Esperion Therapeutics: Why IMPROVE-IT Hasn't Cast A Shadow Over ETC-1002

| About: Esperion Therapeutics (ESPR)


Failure of IMPROVE-IT doesn't create statistical powering risk for ETC-1002's CVOT.

The bar for a CD risk reduction label wasn't raised with IMPROVE-IT.

The LDL hypothesis still holds true, but is context dependent.

On February 15th, 2016, Merck (NYSE:MRK) received a complete response letter from the FDA for Zetia and Vytorin. The letter denied a proposed label extension for the reduction of cardiovascular risk based on the results of the IMPROVE-IT outcomes study. The reasons for the denial centered on the modest additional benefits shown in patients on Vytorin in comparison to patients on statin monotherpy, along with inconsistencies in the trial's data set. Post rejection, investors began to focus on the reasoning of the EMDAC to find clarity on the decision. IMPROVE-IT was a seven-year, 18,000+ patient population trial that didn't achieve adequate statistical powering, and this might create real doubts about Esperion's (NASDAQ:ESPR) ability to conduct a similar CVOT with only a proposed 12,500 patients. The fear being that if a non-statin drug cannot show clinically and statistically significant results from such a robust trial, then this places an insurmountable mountain for ETC-1002. However, a reasoned examination of the FDA guidance from the EMDAC meetings for IMPROVE-IT, Sanofi (NYSE:SNY) and Regeneron's (NASDAQ:REGN) Praluent and Amgen's (NASDAQ:AMGN) Repatha show that instead of the bar being raised for ETC-1002, the result just means that the FDA won't allow the bar for proving CD risk reduction to be set as low as the standard produced with IMPROVE-IT.

What is the LDL hypothesis?

One of the central developments in the cholesterol-lowering field since the release of IMPROVE-IT and the approval of Praluent and Repatha is that the long-standing LDL hypothesis might no longer be a sufficient basis for approval for non-statin, LDL-C lowering drugs.

The hypothesis is based on the belief that elevated levels of circulating LDL-C in the liver are a major causative factor in the onset of cardiovascular disease (CD). Meta-analysis of nearly 30 trials (equating to approximately 175,000 participants) by the CTT has shown that:

1) Reduction of LDL-C using statin therapy substantially reduces the risk of major vascular events and vascular mortality by about 20% for each 1 mmol/L reduction in LDL-C achieved.

2) Further reductions in LDL-C with more intensive statin therapy produce further reductions in the incidence of adverse events.

As such, LDL-C is recognised as a valid surrogate endpoint marker by the FDA. The concern in the validity of the hypothesis arose from two areas. One was the unfounded confidence placed by the FDA in Zetia/ezetimibe by approving it for LDL-C lowering alone when its first two outcomes studies raised efficacy doubts and a safety concern (ENHANCE and SEAS). The second is that when the non-statin trials are examined together, it proves difficult to make a connection between the relationship of different levels of LDL-lowering relative to the overall treatment intervention. This might raise the question that the effect of a non-statin on top of a statin is producing a different effect than a statin on top of placebo.

What was said at IMPROVE-IT's EMDAC meeting

With IMPROVE-IT, Merck was attempting to rely on long-term outcomes data to broaden Zetia's/Vytorin's label based on an indication for the reduction of cardiovascular events. The committee voted against the label extension, determining that the modest benefit seen in patients using Zetia with simvastatin over using simvastatin alone wasn't enough to show a statistically and clinically meaningful benefit. Vytorin achieved an absolute risk reduction of 2% and a relative risk reduction of 6.4% over the trial period.

The reasoning of the committee members focused on the nature of the benefit produced while the sponsor's construction of the trial was praised. One reason was that the proposed label was too broad and wasn't linked to any cholesterol measurement or criteria, and essentially would allow it to be indicated for use by any patient with coronary heart disease.

"It could be a past MI or it could be something else. It could be someone who has a segmental wall abnormality and no history of an MI. It's just an incredibly ambitious proposal based on the patients enrolled in IMPROVE-IT."

Another reason was that the evidence of treatment effect only emerged 6-12 months after initial treatment during a chronic disease phase, reducing the data set's efficacy. A central reason for the committee's doubt in the strength of the results was that the trial's "design goalposts" were moved several times by the sponsor. At commencement in 2008, the trial's aim was to reduce cardiovascular outcomes in patients with acute coronary syndromes; by 2014, it moved towards improving CD outcomes in patients after coronary syndromes. While the final purpose was aimed at patients with coronary heart disease. This level of "trial creep" raised questions that the selected population was chosen so that it could be "enriched and invested."

Added to this issue was the fact that treatment effect wasn't documented across the whole trial population. Diabetics comprised 20% of the population, and had 36% of the primary events compared to 27% for non-diabetics. Patients older than 75 years comprised only 15% of the population, but had 36% of the primary events compared to 28% for the younger group. The group's entry level values for LDL and CRP were deemed unlikely to account for the differences. This meant that the chances of benefiting from treatment were less for those aged below 75 and non-diabetic. In the past, the FDA has eliminated population cohorts to enhance overall risk/benefit balances, providing that treatment effects are significant in the overall cohort. With IMPROVE-IT, it seemed that the sponsor was selecting a group post-hoc and identifying it as the driver of the treatment benefits.

The translation of the statistical outcome brought home the modest benefit - a 2% reduction in risk for heart attack and stroke translated to a 1 in 350 chance that a very-high risk patient would be prevented from having an adverse event by taking Vytorin for one full year. This small effect is one of the smallest effects ever observed in statin trials.

One commentator present at the meeting summed up the lack of statistical meaningfulness:

"…this study achieved statistical significance simply because of the enormous sample size. In order to show such a small difference, the investigators had to study more than 18,000 patients with a very high baseline risk. In fact, the protocol had to be amended to increase the sample size and extend the length of the trial. If they hadn't made that post hoc change, the tiny absolute benefit of Vytorin would not have been statistically significant…"

What IMPROVE-IT means for ETC-1002

Investors' knee-jerk reaction to IMPROVE-IT was that rejection meant that the bar would be raised for all new LDL-lowering drugs seeking approval for cardiovascular risk reduction. For ETC-1002, this would mean that if it cannot show better outcomes data in its CVOT, then any potential to broaden its statin intolerant label after approval will be nullified. On paper, ETC-1002 should produce strong outcomes data based on its similar mechanism of action to statins, its inertness to metabolism, its pro-drug focus inside the liver and its superior ability to lower hs-CRP. This is only theoretical however and Esperion has to generate the data in its Phase III program to demonstrate these effects. The view also emerged that the FDA had changed its stance on LDL-C as an approvable surrogate endpoint and that results based on LDL-C lowering alone would not be enough for new non-statin drugs to gain approval.

A truer reflection of the briefings shows that the landscape for approval will now be based on context. This new context dependency view will be a positive risk/benefit relationship, showing a safe and well-tolerated drug designed to treat a high unmet medical need, where the drug's LDL-C lowering is expected to impact the clinical endpoint prior to being demonstrated with outcomes data. This is most likely to be what Esperion learned from the end-of-Phase II meeting minutes and why it then pivoted ETC-1002 towards initial approval to treat statin intolerant patients. This position reduces the approval risk and places ETC-1002 inside the new FDA guidance framework. By guiding towards the treatment of statin intolerant patients, the FDA was recognizing statin intolerance as a high unmet medical need with a compelling risk/reward relationship. Approval at the end of ETC-1002's Phase III trials is then more focused on safety and tolerability than demonstrating positive long-term outcomes.


In conclusion, IMPROVE-IT is likely to impact on two important treatment patterns; 1) the use of combination therapies and 2) the push to reach lower LDL-C goals. The FDA is guiding for the use of lower-cost combination therapies where new drugs can complement the existing standards of care. Regardless of the modesty of the IMPROVE-IT results, they showed for the first time that a non-statin added to a statin could further reduce LDL-C and have a positive impact on cardiovascular outcomes. IMPROVE-IT showed that achieving lower levels of LDL-C produced better outcomes for patients. Patients who reached an LDL of 50mg/dl achieved better outcome results than the patients who achieved an LDL of 70mg/dl. With the availability of new combination therapies along with a strong existing low-cost standard of care, physicians are likely to push their patients to achieve lower LDL-C goals.

Disclosure: I am/we are long ESPR.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.