FDA Panel Nixes Novel Alzheimer's Therapy

Pauline Anderson

March 25, 2019

A US Food and Drug Administration (FDA) advisory panel has nixed the approval of a novel non-invasive therapy for mild to moderate Alzheimer's disease (AD).

Known as neuroAD Therapy System (Neuronix), the treatment combines transcranial magnetic stimulation (TMS) with online cognitive training — an intervention already in use elsewhere in the world.

At a recent meeting of the Neurological Devices Panel of the FDA's Medical Devices Advisory Committee, panel members agreed that the device is safe, but that the available evidence fails to show clinically meaningful benefit.

The panel was considering a new application for neuroAD. The company is proposing that the device be used in patients with mild to moderate AD who have a baseline AD Assessment Scale-Cognitive Subscale (ADAS-Cog) score of no more than 30.

In 2016, neuroAD was granted an Expedited Access Pathway (EAP) designation. Such status is intended to speed approval of devices in the treatment of life-threatening, or irreversibly debilitating, diseases to address an unmet need.

Following meetings with the sponsor and a request for additional information, the FDA issued a denial letter to the company last June. The sponsor appealed this decision and, in response, the FDA reopened the submission.

TMS primes specific brain networks to enhance plasticity. Cognitive training engages these primed networks, with training progressing based on individual performance, the sponsor noted

The idea is that the combination of these two approaches results in consolidated benefit over time.

TMS in Widespread Use

TMS is FDA-cleared for major depression, migraine, and obsessive compulsive disorder. It's used worldwide and has been shown to be safe in the elderly.

The neuroAD device is approved and in clinical use in Europe, Israel, and Australia.

During the advisory committee meeting, the sponsor presented data from its pivotal US study. The trial included 130 patients with mild to moderate AD at nine sites in the US and one in Israel. About 80% of subjects were on stable AD medications, including cholinesterase inhibitors and memantine.

Patients underwent structural magnetic resonance imaging (MRI) to mark targeted brain regions and identify excluded disorders and non-AD brain pathology.

Study participants were randomized to receive active TMS and active cognitive training, or sham TMS and sham cognitive training.

The treatment course included five daily, 1-hour sessions per week for 6 weeks. The study had a 6-week follow-up period.

The study was blinded, but each site had two unblinded subjects receiving active treatment. These data were only included in the safety analysis.

The study's primary efficacy endpoint was change from baseline to 7 weeks on the ADAS-Cog. The study failed to meet this endpoint. In fact, at 7 weeks, the between-group difference on ADAS-Cog was 1.45, favoring sham.

"Dramatic Benefit"

A secondary endpoint was change in ADAS-Cog at 12 weeks to assess durability. Here, the study showed a difference of -0.42, favoring active treatment.

Other secondary endpoints included AD Cooperative Study-Clinical Global Impression of Change (ADCS-CGIC) scores at 7 and 12 weeks.

A post-hoc analysis of the US pivotal study used an ADAS-Cog cutoff of 30. About 85% of study participants had a score of 30 or less — a patient population that tends to deteriorate less rapidly than those with higher scores.

In this population, there was a difference of -1.61 points on the ADAS-Cog favoring the active treatment at 12 weeks. More than 40% of active subjects showed at least a 3-point improvement, and more than 70% showed some improvement at 12 weeks compared with baseline.

For the CGIC endpoint, there was also a clinically meaningful benefit, with the active group outperforming the sham by -0.40 points, said the sponsor.

When combining the CGIC and ADAS-Cog endpoints in this population, the active group performed statistically better than the sham group at 12 weeks (P = .046).

About 64% of active subjects improved or had no change on both measures compared to 43% of sham subjects. Only 7% of active subjects deteriorated on both measures compared to 23% of sham subjects.

"NeuroAD delivers dramatic benefit for about a third of patients, and some benefit for an additional one third of the population in addition to standard of care," said Alvaro Pascual-Leone, PhD, MD, professor of neurology and associate dean for Clinical and Translational Science, Harvard Medical School, Boston, Massachusetts, and a principal investigator for one of the sites in the US pivotal study.

Meaningful Benefit, Minimal Risk?

It's unclear why patients continue to improve after therapy stops but company representatives speculated that it may take time for consolidation of its effect.

Sponsor representatives presented data from supplementary studies, including two from Korea, which they maintained showed consistency of the positive effect and a clear benefit for some patients.

The device was shown to be safe. About 14% of the active group had a related adverse event (AE) such as headache, muscle twitching, fatigue, and neck pain. These AEs were mild and transient, and no patient withdrew due to AEs.

The adherence rate was high. Over 90% of subjects attended at least 90% of the in-clinic sessions.

"The totality of evidence strongly supports that there is a meaningful clinical benefit that outweighs the minimal risk," said Marwan Sabbagh, MD, director, Cleveland Clinic Lou Ruvo Center for Brain Health, Las Vegas, Nevada, and a principal investigator at one of the US pivotal study sites.

However, the FDA panel had a different take on the research findings. One concern was that the secondary endpoints only begin to favor the active group at the 12-week mark, 6 weeks after the end of treatment.

FDA officials questioned the use of the ADAS-Cog cutoff point of 30 and suggested that the post-hoc analysis of the US pivotal study should be viewed only as hypothesis-generating.

FDA panel members disagreed that the results of the Korean studies are equivalent to those of the US study. They raised concerns about the small size and high uncertainty level of these and other supplementary studies.

Uncertain Efficacy

Laura Thompson, PhD, a mathematical statistician at the FDA, noted that collectively, the supplemental studies showed different patterns of treatment benefit over sham compared to the US study across follow-up assessments, as well as across the under-30 and over-30 subgroups.

"The key question is whether the enhanced effect in the US subgroup is merely a chance finding from the pivotal study," said Thompson.

She pointed out that the ADAS-Cog cutoff point of 30 was chosen after study data were available and then tested on those data. "So the subgroup results may be inflated and P values not adjusted for multiplicity."

Taken together, "these points lead to uncertainty in concluding device effectiveness in a US population," said Thompson.

Claudette Brooks, MD, neurologist and clinical reviewer with the FDA's Division of Neurological and Physical Medicine Devices, maintained that a change of 1.61 on the ADAS-Cog is not clinically meaningful.

Brooks expressed concerns with using ADAS-Cog to select patients for this treatment.

Asked how long the treatment effects last, Pascual-Leone said that's not clear, as the studies focus on the acute effect. But he noted that in depression, the effects of TMS span an average of about 6 to 9 months.

"This is a completely different disease," stressed David Knopman, MD, professor of neurology at the Mayo Clinic in Rochester, Minnesota. "I think that the absence of any evidence presented about what happens after 12 weeks is a major deficiency."

Need for "Real-World" Assessments

A number of presentations from patients, caregivers, and families during the day-long meeting underlined the clear unmet need for additional effective treatment options for AD. Some families have traveled to countries where the treatment is offered.

During afternoon discussions, panel members determined that there were no significant safety concerns about the US pivotal study, although some members noted that psychiatric side effects were not well characterized and some worried about cognitive worsening in patients with moderate AD. They agreed that there were not enough data in the other trials to decide if there were safety issues.

Many panel members believe the US study did not show clinically meaningful benefit as an adjunct treatment for AD.

"On the prespecified primary outcome and the prespecified three secondary outcomes, there is no significant benefit, nor is there any signal of benefit," said Earl Ray Dorsey, MD, professor of  neurology at the University of Rochester Medical Center, New York.

"Based on that, I don't see how you can conclude anything other than the treatment is not effective."

He argued against using "subjective episodic clinic-administered scales" like the ADAS-Cog.

"The definition of insanity is doing the same thing over and over again, and expecting different results," Dorsey said.

The fact that 99% of clinical trials in AD have failed "is good evidence that we should change the way that we're assessing drugs and the way we are conducting clinical trials," Dorsey added.

Many panel members agreed that instead of using ADAS-Cog to identify patients for therapy, clinicians should use clinically relevant examinations such as the Mini-Mental State Examination (MMSE), a test that's more familiar to clinicians.

"We need to move toward more objective, real-world assessments of AD," said Dorsey. "I'm far more concerned about whether someone can walk down a dock or aisle than whether they can recall 8 words or 10 words in a clinic in an arbitrary time frame and in an artificial setting.

"We have the tools and technologies to measure social engagement, to measure conversations, to measure sociability and activity; we should be using those and stop using scales that we know don't work."

Back to the Drawing Board?

But some panel members, including Wayne Goodman, MD, chair of the department of Psychiatry and Behavioral Sciences at Baylor College of Medicine in Houston, Texas, did see a signal in the post-hoc analysis. He added, though, that it's difficult to "disentangle" how much of the effect was due to cognitive training alone.

He suggested that a future study should look at active cognitive training in both intervention and control groups.

In general, panel members agreed that a clinically meaningful benefit would be a minimal improvement on the ADAS-Cog of 2 and possibly 3 to 5.

The panel concluded that although the post-hoc analysis didn't confirm clinical benefit, it may provide a good starting point for the next study.

"I have been involved with clinical trial design and implementation for two decades, and the one thing I've learned is that when your prespecified primary endpoint is missed, the only value of the post-hoc analysis is designing the next study, and I think that's where we are," said Goodman.

Some panel members said it would be useful to have prespecified time points that extend beyond 12 weeks, perhaps as far out as a year or 18 months.

Many members sympathized with patients who are enormously frustrated about the limitations of current treatments.

"This is a difficult topic, and we want very much for there to be a treatment that's effective," said meeting chair Mary E. Jensen, MD, professor of neurology at the University of Virginia Health Center, Charlottesville.

"We really hope you continue to bring forward trials that give us the data that allow all of us at this table to say resoundingly, 'Yes' " Jensen told FDA reps.

For more Medscape Neurology news, join us on Facebook and Twitter .


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.