Despite lingering concerns over post hoc analyses, palovarotene won the support of an FDA panel as a treatment for abnormal bone growth due to fibrodysplasia ossificans progressiva (FOP).
By a vote of 10-4, members of the agreed the oral selective retinoic acid receptor gamma agonist was effective at reducing the rate of abnormal bone growth, or heterotopic ossification (HO), associated with the ultra-rare disease. The committee also agreed by an 11-3 vote that palovarotene's benefits outweigh its risks.
Committee chair Cecilia Low Wang, MD, of the University of Colorado Anschutz Medical Campus in Aurora, said these analyses, while not ideal, support palovarotene's efficacy.
"Despite problems with post hoc analyses in general, and the fact that these were performed after the data were unblinded, I felt that they were appropriate in this instance," said Wang. "The results were convincing and consistent as performed by the sponsor, and especially because these were confirmed by the FDA."
When sponsor Ipsen first began to study palovarotene, it gathered basic disease information with a natural history study (NHS). The NHS became the comparator for the drug's single phase III trial, , rather than a placebo group. Efficacy was based on data from about 100 participants in each group. This lack of a placebo comparator was a hard pill to swallow for both FDA reviewers and committee members, although there was agreement that the drugmaker controlled for confounders and matched the groups as well as could be done.
But MOVE failed to meet the primary endpoint of a reduction in mean annualized new HO relative to the NHS (median ratio 0.95, 95% CI 0.74-1.22, P=0.65). This was due to a flawed prespecified statistical plan, according to the drugmaker -- something the FDA and committee members agreed was so. But since it's so hard to conduct an FOP trial, given the limited patient population and the struggles with mobility and clinic visits, another phase III wasn't practical.
After data lock, the company conducted six post hoc sensitivity analyses. Only two of those were statistically significant, but all of them, no matter what the statistical method, found that palovarotene reduced new HO by 52% to 77% relative to the NHS patients. The average between-group difference was about 10,000 mm3, but the range of effect size was very large, from just a few hundred to 20,000 mm3.
In its , the FDA came up with similar findings, putting the agency and its committee members in the unusual position of being asked to pass a potentially market-approving judgment based on secondary data.
Robert Greevy Jr., PhD, a biostatistician at Vanderbilt University in Nashville, Tennessee, who voted in support of the drug, pointed this out.
"In different circumstances, this would be the kind of evidence that would make me say 'Oh, wow. Yes, please do the next study' as opposed to 'Yes, we want to approve this medication,'" said Greevy.
Ipsen bolstered its efficacy claim with results from a small group of 17 patients whose palovarotene treatment was interrupted by COVID-19 restrictions, then restarted when the pandemic eased. Over a mean of 20.9 months on treatment, relative to the NHS, the mean annualized HO was about 5,000 mm3. During a 7-month off-treatment phase, the mean annualized new HO soared to 29,767 mm3. But after patients were back on treatment for another 22 months, that rate fell again to 7,728 mm3.
Even the naysayers supported palovarotene provisionally, saying the drug appears to hold promise even if the phase III data weren't strong enough to be considered pivotal.
"By all accounts, and in almost any other scenario, we would consider this to be data that justify a pivotal trial, rather than this being the pivotal trial for a new therapy," said Michael Blaha, MD, a professor of cardiology and epidemiology at Johns Hopkins University in Baltimore. "So I just hope this isn't the last word on this therapy as far as high-quality data."
Palovarotene's proposed indication is for the prevention of HO in adult and pediatric patients with FOP (ages 8 years and older for girls; 10 and up for boys).
On the risk-benefit calculus, two issues appeared most concerning: the suggestion that treatment might increase the risk of a new flare (the inflammatory incident presaging new HO), and palovarotene's effect on normal bone health. Since the drug decreases bone growth signaling, it can also lead to premature physeal closure and reduced vertebral bone mineral density.
Elizabeth Chrischilles, PhD, felt the risk-benefit ratio was positive for palovarotene. She agreed with the sponsor that the NHS flare rate might have been under-reported, leading to a false signal of treatment-related flares.
"There's reason to suspect that the apparent increase in flare rate was either not real, or was real but not clinically meaningful in terms of disease progression," said Chrischilles, a professor of epidemiology at the University of Iowa in Iowa City.
The concerns about compromised bone health could be managed by individual risk-benefit counseling, she said. Ipsen's planned palovarotene treatment registry should help providers identify and manage problems like this.
But a sponsor-run registry is troubling, said Suzanne Robotti, president and founder of the .
"I don't believe registries belong with a pharmaceutical company," she said. "Registries are too important to be housed where there is such a clear conflict of interest in sharing data."
She also suggested that the FDA should enter palovarotene into its program, which is designed to manage the risks associated with isotretinoin.
"We're putting out this drug based on some assumptions," Robotti said. "I ask that the FDA please require confirmatory trials and please require that they be done on a reasonable schedule with clear deadlines and clear endpoints."
While the FDA is not required to follow the advice of its advisory committees, it typically does.