More Positives for PCSK9 inhibitors and the LDL-C Hypothesis

An Interview With ODYSSEY Co-investigator Gregory Schwartz

Interviewer: John M. Mandrola MD  Interviewee: Gregory G. Schwartz, MD, PhD


March 14, 2018

John M. Mandrola, MD: Hi, everyone. This is John Mandrola from | Medscape Cardiology, and I am at the 2018 ACC meeting in Orlando where I am pleased to welcome the co-chair of the ODYSSEY outcomes trial, Dr Greg Schwartz. Dr Schwartz, thanks for being with us.

Gregory G. Schwartz, MD, PhD: Thank you, John.

This is a biologic, it is expensive, so we need to be realistic here and try to find patients who are going to derive the most benefit.

Dr Mandrola: The ODYSSEY outcomes trial is an exciting trial, a late-breaking clinical trial looking at PCSK9 inhibition. Tell us about the background and then we will just get going.

Dr Schwartz: I think the trial is well named because we really have been on an odyssey—for a quarter of a century or more—to find things that worked to reduce the risk of our patients with acute coronary syndrome, who are arguably among our highest-risk patients with coronary disease. Statins have been the cornerstone of that approach for many years. We know that risk after acute coronary syndrome is related to levels of LDL cholesterol (LDL-C) and associated atherogenic lipoproteins. Obviously, statins lower those, and we have seen, without a doubt, that statins reduce that risk.[1]

A few years ago, we had an important proof of concept from the IMPROVE-IT trial,[2] which showed that if you added nonstatin therapy ezetimibe on top of a statin, you had a further lowering of LDL-C and a further diminution or risk, although the effects were small because the agent wasn't that potent. We tried to leverage that proof of concept from IMPROVE-IT, using a much more powerful tool to reduce levels of atherogenic lipoproteins, a PCSK9 antibody, to see whether we could achieve a further reduction in cardiovascular events, morbidity and mortality, after acute coronary syndrome.

Dr Mandrola: You chose patients with acute coronary syndrome, a higher-risk population than in the FOURIER trial.[3]

Dr Schwartz: The FOURIER trial patients were also high risk, and FOURIER, for your listeners, was really a landmark study that established the efficacy of this approach, of PCSK9 inhibition in patients with stable vascular and polyvascular disease. They had specific requirements for risk enhancers in that trial: age, diabetes, etc. We had a more easily described and homogenous population of acute coronary syndrome patients. The risk was a little bit higher than in FOURIER. I would consider both populations high risk but for different reasons or different descriptors.

Dr Mandrola: This was a huge trial; I noticed that it is truly global, isn't it?

Dr Schwartz: Yes, it was a big trial—almost 19,000 patients from 57 countries and six continents. Apart from the scale of the trial, the importance of it being truly global is that it involved people from many different countries, ethnic backgrounds, and healthcare systems. I think that adds to the validity of the results.

Maximal Statin Therapy

Dr Mandrola: You had to have acute coronary syndrome to be enrolled, but you also had to have an abnormal lipid profile, despite statins, and most were on pretty good statin dosing.

Dr Schwartz: Yes, that is correct. Patients had to have one of three lipid criteria fulfilled in order to qualify for the trial: an LDL-C of at least 70 mg/dL (1.8 mmol/L); a non-HDL-C of at least 100 mg/dL; or an apoB level of at least 80 mg/dL. They had to have those levels on top doses of atorvastatin or rosuvastatin, or the maximum tolerated doses of one of those two drugs.

That is slightly different from FOURIER, and as it turned out, 89% of the patients were on one of the top two doses of atorvastatin or rosuvastatin, what we could call intensive statin therapy. In FOURIER, that percentage was a little lower, at 69%. Both trials had a low use of ezetimibe because the trials began, and the patients were enrolled, before the IMPROVE-IT trial had presented its results. We wanted to keep lipid therapy stable after randomization— in both trials; therefore, only 3% of our patients were on ezetimibe.

Dr Mandrola: One quick thing before results is that in ODYSSEY, they also had to target below a certain LDL-C level.

Dr Schwartz: That is absolutely right. We targeted what we consider a low but still physiologic range of LDL-C. We used two different doses of the PCSK9 monoclonal antibody alirocumab, which was the test agent. We used a 75-mg and a 150-mg dose, with blinded titration between those doses to try to maximize the number of patients with LDL-C levels between 25 mg/dL and 50 mg/dL.

We allowed levels to go as low as 15 mg/dL, and in that 15-25 mg/dL range, patients had special safety monitoring by a blinded independent safety monitor. If the levels got below 15 mg/dL on two consecutive determinations on the lower dose of alirocumab, patients were blindly titrated to placebo for the duration of the trial.

One might argue that that could have diminished the efficacy that we saw in the trial, which we will get to. I think this design mimics the decisions that physicians would most likely make in actual application of this drug, where patients and their physicians actually see the lab results (versus a blinded trial). It's unlikely that patients who go down to levels of LDL-C in the single digits would just be allowed to stay, or would choose to stay, on the therapy without any changes made.

Dr Mandrola: Let's get to the results. The primary outcome was a composite.

Dr Schwartz: The primary outcome was a four-part composite, MACE composite. It consisted of coronary heart disease death, nonfatal myocardial infarction, ischemic stroke, and unstable angina, with a very stringent definition of the latter. In the placebo group, in these high-risk patients at 4 years of follow-up, there was almost a 15% occurrence of this composite MACE. That was reduced by 15%, an absolute risk reduction of approximately 1.6% with alirocumab, and that was statistically significant with a high level of significance.

Of the components of the MACE composite, coronary heart disease death was not significantly affected, but nonfatal MI, stroke, and unstable angina all were. There was a fair amount of consistency in the hazard ratios for all of those components.

All-Cause Mortality and the Nominal P Value

Dr Mandrola: You looked at a bunch of key secondary endpoints. I am going to ask you about the order in which you looked at them, statistically.

Dr Schwartz: We had seven main secondary endpoints; they were assessed in a hierarchical sequence. In order to declare statistical significance for category number two, there had to be preceding significance for category number one and then on down the line, for a total of seven of these. The first four were large composites: coronary heart disease events; cardiovascular events; the composite of death, MI, and stroke. All of those were very significantly reduced with alirocumab.

Number five and number six on that list were coronary heart disease death and cardiovascular death. Although there were numerically fewer of those events in the alirocumab arm compared with the placebo arm, the difference did not reach statistical significance—probably because there were relatively fewer of those events in number and there just wasn't the power there.

When we got to number seven, which was all-cause death, it was significantly reduced with alirocumab—again, with a 15% reduction. The P value associated with that comparison alone was .026, but we have to call that nominally significant. In that hierarchical sequence, it followed two assessments that did not reach significance, so we have to qualify that result. Nonetheless, there were indisputably fewer people who died with alirocumab than with placebo.

Dr Mandrola: The next (and perhaps the most talked about) part of this is the subgroup analysis. You chose many subgroups, but the most interesting one is baseline LDL-C. You chose three categories of baseline LDL-C.

Dr Schwartz: This is a lipid-lowering trial, so obviously it makes most biological sense to look at the effective treatment according to baseline levels of LDL-C. We categorized patients roughly in thirds, according to cut points of < 80 mg/dL, 80-100 mg/dL, and then 100 mg/dL or more. What we saw is that the effect was greatest in the patients with LDL-C levels of 100 mg/dL or more at baseline. In those patients, there was a larger reduction in both MACE and mortality, although the analysis of death, according to those subgroups, was post hoc (ie, not prespecified).

Dr Mandrola: When I looked at the < 80 mg/dL and 80-100 mg/dL group, the Kaplan-Meier curves were identical—very close. It almost looks like all of the benefit is in the group with baseline LDL-C levels of 100 mg/dL or more.

Dr Schwartz: Certainly one could interpret the trial in that way. I think there are two ways one could look at a big trial like this. One is that we drew a big circle around a broad population of patients, and within that big circle we found that there was a 15% reduction in terms of a composite MACE and mortality. End the story there and say, if you have a patient who would fit inside that same broad circle, that patient is likely to benefit from the treatment.

If this were a treatment that cost the same as aspirin or a beta-blocker, we probably would take that approach, but it is obviously not. This is a biologic and it is expensive, so we need to be realistic here and try to find patients who are going to derive the most benefit.

In the subgroup with LDL-C of 100 mg/dL or more, the number needed to treat to reduce one MACE event was 29. The number needed to treat to avoid one death was approximately 60 patients. In that subgroup, the numbers needed to treat are fairly robust.

For the trial as a whole, to reduce one MACE event required treating about 63 patients. For reduction of death, about 160 patients had to be treated for the average length of exposure in the trial, which was about 2.8 years. It becomes a more efficient strategy when you try to find the groups that are going to derive greater benefit. That is important when you are dealing with an expensive tool.

Clinical Implications

Dr Mandrola: I talked to a bunch of my partners who take care of more coronary patients than me, and they both said that it is pretty uncommon to have a post-ACS patient on high-dose lipid-lowering therapy, like statins with an LDL-C above 100 mg/dL. How many of these patients are there out there in the real world? My second follow-up is, what about patients who are statin-ineligible, or the statin side effects—what are the implications here?

I think the LDL hypothesis is nailed.

Dr Schwartz: To answer the first question, if you have a patient who can take intensive statin therapy at the highest doses, how many of those patients would still be above 100 mg/dL? That is a relatively small percentage. If we look at the distribution of patients in the PROVE-IT trial[4] way back, 15 years ago, who were assigned to atorvastatin 80 mg, it is only about 13% or 15% that lie to the right of 100 mg/dL.

However, 15% of all patients with acute coronary syndrome is still a lot of people; it is probably 100,000 to 200,000 per year in this country alone. Now, if you add in all of the patients who can't take intensive statin, or perhaps any statin, you get to a much larger number of patients who might still remain above that level. I think that if we had applied ezetimibe to everyone, obviously the numbers staying above those thresholds would have been even less. As I said, at the time the trial was designed, ezetimibe was not yet evidence-based.

Dr Mandrola: The second question I wanted you to think about or answer is, there is still a question about the LDL hypothesis. To me, it seems like this is further evidence that the lower the better for LDL-C; LDL is a critical risk factor. What are your thoughts on this?

Dr Schwartz: I agree with you, John. I think that it is a necessary condition to establish and perpetuate atherosclerosis to have elevated levels of atherogenic lipoproteins. With the data that we have from statins, with the data that we have from IMPROVE-IT, and now data from both FOURIER and this trial, I think the LDL hypothesis is nailed. There is really no light left in terms of doubt that this is not only a key risk factor, but a key target for intervention to reduce the risk of our patients—patients whom we see every day in our practice.

Dr Mandrola: Very good. Thank you for being with us. It was excellent having you.

Dr Schwartz: Likewise. Thank you for having me.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: