ROCKVILLE, Md. -- An FDA advisory panel voted 12-11 Tuesday to recommend expanding the indications for the type 2 diabetes drug empagliflozin (Jardiance) to include prevention of cardiovascular events.
The Endocrine and Metabolic Drug Advisory Committee (EMDAC) drew nearly even on the question of whether a large post-market study offered "substantial evidence to establish that the drug reduced cardiovascular mortality in the population studied." The split vote is effectively a null recommendation.
The vote followed an in-depth discussion of the design and conduct of the EMPA-REG trial, which included criticisms involving interim unblinding, changes to the protocol and to the primary endpoint definition, and duplicative as well as missing data.
On a less controversial question, the panel voted unanimously that the drug did not show an "unacceptable increase in cardiovascular risk."
Empagliflozin is a sodium glucose co-transporter 2 (SGLT-2) inhibitor that lowers blood glucose by increasing urinary sugar excretion.
In August 2014, the FDA approved empagliflozin with the caveat that developer Boehringer Ingelheim conduct a post-market trial in accordance with for all new diabetes drugs. The study was submitted to the agency in November 2014.
The EMPA-REG study was initially intended merely to rule out an increase in cardiovascular risk. However, the sponsor chose to develop a randomized, double blind, placebo-controlled trial with a broader sample of patients -- 7,020 in total -- with type 2 diabetes and heightened cardiovascular risk to concurrently pursue a claim of event reduction.
In its application, the developer proposed the following additional indication for empagliflozin: "In adult patients with type 2 diabetes mellitus and established cardiovascular diseases empagliflozin is indicated to reduce the incidence of cardiovascular death."
EMDAC member , who expressed skepticism about the expanded indication at the start of the meeting, said he was won over by the reduction in cardiovascular deaths, which "withstood all sensitivity analyses, including the worst cases."
"I would hate to deny a claim based on the unique way this trial was run, and how this claim was evolved, because I believe at that, end of the day, the result is true," said Hiatt, of the University of Colorado School of Medicine in Aurora, referring to a study protocol that was amended multiple times.
On the other hand, , at the University of Washington in Seattle, voted not to support "substantial evidence" of reduced cardiovascular deaths.
Heckbert said she was concerned about the number of "inassessible" deaths and the fact that cardiovascular deaths were not the primary endpoint; and she stressed that the agency generally requires two well-designed clinical trials.
"Although these data are intriguing and promising, a second study is needed" before the indication can be considered valid, she said.
In his opening remarks, director of the division of Metabolism and Endocrinology Products at the FDA, noted that while the legal standard for "substantial evidence" of effectiveness calls for "evidence consisting of adequate and well-controlled investigations" -- and that definition has been interpreted by the agency to indicate a requirement of at least two trials -- under certain circumstances, a single study could constitute "substantial evidence."
As the FDA noted, the EMPA-REG study results were the first to show that a type 2 diabetes drug produces a cardiovascular benefit.
The trial's stated primary endpoint compared the drug against placebo while monitoring the time to occurrence of three major adverse cardiovascular events (three-point MACE):
- Adjudicated cardiovascular death
- Nonfatal myocardial infarction (MI)
- Nonfatal stroke
The sponsor's chose "MACE-plus" as its secondary endpoint, incorporating all three-point MACE events as well as hospitalization for unstable angina.
The trial found that empagliflozin was both noninferior and superior to placebo for three-point MACE (HR 0.86, 95% CI 0.74-0.99, P=0.04 for superiority). Results also showed that empagliflozin was noninferior, but not superior, to placebo for the secondary endpoint of MACE-plus (HR 0.89, 95% CI 0.78-1.01, P=0.08).
Importantly, the study showed a relative risk reduction of cardiovascular death of 38% and a relative risk reduction in all-cause morality of 32%. The agency noted that the greatest difference between the two treatment arms were driven by cardiovascular death, which also heavily influenced all-cause death.
However, the FDA's technical staff called much of the data collection and its interpretation into question.
The panel discussion highlighted concerns around missing patient data -- 211 participants discontinued the trial before its completion -- and the decision to exclude silent MI from the primary endpoint. The panel also explored the impact of the study being unblinded for 230 individuals.
"It's kind of inconceivable to me that 230 people can keep their mouths shut for 2 years, or whatever," said EMDAC member . But Konstam, from Tufts Medical Center in Boston, ultimately voted in support of the drug's cardiovascular benefit.
The FDA staff said that the composite primary endpoint initially included silent MIs until a protocol amendment removed them.
"Anytime the sea changes, there are red flags that go up," said , a medical officer for the FDA. Hicks was highly critical of how the sponsor conducted the trial and its data collection processes.
The issue of "inassessible" or undetermined deaths was another problem, Hicks stated, pointing out that "124 patients with an undetermined death is not a small number of patients. When we see this many undetermined deaths, we know that there's missing data."
Nonetheless, Hicks noted that "all of that said, if you do extreme analysis and exclude all of these uninterpretable deaths from the analysis, the effect on cardiovascular deaths is still robust."