An FDA advisory committee's strong recommendation against approval of cardiovascular risk reduction claims for Merck & Co. Inc.'s Zetia (ezetimibe) and Vytorin (ezetimibe/simvastatin) on the basis of the IMPROVE-IT trial puts the agency's senior review officials between a rock and a hard place, and possibly searching for an alternate route.

Approving Merck's efficacy labeling supplements would run counter to the advice of 10 of the Endocrinologic and Metabolic Drugs Advisory Committee's 15 members who vetted the drug Dec. 14.

Those objecting to approval cited ezetimibe's modest magnitude of benefit, doubts about its clinical meaningfulness, and concerns about statistical robustness due to a large amount of missing follow-up data in IMPROVE-IT. Even some who favored approval characterized the effect on CV risk reduction as "weak" and "not overwhelming."

FDA could opt for a middle-of-the-road approach: approve a narrower claim, or add the IMPROVE-IT efficacy data to labeling without a claim.

Yet, if FDA were to deny Merck's sNDAs, the agency essentially would be concluding that IMPROVE-IT – the first trial to demonstrate a CV risk reduction benefit with a lipid-altering agent added on top of statin therapy – did not provide the substantial evidence needed to support a new efficacy claim.

The Division of Metabolism and Endocrinology Products' decision is not likely to be made any easier by the fact that Robert Califf, President Obama's nominee to be the next FDA commissioner, was personally involved in the study. Califf co-chaired the study's steering committee during his time at the Duke Clinical Research Institute.

FDA approval could spur complaints from agency critics alleging that the decision was based on politics, not science. In turn, rejecting Merck's request for a labeling claim could be seen as an embarrassment to the commissioner nominee, who had touted the trial's success before joining FDA in early 2015 as deputy commissioner for medical products and tobacco (see sidebar).

[Sidebar#00151221010]

The review division ultimately could opt for one of the middle-of-the-road approaches suggested by some advisory committee members. These include approving a narrower claim limited to the post-acute coronary syndrome population studied in IMPROVE-IT, or denying a new efficacy claim but adding the trial's efficacy data to labeling ([A#14151214005]).

An Answer To The 'LDL Hypothesis'?

Merck submitted the IMPROVE-IT data in April with labeling supplements requesting claims that Zetia, when administered in combination with a statin, and Vytorin reduce the risk of CV events in patients with coronary heart disease. This would encompass patients with stable chronic disease.

The study's results, first unveiled at the American Heart Association annual meeting in November 2014, have been hailed by many as validating the so-called "LDL hypothesis," establishing that a reduction in LDL-cholesterol is a reliable surrogate for CV benefit with non-statin agents ([A#14141117002]).

This hypothesis underpinned FDA's recent approvals of the first PCSK9 inhibitors, Sanofi and Regeneron Pharmaceuticals Inc.'s Praluent (alirocumab) and Amgen Inc.'s Repatha (evolocumab). However, FDA approved both for narrower populations than the sponsors had requested, deciding that broader indications must await the results of ongoing CV outcomes trials ([A#14150827010]).

Advisory Committee Vote

Do the efficacy and safety data from the IMPROVE-IT trial provide substantial evidence to support approval of a claim that adding ezetimibe to statin therapy reduces the risk of cardiovascular events? Y -10, N - 5

Nevertheless, the validity of the LDL hypothesis had become the subject of much scrutiny in recent years due to a number of unsuccessful outcomes trials with non-statin, lipid-altering agents, including niacin, fenofibrate and investigational CETP inhibitors, as well as two unsuccessful studies with ezetimibe – ENHANCE and SEAS.

While many had been looking to IMPROVE-IT to confirm the LDL hypothesis once and for all, the advisory committee's negative review of the study's results and FDA's ultimate verdict on the CV risk reduction claim for Zetia and Vytorin may well reinvigorate the debate.

'Conclusion On Missing Data Was Missing'

The IMPROVE-IT trial began in 2005 and enrolled 18,144 patients with recent acute coronary syndromes in 39 countries. All subjects were to be followed for a minimum of 2.5 years.

The study's composite primary endpoint was reduction in the risk of the occurrence of: CV death; non-fatal myocardial infarction; documented unstable angina that required hospital admission; all coronary revascularization with either percutaneous coronary intervention or coronary artery bypass grafting occurring at least 30 days after randomized treatment assignment; and non-fatal stroke.

The ezetimibe/simvastatin combination reduced the risk of the primary composite endpoint by 6.4% (HR 0.94; 95% CI, 0.89-0.99, p=0.016) compared to simvastatin alone, with the results driven by fewer non-fatal MIs and non-fatal strokes.

FDA's statistical reviewers raised concerns about the robustness of the results given that approximately 11% of the subjects in the study discontinued follow-up for the primary endpoint early. The agency also questioned the lack of efficacy in two subgroups: non-diabetics and patients under the age of 75 ([A#14151210002]).

To assess the impact of the missing follow-up data, FDA and Merck statisticians presented various sensitivity or "tipping point" analyses, using different assumed event rates and hazard ratios to show the point at which the missing data would tip the results from statistically significant to non-significant for the primary endpoint.

Merck asserted the statistical significance of its primary endpoint results held up under most conservative analyses. "We don't believe missing data is a major concern regarding the validity of the results," Merck Biostatistics Director Paul DeLucca said.

FDA's statistical analysis did not draw any conclusions about the impact of the missing data on the efficacy results, a fact which did not go unnoticed by some advisory committee members.

Participating by phone, committee member William Hiatt, a cardiologist at the University of Colorado, said the sensitivity analyses on missing data were appreciated "but there's really no conclusion from the FDA review as to whether you believe that this is posing a problem in terms of the overall validity of the results or not."

"I find it ironic that your conclusion on missing data was missing," Cedars-Sinai cardiologist Sanjay Kaul said.

FDA statistician Mark Rothmann said both the primary and sensitivity analyses show "something that's not overwhelming. It's sort of fairly borderline, so that's what makes it really difficult to come to a firm conclusion."

Modest Benefits Of Questionable Significance

Although the agency offered no clear conclusions about the impact of the missing data, some panelists said the lost follow-up just added more uncertainty to what was already an efficacy benefit of questionable magnitude.

"I had questions about the clinical meaningfulness of the treatment effect," Kaul said. "I think the totality of evidence suggests that there is potential for missing data to have a material impact on trial interpretation. So when you factor in the effect size, you factor in the robustness – both qualitative as well as quantitative – the clinical meaningfulness and potential impact of missingness, I'm not convinced that this is a robust … clinically meaningful treatment effect."

Kaul noted that a robust 10% risk reduction was seen on an exploratory endpoint comprising CV death, non-fatal MI and non-fatal stroke. This is the same three-pronged "MACE" endpoint favored by FDA for CV safety trials of new diabetes drugs. However, since it was exploratory in IMPROVE-IT it would not be approvable for a claim, Kaul said.

Panel chairman Robert Smith, an endocrinologist at Brown University, said ezetimibe demonstrated a "very modest clinical effect."

"I'm not unenthusiastic about the drug, but … I don’t see substantial clinical benefit, and I have some questions about the strength of the conclusion that this is a statistically significant effect," Smith said.

On the other side of the vote, Hiatt said he supported approval because the study met its endpoint, adding that it's up to the Centers for Medicare and Medicaid Services "and other bodies to think about whether it's worth it or not."

While the benefit is "rather weak and clinically unimpressive … I think it's a positive study that withstands the sensitivity analyses sufficiently to keep it just marginally in a statistically significant range," committee member Hiatt said.

"It makes me a little uncomfortable to take a positive trial on the primary endpoint and sort of disclaim that based on what is substantial [evidence of efficacy] and what is not, because I don't think that's clinically defined," he said.

While the benefit is "rather weak and clinically unimpressive … I think it's a positive study that withstands the sensitivity analyses sufficiently to keep it just marginally in a statistically significant range," Hiatt said.

Another vote for approval came from Brendan Everett, a cardiologist at Brigham and Women's Hospital, who cited the benefit seen on the non-fatal MI and stroke components of the composite primary endpoint.

"This conversation might be very different if the result was being driven by a statistically significant reduction in revascularization, but not MI and not in stroke," Everett said. "I think we'd all feel that at least the clinical relevance of that was less substantial than what we've seen."

"I think it's fair to say …that these benefits, while not overwhelming, are nonetheless clinically relevant to my practice," Everett said. The efficacy data from IMPROVE-IT "are perhaps … the best estimate that we're going to get," he continued. "This is a remarkable feat, a Herculean feat of a trial that is not going to be repeated, so we're not going to be able to answer this question a second time."

Precedent For Labeling Change Without Claim

Most panelists said the benefit seen in IMPROVE-IT should not be extrapolated to the population of patients with stable coronary disease, and panelists on both sides of the approval vote suggested some support for limiting the indication to the post-ACS patients studied in the trial.

"I would advocate for a label endorsing modest cardiovascular risk-lowering in high-risk patients with recent ACS … on top of maximally tolerated statin," said Cleveland Clinic cardiologist Michael Blaha, who voted for approval. He noted this language would be consistent with the labels for the recently approved PCSK9 inhibitors Praluent and Repatha.

Baylor University cardiologist Milton Packer recommended against approval of a CV risk reduction claim but said the outcomes trial data should be described in the Clinical Trials section of labeling. In addition, language in the current indication statement that no incremental benefit has been demonstrated for ezetimibe on CV morbidity and mortality should be removed, he said.

The agency's previous actions on ezetimibe offer some precedent for adding the efficacy data results but not a formal claim.

In 2012, FDA added results from the SHARP study, which demonstrated a reduced risk of major vascular events and major atherosclerotic events in chronic kidney disease patients, to Vytorin's labeling. However, the agency did not allow a claim for risk reduction because the trial did not assess the independent contributions of ezetimibe and simvastatin to the observed effect Also see "The SHARP “Approval”: Merck Gets Vytorin Labeling Change, But Not The CV Outcomes Claim It Sought" - Pink Sheet, 13 February, 2012..

FDA's decision to invoke the combination product rule and deny Merck's request for a formal claim came despite a 16-0 advisory committee vote in favor of approval for patients with pre-dialysis chronic kidney disease Also see "Merck's Vytorin/Zetia Headed For CV Outcomes Claim, But Will Dialysis Patients Be Off-Limits?" - Pink Sheet, 2 November, 2011..

IMPROVE-IT Study: Negative Panel Review Leaves FDA With Tough Choice

More from United States

More from North America