More support for Pfizer Inc.’s orphan drug Vyndaqel (tafamidis meglumine) was to be found on FDA’s Peripheral and Central Nervous System Drugs Advisory Committee than might have been expected from the agency’s presentations and background materials questioning whether there was substantial evidence of efficacy against familial amyloid polyneuropathy (FAP).

In back-to-back votes during its May 24 meeting, the panel said by a 13-4 margin that the findings of the pivotal Study 005 were, in FDA’s words, “sufficiently robust to provide substantial evidence of efficacy similar to that usually provided by two supportive studies for a biomarker endpoint that is reasonably likely to predict a clinical benefit,” but, by the exact same margin, that there was not substantial, direct evidence for a clinical benefit.

Russell Katz, director of FDA’s Division of Neurology Products, explained in response to a question from Jeffrey Cohen, Dartmouth Medical School, about the surrogate endpoint question that while the choice of the word “biomarker” may have been an awkward one in the context, “we’re looking for a surrogate, usually a lab test or a clinical endpoint that’s not the real one you care about but predicts the one you do care about.”

Different panel members focused on three such endpoints for which they said there was substantial evidence, although not all of those who voted yes on the question agreed that the evidence was equally strong for all three: small nerve fiber function, muscle weakness and transthyretin (TTR) stabilization (FAP is also called transthyretin amyloid polyneuropathy). These nuances were important enough for the agency that Katz asked the members who had voted yes to state specifically which surrogate endpoints they found most convincing.

Committee members Eric Logigian, University of Rochester School of Medicine and Dentistry, and Michelle Mielke, Mayo Clinic, voted what Logigian called “a weak yes” based on the muscle weakness endpoint. Muscle weakness is one of three subscales of the Neuropathy Impairment Score-Lower Limb scale that was used as one of the two primary endpoints.

According to Pfizer’s background materials, the placebo group experienced “a statistically significant worsening in muscle weakness compared with the tafamidis group. As anticipated for a length-dependent neuropathy, the placebo group demonstrated worse muscle weakness distally with statistically significant differences for the toe and ankle joints (p-values= 0.0091 and 0.0157, respectively) and numerical advantages in the knee joint.

On the other hand, for Samuel Frank, Boston University School of Medicine, the TTR stabilization was “most convincing to me,” as it was for David Preston, University Hospitals-Case Medical Center, who added that he was “not impressed” by the muscle weakness and small nerve fiber results. However, the TTR results were “compelling” and a longer trial could well show evidence of clinical benefit, he said in explaining his yes vote.

Too Flexible?

A minority thought that Study 005 did provide evidence of clinical benefit, if not as strongly as could be hoped. “On a dark night, even small sliver of the moon is better than no moon at all,” argued Robert Clancy, Children’s Hospital of Philadelphia. FAP is invariably fatal if untreated, and the sole available treatment, liver transplant, doesn’t always work either, as the 17 patients, relatives and advocacy group members who spoke during the public comment period pointed out, while also talking passionately about the pain and disability the disorder causes.

Clancy argued that this unmet need in an orphan condition justified showing regulatory “flexibility,” a buzzword that was bandied about throughout the discussion. Last year, the National Organization for Rare Disorders said in a report that most of FDA’s approval decisions for orphan products use unconventional efficacy standards Also see "Orphan Products Mostly Approved With Regulatory Flexibility, NORD "Catalog" Finds" - Pink Sheet, 17 October, 2011..

But for the majority on the committee, the evidence of clinical benefit just wasn’t compelling enough to show such flexibility regarding the results of Study 005. “Flexibility in orphan diseases doesn’t change your need for certainty or at least confidence that what you’re seeing is real,” argued Jeremy Shefner, State University of New York Upstate Medical University. He voted no on both the clinical and surrogate endpoint questions.

The co-primary endpoints in Study 005 were the NIS-LL responder analysis, in which patients who had less than a two-point increase on the scale over 18 months were judged to have no disease progression, and Total Quality-of-Life score change from baseline to month 18. “The protocol-specified primary analysis of both co-primary endpoints (with treated patients who discontinued due to the availability of liver transplants considered non-responders in this analysis) demonstrated that 45.3% of the tafamidis patients in the ITT [intent-to-treat] population had no disease progression at 18 months compared with 29.5% of placebo patients (tafamidis vs placebo, p=0.0682).”

But the study was very small, with only 125 patients in the ITT population (64 randomized to study drug and 61 to placebo). This resulted in an “efficacy evaluable” population of only 87 (45 on study drug and 42 on placebo). “It seems to me ultimately you were underpowered,” committee chair Nathan Fountain, University of Virginia School of Medicine, said.

“I voted no with a heavy heart” on the clinical efficacy question, Preston said. “The drug has low toxicity, but looking at the question, there is no evidence for [clinical] efficacy, certainly not robust evidence. The study was underpowered.”

Pfizer’s Vyndaqel Splits Advisory Committee; Now It Must Move FDA

More from United States

More from North America