SILVER SPRING, Md. -- An FDA advisory committee voted 9-4 against recommending approval for a not-so-new antidepressant, gepirone hydrochloride extended release (ER), that has already been denied three times by the agency.
The Psychopharmacologic Drugs Advisory Committee (PDAC) did agree, in an 11-2 vote, that the drug is adequately safe. But most members indicated that "substantial evidence of effectiveness" in treating major depressive disorder (MDD) was lacking.
Gepirone's current sponsor is Houston-based Fabre Kramer Pharmaceuticals -- the third company to own rights to the drug in its roughly 25-year history. The panel's negative vote follows three previous complete response letters sent to sponsors, which included Bristol-Myers Squibb and Organon. (GlaxoSmithKline also licensed the drug for a period.)
Gepirone ER, of the azapirone drug family, targets a different receptor than most first-line antidepressants, a fact that the sponsor suggested may lead to a reduce risk of certain side effects -- including sexual dysfunction -- associated with selective serotonin reuptake inhibitors.
But these reduced risks will mean very little if the drug fails to relieve depression. There are currently 18 drugs available to treat MDD.
Much of the meeting was taken up in discussion of a broader issue: the FDA's evidentiary standards for weighing new drug approvals. The FDA had asked the panel to offer insights on a formal dispute resolution requested by the sponsor in April 2014 and approved by the agency in January 2015.
The FDA called on PDAC to resolve a complex concern: how to evaluate a drug that meets the traditional evidentiary standard -- two successful adequate, well-designed control trials -- when that drug also has a significant number of failed or negative trials. In other words, the agency wanted to know, "how much and what type of 'negative evidence' from other negative or failed trials would it take to undermine a finding of substantial evidence of effectiveness?"
Additionally, the panel was charged with discussing possible strategies for "synthesizing evidence across positive and negative/failed trials" and determining whether post hoc analyses that use primary endpoints different than those prospectively specified in a study produces valuable information.
On the first issue of how best to determine "substantial evidence of effectiveness," , deputy director for clinical sciences at the FDA, explained his reluctance to accept the two positive trials on their own.
"If it's two positive [studies] out of 14 or 12, there is a certain chance that the appearance of a positive result is a chance occurrence," Temple said.
Indeed, only two of gepirone ER's 12 short-term clinical trials were positive, according to both the FDA and the sponsor. Both the sponsor and the agency agreed that another three trials had failed and were therefore "uninformative." But things start to get hairy when the agency and the sponsor address the seven remaining short-term trials.
FDA staff noted that in depression trials, 50% of "known effective drugs" still fail to show an effect. This may be due to the placebo effect or the simple fact that depression is cyclical -- i.e., people recover on their own, at least for a time. Because of this high failure rate, the agency recommends using studies with active controls in addition to placebo to help distinguish between ineffective drugs and ineffective study endpoints.
Once again, the agency and sponsor reached consensus on two positive and three failed trials out of 12. The core dispute centered on four of the seven remaining short-term trials, all of which used an active comparator (fluoxetine, paroxetine, or imipramine), per the agency's recommendation. In these trials, gepirone's performance was "numerically worse than placebo" while the active comparator was "numerically better than placebo," the FDA said.
However, the FDA's review team reached this conclusion in a post hoc analysis that employed a depression rating scale, the , that was not the prespecified primary endpoint in some of the trials, but a secondary one. Four used either the HAMD-25 or the Montgomery Asberg Depression Rating Scale (MADRS) as their primary endpoint.
In three of these trials using HAMD-25 or MADRS scores as the primary endpoint, the active control also failed to show benefit.
Panel member , adjunct professor of psychiatry at Johns Hopkins University School of Medicine in Baltimore -- one of the minority favoring approval for gepirone ER -- expressed frustration at the agency's decision to change a primary endpoint in ad hoc analysis.
"The first thing I learned about the FDA is that you've got to call the shot or the ball comes back out," he said, adding that he wished the agency had not performed its post hoc analysis at all.
Temple noted that "HAMD-17 was the one common denominator" among all of the trials, and was not picked at random.
Pickar continued, "This felt like it was tampering after the fact.... I hear your explanation. I don't buy it."
, associate professor of radiology and psychiatry at the University of Pittsburgh, who voted against recommending approval, said that the FDA's approach seemed "very reasonable," given that HAMD-17 was the primary outcome measure in many of the trials.
Panel members also questioned the sponsor's handling of post hoc analyses. In particular, one change redefined the patient population by using a different threshold to qualify patients as depressed, that ultimately excluded half of the patient in two short-term studies.
a professor of biostatistics at the New York University School of Medicine, also voted against approval for gepirone ER.
"It's old data, it's old methods of analysis, there are analytic problems and I think it's not convincing enough to say that it's irrefutable," she said.