Almost every recent advance in Duchenne muscular dystrophy has aimed at the same target: the faulty dystrophin gene. Exon-skipping drugs and gene therapy try to restore some version of the missing protein, and they work only for patients whose mutation matches.
A Phase 3 trial published in The Lancet took a different approach entirely. It left the mutation alone and tried to slow the damage instead, in boys and young men whose disease was already advanced.
The therapy, deramiocel, is made from human donor heart cells and infused into a vein every three months. It met its primary endpoint. On the same day the trial was published, an FDA advisory committee voted against it.
What HOPE-3 Measured and What It Found
HOPE-3 was a multicenter, randomized, double-blind, placebo-controlled Phase 3 trial conducted at neuromuscular centers across the United States. It randomized 106 boys and young men aged 10 and older with advanced Duchenne to quarterly intravenous infusions of deramiocel or placebo over 12 months. The population was largely non-ambulatory, meaning participants had already lost the ability to walk.
The primary endpoint was the Performance of Upper Limb scale, version 2.0, a functional measure of what a person can actually do with their arms and hands. This population is not an abstraction. Upper limb function determines whether someone can feed themselves, groom, or operate a phone or wheelchair controls.
At 12 months, the least-squares mean difference favored deramiocel by 4.55%, with a 95% confidence interval of 0.47 to 8.63 and a p-value of 0.029. The sponsor characterized it as a 54 percent slowing of upper-limb decline relative to placebo. Craig McDonald, chair of physical medicine and rehabilitation at UC Davis Health and the trial's lead investigator, described the finding as evidence of benefit across multiple systems affected by the disease.
No deaths occurred during the study, and serious adverse events were uncommon.
The Cardiac Endpoint Is Where the Argument Is
Duchenne kills primarily through the heart. Cardiomyopathy is the leading cause of death, and it has been the hardest part of the disease to influence.
The trial's key secondary endpoint was left ventricular ejection fraction, a measure of pumping function. This is the number the dispute turns on. Results initially reported by the company showed an absolute treatment difference of roughly 2.4 percentage points with a p value of 0.041. After discussions with the FDA and journal peer reviewers, the company updated the statistical model, resulting in a 1.8 percentage-point difference and a p-value of 0.09. Under FDA's own primary analysis, the endpoint did not reach statistical significance.
Cardiac MRI was used to assess myocardial fibrosis, the scarring that progressively replaces heart muscle. Imaging showed evidence of reduced progression of that scarring, and multiple analyses of cardiac function favored the treated group, suggesting stabilization rather than reversal. Investigators have framed this as potentially the first clinical evidence that the antifibrotic effects observed in preclinical Duchenne models translate into measurable benefits in patients. Those cardiac measures remain secondary and exploratory rather than confirmatory.
A Therapy That Ignores the Mutation
Deramiocel consists of allogeneic cardiosphere-derived cells, meaning cells grown from donated human heart tissue rather than from the patient. They are not intended to graft into muscle and become part of it.
The proposed mechanism is indirect. The cells are thought to release extracellular vesicles, sometimes called exosomes, that shift immune cell behavior toward a tissue-healing pattern and dampen fibrosis.
That design has a specific practical consequence: the approach is agnostic to the precise underlying genetic lesion. A patient whose mutation makes them ineligible for exon-skipping therapy is not excluded on genetic grounds.
It is also delivered as a quarterly outpatient infusion rather than a one-time procedure, which is a meaningfully different logistical proposition for families than gene therapy.
What Happened at the Advisory Committee
The FDA's Cellular, Tissue and Gene Therapies Advisory Committee took up the application on July 29 to consider its use in treating cardiomyopathy in Duchenne. It voted 9 to 3, with no abstentions, that the available evidence did not provide substantial evidence of effectiveness.
The vote is non-binding, and the committee's question addressed a narrower indication than the company originally proposed. But the reasoning matters. As Medscape reported, members who voted no cited concerns that results on the ejection fraction endpoint, and to a lesser extent the upper limb endpoint, appeared highly sensitive to how missing data were handled and which analytic assumptions were applied. FDA reviewers laid out a history of four successive versions of the statistical analysis plan and relied on the earliest one for their own analysis. Committee feedback on upper limb function was described by the company as directionally supportive.
This is a second-cycle setback. The application previously received a Complete Response Letter, the FDA's formal notice that an application cannot be approved in its present form, and HOPE-3 data were incorporated into the resubmission accepted earlier this year.
The Caveats, and What Happens Next
The trial ran 12 months in 106 participants. That is a substantial study for a rare disease and a small one by the standards of most drug approvals.
The confidence interval for the primary endpoint approaches zero at its lower bound, suggesting the true effect could be considerably smaller than the point estimate. Whether a 4.55 percentage point difference on a functional scale over one year translates into a difference families notice in daily life is a separate question the trial was not designed to answer, and durability beyond 12 months is unknown.
The trial was sponsored by Capricor Therapeutics, the company seeking approval. Peer review in The Lancet provides independent scrutiny of the design and statistics, which the company has emphasized, and it does not remove the sponsorship.
The Prescription Drug User Fee Act target action date is August 22, 2026. FDA is not bound by the committee's recommendation, though it is unusual for the agency to approve a product after a negative vote.
Duchenne muscular dystrophy affects roughly 1 in 3,500 to 5,000 boys worldwide. There is no cure, and no therapy stops the disease.
Families weighing what this means should discuss it with a neuromuscular specialist rather than relying on trial numbers alone, and should be aware that a target action date is a deadline for a decision, not a guarantee of what the decision will be.
Key Questions Answered
What is deramiocel? It is an investigational cell therapy made from allogeneic cardiosphere-derived cells, meaning cells grown from donated human heart tissue. It is given as an intravenous infusion every three months and is designed to reduce inflammation and fibrosis rather than to correct the dystrophin mutation.
What did the HOPE-3 trial show? In 106 boys and young men with advanced Duchenne, deramiocel met its primary endpoint on the Performance of Upper Limb 2.0 scale, with a least-squares mean difference of 4.55 percent, a confidence interval of 0.47 to 8.63, and a p-value of 0.029. No deaths occurred, and serious adverse events were uncommon.
Did the trial show a heart benefit? That is contested. The key secondary cardiac endpoint, left ventricular ejection fraction, did not reach statistical significance under FDA's primary analysis, with a p-value of about 0.09 after the statistical model was updated. Cardiac MRI showed reduced progression of myocardial scarring, but those measures remain secondary.
Why did the FDA advisory committee vote against it? The committee voted 9 to 3 that the evidence did not provide substantial evidence of effectiveness for treating Duchenne cardiomyopathy. Members cited concerns that the cardiac results, and to a lesser extent the upper limb results, appeared sensitive to how missing data were handled and which statistical assumptions were used.
Does that vote decide the outcome? No. Advisory committee votes are non-binding, and the committee's question covered a narrower indication than the company proposed. FDA has a target action date of August 22, 2026. It is unusual, though not impossible, for the agency to approve a product after a negative vote.
Who could this therapy potentially help? Because it does not target a specific genetic mutation, the approach is not limited to patients whose mutation matches an exon-skipping drug. The trial population was largely non-ambulatory boys and young men aged 10 and older.
What should families do with this information? Discuss it with a neuromuscular specialist rather than relying solely on trial numbers. A target action date is a deadline for a decision, not a guarantee of approval, and the durability of any benefit beyond 12 months is unknown.
