor—though fortunately not in major—bleeding. Those are 2 things to be aware of. But overall well tolerated.

James A. Underberg, MD, MS: And no increase of stroke in the AFib patients.

Deepak L. Bhatt, MD, MPH: No, in fact there was a 28% reduction overall in stroke in the trial, and consistent findings, even in those with a history of AFib at baseline, and even those who developed AFib during the trial. Overall, a reassuring safety.

Steven E. Nissen, MD: Deepak, you know the bleeding because we’ve always known that the fish oils have effect. What I didn’t expect or understand is mechanistically, do you have any insights about why atrial fibrillation goes up?

Deepak L. Bhatt, MD, MPH: No, not really. It’s something that’s been described before in the omega-3 world, so it’s not the first time. It does make it a little more likely that it’s a real finding and not just a spurious finding.

The AFib was prespecified and adjudicated. But the absolute increase in that end point was small. It was a 1% absolute increase over an average of 5 years. You know, the chair of our Data and Safety Monitoring Boards, Brian Olshansky, is an electrophysiologist. Mina Chung is another electrophysiologist on the Data and Safety Monitoring Board. You might wonder, why do we have electrophysiologists on the Data and Safety Monitoring Board if it’s lipid clinic? Because early on I was thinking, there might be arrhythmic potential or anti-arrhythmic potential. In either case, I thought it couldn’t hurt to have some electrophysiologist support.

Their interpretation of the AFib signal is they actually didn’t think it was that big of a deal. In fact, when a trial is showing a reduction in sudden cardiac death, a proportion of which might be malignant ventricular arrhythmias that are reduced, they said, “Look, put it in context. Not a big deal.”

Steven E. Nissen, MD: If I could trade a reduction in sudden death for an extra few atrial fibrillations, I’d take it every day of the week if it were me.

Deepak L. Bhatt, MD, MPH: One of the things that we presented at the FDA—we’d actually get all this stuff peer reviewed and published, but we did present it at the FDA—was that in terms of the increase in AFib hospitalization, it was largely in people who already had an identified history of AFib. De novo AFib rate was extremely low. Overall I think what Jamie said is quite accurate: it’s a safe drug. Overall in the trial—in a blinded trial, placebo-controlled trial—it was as well tolerated as the placebo. That is, if you look at significant adverse events, either in a very sensitive fashion or a very specific fashion, like leading not only statistically but numerically almost identical.

Steven E. Nissen, MD: I guess the only other thing is there’s just a little bit of GI [gastrointestinal] complication; with fish oils there’s often just a little GI intolerance. It almost always goes away over time. One of those therapies, like metformin, you have to just wait out that adverse GI effect, and it will tend to diminish over time.

Deepak L. Bhatt, MD, MPH: Yeah, I know, that’s another important point. Maybe you could just say a few words, because we talked about other trials, about STRENGTH. If you just want to update the largely primary care audience that’s listening.

Steven E. Nissen, MD: Yeah. STRENGTH is using a carboxylic acid derivative of both EPA and DHA. The punitive advantage is that it is better absorbed. It gets higher blood levels—I’m glad to hear that you found blood levels to be correlating. Although it doesn’t have pure EPA, it actually gets similar EPA levels to what you saw in REDUCE-IT. But it also has DHA. So it’s an enhanced bioavailable product.

We’re studying a population with high HDL—I’m sorry, high triglycerides, low HDL. It’s a little higher fraction of primary prevention, so it may get more information about the primary prevention. But it’s a very similar trial. The patients are probably not quite as high a risk; it’s 13,000 patients. The same number of end points are required to terminate the trial so it’ll have comparable statistical power, and it’ll be another opportunity to see whether in this population a different fish oil—derived product, with a little different pharmacologic profile, proves similar or dissimilar effects.

As you know, you can’t compare across trials because they’re different populations. But it would be very good for the field obviously to have confirmatory trials. It’s always good to have confirmatory trials. We’ve done this in statins. We took I don’t know how many trials before we decided we could no longer do placebo-controlled statin trials. We’re going to get there now with 2 reasonably well-powered trials.

Transcript edited for clarity.