Recent clinical insights into icosapent ethyl

It's really a pleasure for me to provide a recap of the REDUCE-IT trial and highlight some of the

more recent findings that we've learned in the past three years since the initial presentation, publication of the data. Before I get started, these are my disclosures, relevant to this talk is research funding from Amarin who sponsored the REDUCE-IT trial, that's paid to Brigham and Women's Hospital.

Let me quickly review the design of REDUCE-IT. REDUCE-IT consisted of patients with secondary prevention, and by that I mean established coronary, cerebral, or peripheral artery disease. 70% of the trial population were such secondary prevention type patients and another 30% of patients were what I'll call high-risk primary prevention, what that means is diabetes plus at least one additional cardiovascular risk factor. All the patients were treated with a statin. Statin treatment was documented. We know this because the LDL cholesterol was well controlled at baseline. In terms of the actual numbers needed to get into the trial, the triglycerides had to be at least mildly or moderately elevated, a level of about 135 mg/dL or above. I say "about" because some patients with normal triglycerides did end up in the trial, as you know, there's some variability to triglyceride measurement, and the LDL had to be between 40 and 100 milligrams per deciliter, which again, as I mentioned before, illustrates good cholesterol control. These patients, in addition to receiving dietary counseling, in addition to being on a statin, were then randomized to either icosapent ethyl 4 grams a day, 2 grams twice a day specifically, or to a matching placebo. The placebo we picked was pharmaceutical-grade mineral oil given at a rather small quantity of 2 cc twice a day. We picked that particular placebo because we wanted to match the colorless and odorless icosapent ethyl capsules

such that the trial was truly blinded, and unlike, say, prior trials of icosapent ethyl, JELIS,

that were unblinded and open-label trials. That's what led to the considerations of trial design. The patients were followed for an average of 4.9 years. The primary endpoint was five-point MACE, major adverse cardiovascular events.

That endpoint is shown here, as you can see, a significant reduction in MACE, about a 25% relative, about a 5% absolute risk reduction, a number needed to treat of only 21, highly significant P-value.

The key secondary endpoint of cardiovascular death, MI, stroke, the triple ischemic endpoint, hard MACE sometimes people call it, was also significantly reduced with a 26% relative risk reduction. Again, a large absolute risk reduction of about 4%, a number needed to treat of 28, once more a low NNT, and a highly significant P-value. The overall trial robustly positive for the primary and key secondary endpoints.

There was consistency across multiple pre-specified subgroups for the primary and also for the key secondary endpoint, including consistent benefits in the 10% of patients with normal triglycerides. As I mentioned, some patients did sneak into the trial with normal triglycerides. Again, that's just because the triglyceride measurements can be quite variable, we had two different measurements we averaged.

We examined a number of individual and composite endpoints in a pre-specified sequence, and the bottom line is everything in the green box was statistically significantly reduced, and in the blue box, all-cause mortality, there was a trend towards that being lower, which I think would have been significant if we could have followed the patients a bit longer. I say that because cardiovascular death was significantly reduced by 20%, and non-cardiovascular death, well, there, the hazard ratio was 1, showing neither benefit nor harm. Presumably, the reduction in cardiovascular death should have translated into reduction in all-cause mortality had we just kept following the patients a bit longer, accrued a few more events. That's the main REDUCE-IT trial results as published in the New England Journal of Medicine.

What I'm showing you here is a subsequent analysis where we looked at first events, that's what's in green here, the 25% relative risk reduction, and we looked at recurrent events, that's what is in orange. Sometimes people say subsequent events or recurrent events. When you add first and recurrent events, you get total events. This is a total events analysis looking not just at the first non-fatal MI but also looking at the subsequent ischemic stroke or bypass surgery or whatever the event might be, with some patients having even four or more repeat events. What we find is that overall somewhere between 30% to 32% reduction in total events with P-values that are extremely significant, 10 to the -22, by our pre-specified methodology, so large relative, also large absolute risk reductions with about 500 fewer events in the 4000 treated patients.

As far as side effects, overall, the drug was tolerated as well as the placebo, and the placebo was tolerated as well as the drug with no significant difference in overall rates of treatment-emergent adverse events, and this tells us the drug is very well tolerated, it's very safe. For anyone that was at all concerned about our choice of placebo here, it also tells us that the placebo is quite safe.

We also looked more deeply at potential side effects based on the prior literature and found a significant increase in minor but not major bleeding with icosapent ethyl. No significant increases in intracranial bleeding, in fatal bleeding, in gastrointestinal or genitourinary bleeding. This is including in patients on anticoagulants, on dual antiplatelet therapy. As far as bad bleeding, we didn't see any evidence of that. We did see a significant increase in minor bleeding, however, but this is mostly things like skin bruising.

We also found a significant increase in hospitalization for atrial fibrillation or flutter. That went from 2.1% to 3.1%, again, a statistically significant difference. However, for really bad Afib, serious Afib, Aflutter treatment-emergent adverse events, that rate was 0.5% in each arm, so virtually identical.

It wasn't really the bad sort of Afib or sequelae that were an issue here, it was more hospitalization for Afib or flutter. I'll point out that in the overall trial and in the subgroups of patients who had atrial fibrillation at baseline or who developed it in the trial, the rates of stroke weren't increased. In fact in the overall trial, there was a significant 28% reduction in stroke.

Now, very quickly let me share with you a number of subsequent analyses beyond the main results from REDUCE-IT. Let me start with this analysis by triglyceride tertile showing a consistent and indeed significant benefit of icosapent ethyl versus placebo across the full range of triglycerides enrolled in the study.

We also examined the benefit of icosapent ethyl versus placebo as a function of a variety of baseline biomarkers, triglycerides at various cut points, LDL, CRP. Bottom line is consistency of benefit regardless of what the baseline biomarkers showed. I will point out that if we look at the tertile of patients in the lowest tertile of LDL cholesterol at baseline, there as well is a consistent benefit favoring icosapent ethyl versus placebo, with no signal of attenuation of benefit.

We also in a very recent publication looked at the effects of icosapent ethyl versus placebo as a function of baseline statin, finding consistent benefits regardless of whether patients were on statins such as rosuvastatin or atorvastatin, or statins such as pravastatin or simvastatin. Of course, ideally I would have said these patients should be on high-intensity statins, but not every patient can tolerate that, so we studied exactly what the impact of baseline statin might be, and the good news is

if you're using high-intensity, medium-intensity statin, in either case, there is a consistent benefit of the drug, though once more I'll say high-statin, intense-statin is the way to go, but in those subgroups, if anything, the data once more looks stronger. It certainly appears that combining a potent statin with icosapent ethyl is a very promising strategy to reduce global cardiovascular risk.

Another part of this analysis is that we looked at the lipophilicity of the statin, hydrophobic or hydrophilic, and found consistent benefits regardless of which statin subtype the statin fell under. Why do I even mention this? Well, again, some people have said, "Is there any possibility that the pharmaceutical-grade mineral oil placebo used in this trial was interfering with statin uptake and that might have led to differences in LDL in the treatment arms not due to the drug but due to the statin interaction?" Well, if such an interaction were occurring, I would have expected the lipophilicity of the statin to influence it, and there's no evidence of that. I think this is further evidence that the choice of placebo here didn't have anything to do with our very robust trial results.

Let me share with you quickly the results from REDUCE-IT EPA. We looked at achieved on treatment levels of EPA and found significant correlations, associations with lower rates of ischemic events, and the 400% increase in EPA levels in this trial I think accounts for the bulk of benefit that we see with this drug.

Here, let's look across three different trials, the STRENGTH trial, which examined a mixture of EPA and DHA at high dose, 4 grams a day, in a carboxylic acid formulation, and JELIS and REDUCE-IT, which both looked at prescription icosapent ethyl, a pure eicosapentaenoic acid or EPA sort of drug and a prescription drug. With JELIS, it was a dose of 1.8 grams a day. With REDUCE-IT, it was a dose of 4 grams a day. Higher dose of use of icosapent ethyl in the Western REDUCE-IT trial versus the Japanese JELIS trial. The bottom line here is that in each of those trials, JELIS or REDUCE-IT, the achieved levels of EPA were about 170, though the starting levels in JELIS in that Japanese population were higher than in the Western REDUCE-IT trial. What is interesting is that in the STRENGTH trial the achieved levels of EPA were where the Japanese patients started in JELIS. Yes, the EPA levels did go up in STRENGTH and in some patients went up even more than the average levels here, but it didn't attain the levels that we're seeing in JELIS or REDUCE-IT. Part of the reason JELIS and REDUCE-IT were both positive and STRENGTH was negative, I believe, is the use of pure EPA. The ethyl ester formulation versus a carboxylic acid might have also had something to do with it.

Let me share with you now a REDUCE-IT REVASC, where we looked at the endpoint of coronary revascularization and found a significant benefit of icosapent ethyl versus placebo, 34% relative risk reduction, and the time to this benefit was pretty quick. That is, by about a year, these patients, randomized icosapent ethyl, had a consistently statistically significant benefit, actually by 11 months to be precise. This shows the early benefit on this particular endpoint.

We also examined elective, urgent, and emergent revascularization, all three categories were significantly reduced with icosapent ethyl versus placebo, the need for PCI and even the need for CABG was significantly reduced with icosapent ethyl.

Now, let me just share with you some final subgroup data for three different clinically easily identifiable subgroups, REDUCE-IT PCI examined patients with a history of PCI, finding a significant reduction in the primary as well as the key secondary endpoint.

REDUCE-IT CABG looked at patients with a history of previous CABG, 1800 patients. That's larger than most bypass surgery trials, and here as well for the primary endpoint and also the key secondary endpoint, significant reductions. Again, 30% or so relative risk reductions. I'll point out these are also large absolute risk reductions. In this particular case, is 6% absolute risk reduction in CV death, MI, stroke. That really is something I think it is clinically significant.

The final subgroup data I'll share with you are the patients with a prior MI from REDUCE-IT, where if one looks at first events, total events, the primary endpoint, the key secondary endpoint, in all cases, large and significant reductions.

As well, you can see here in the prior MI patients a lower rate of total mortality, cardiovascular death, sudden cardiac death, and cardiac arrest with icosapent ethyl versus placebo, and significant reductions in sudden cardiac death as well as cardiac arrest, takes about four years for this endpoint to manifest as being statistically significant. Different endpoints are influenced at different times. For an endpoint like sudden cardiac death, it takes about four years of treatment for the effects to really fully kick in.

I'll mention at the European Society of Cardiology in 2022, I presented a REDUCE-IT late breaker, where we showed a 40% reduction in ST elevation MI, and we also showed about a 50% reduction in large myocardial infarctions, large as assessed by multiple fold troponin elevation, say, a 50x or more times the upper limit of normal. In REDUCE-IT, we see that icosapent ethyl is reducing endpoints such as ischemic stroke, cardiovascular death, and now we know even endpoints such as ST elevation MI.

With that, I'd like to thank you for your attention. I hope this rapid but I hope you agree comprehensive review of some of the old REDUCE-IT data and some of the very new hot-off-the-presses REDUCE-IT data was useful to you. Thank you so much.