Duchenne Exon Skipping, From Exondys 51's 0.44% Dystrophin to Del-zota's 25%

On April 25, 2016, a panel of outside experts advising the FDA voted 7 to 6 against approving a drug for Duchenne muscular dystrophy. The 6 votes in favor included the panel's consumer representative and both of its patient representatives. The drug, eteplirsen, had been tested for dystrophin in a 12-boy study. In 8 of them the protein rose by 0.25 percentage points or less. One boy gained more than 1 point. The average went from 0.16% of a healthy person's dystrophin to 0.44% (FDA, 2016 summary review; Exondys 51 label).

Hold on to that number, 0.44%. Five months later the FDA approved the drug anyway, as Exondys 51, over the written objections of its own review team. Everything that has happened in Duchenne since, the 3 sister drugs, the gene therapy that was pulled back after 3 deaths, the confirmatory trial that missed in November 2025, and the 2 new drugs now sitting in front of the FDA reporting 2.87% and 25%, is a 10-year argument about whether 0.44% was enough and what number would be.

What Dystrophin Does and Why 0.44% Split the FDA

Dystrophin is a shock absorber. It is a long protein that bolts the inside of each muscle fiber to the membrane around it, so that when the fiber contracts the membrane does not tear. Boys with Duchenne make almost none of it. Every contraction causes a little damage, the damage outruns repair, and muscle is slowly replaced by fat and scar. Most boys stop walking in their early teens, and the heart and breathing muscles follow.

The reason anyone cares about a fraction of a percent is a related, milder disease. Becker muscular dystrophy comes from mutations in the same gene that leave a shortened but working dystrophin. In 1988 the lab that discovered the protein reported that boys with the severe Duchenne picture had less than 3% of normal dystrophin or none, while boys with Becker had abnormal but present protein (Hoffman et al., 1988). A 2014 Dutch study of 33 Becker patients with dystrophin anywhere from 3% to 78% of normal found that every patient under about 10% had a severe course, and that above roughly 10% the exact level stopped predicting how weak someone would be (van den Bergen et al., 2014). Families in which men carry 29% to 57% of normal dystrophin have no skeletal muscle weakness at all, though the same mutations affect their hearts (Neri et al., 2007).

Those 3 studies draw the zones on the chart you will see later in this story. Under 3% is Duchenne. Three to 10% is severe Becker. Ten to 30% is Becker's milder range. Above 30% there may be no weakness. In 2016 the FDA was looking at a drug that moved boys from 0.16% to 0.44%, and the question was whether the first rung of that ladder counted as a step.

“There often is little consideration of another error, which is failing to approve a drug that actually works. In devastating diseases, the consequences of this mistake can be extreme, but most of these consequences are borne by patients.”

Janet Woodcock, then head of the FDA's drug center, as quoted in the agency's 2016 Exondys 51 decision record

Woodcock overruled her reviewers and granted accelerated approval, a pathway that lets a drug onto the market on a stand-in measurement, here dystrophin, on the condition that a later trial proves it helps patients. Her acting chief scientist filed a dissent saying the evidence did "not support that the levels of dystrophin produced by eteplirsen at the doses studied are reasonably likely to provide clinical benefit" (FDA, 2016 summary review). Both of them turned out to be describing the next 10 years.

The dystrophin gene is the largest in the human body, 79 exons long. An exon is a segment of the gene that gets stitched into the final instructions for the protein; think of each one as a word in a 79-word sentence. In about 68% of boys with Duchenne, one or more words are missing (Bladen et al., 2015). Whether that deletion causes Duchenne or the milder Becker disease depends on something that sounds like a technicality and decides a life: whether the words that remain still line up in groups of 3 letters. If they do, the cell reads a shorter sentence and makes a shorter dystrophin. If they do not, the reading frame is broken, the cell hits a stop sign a few words later, and it makes nothing.

Exon skipping is a patch on the sentence. An antisense oligonucleotide, a short synthetic strand of genetic code, sticks to one more exon next to the deletion and hides it from the cell's splicing machinery, so that exon is left out too. If leaving out the right exon restores the groups of 3, the cell can make a Becker-like dystrophin. The trick is that each drug can hide only one specific exon, so each drug helps only the boys whose deletions sit next to it.

Add the 4 skippable exons together and they cover between 36% and 39% of all Duchenne mutations, depending on whose count you use. The other 61% to 64% of boys, those with duplications, small mutations or deletions in the wrong place, have never had an exon skipping drug to take, and that is the group the gene therapy later in this story was built for.

“PPMD is thrilled that FDA has granted an Accelerated Approval to Exondys 51, a therapy with the potential to treat 13% of people living with Duchenne, marking the first-ever U.S. approval of a drug to treat this disease.”

Pat Furlong, founder of Parent Project Muscular Dystrophy, who lost both of her sons to Duchenne, September 19, 2016

A Boy Born in 2011, Diagnosed at Age 4.9, and the 2.2-Year Duchenne Diagnosis Delay

Picture a boy born in 2011. He is imaginary, but his dates are not. The CDC's muscular dystrophy surveillance network followed 221 American boys with Duchenne and no family history and found that parents first noticed something wrong, often late walking, frequent falls or trouble with stairs, at an average age of 2.7 years. A first blood test for creatine kinase, the muscle enzyme that leaks when fibers are damaged, did not come until 4.6 years, and genetic confirmation came at 4.9. That gap between first signs and a diagnosis, 2.2 years, had not shrunk at all between 2000 and 2015 (Thomas et al., 2022). A study a decade earlier had found the same 2.5-year gap, unchanged for the 2 decades before that (Ciafaloni et al., 2009).

So this boy's parents notice the falls in 2014. His pediatrician reassures them for a while, because that is what the 2.2 years mostly are. His CK comes back in the tens of thousands in 2015 and a genetic test in the spring of 2016 names a deletion of exons 45 to 50. That deletion is skippable with exon 51. He is diagnosed in the same year, nearly the same season, as the 7 to 6 vote.

Change one number on the report and his story forks. A deletion of exon 45 alone is amenable to exon 44 skipping, and nothing on the market has ever been for him, though a drug for him is under FDA review as you read this. Keep both versions of the boy in mind, the exon 51 version and the exon 44 version, because they end the decade in very different places.

Exondys 51, Vyondys 53, Viltepso and Amondys 45: 4 Approvals, 4 Dystrophin Numbers

The exon 51 version of the boy starts weekly infusions of Exondys 51 in the fall of 2016, at age 5. The label calls for 30 mg per kilogram into a vein once a week, 35 to 60 minutes each, through a filter, for as long as he takes it (Exondys 51 label). He gets a port. His parents learn to schedule their lives around Thursdays.

The sister drugs followed, each approved the same way, on dystrophin, with a clinical trial still owed. Vyondys 53 (golodirsen) in December 2019 took 25 boys from 0.10% to 1.02% of normal at 48 weeks (Vyondys 53 label). Viltepso (viltolarsen), a Japanese drug for the same exon, in August 2020, took 8 boys from 0.6% to 5.9% at 25 weeks, the best number of the group, at a dose of 80 mg per kilogram (Viltepso label). Amondys 45 (casimersen) in February 2021 took 27 boys from 0.93% to 1.74%, against a placebo group that drifted from 0.54% to 0.76% (Amondys 45 label).

Look at where the dots sit. Three of the 4 drugs, on average, left boys inside the band that defines Duchenne. The exception was tested in 8 children. None of the 4 has a confirmed effect on walking, breathing or strength; Exondys 51's own label says its 24-week placebo-controlled study found no difference in how far boys could walk in 6 minutes, and that a longer comparison against boys not on the drug "failed to provide evidence of a clinical benefit" (Exondys 51 label).

The price did not wait for the proof. In 2019 the Institute for Clinical and Economic Review put eteplirsen's cost at more than $1 million a year and said it could not calculate a price at which the drug would be worth it, because there was no established benefit to put a value on (ICER, 2019). By October 2026 the exon 51 boy is 15. If he has stayed on the drug the whole time, he has had roughly 500 infusions. The label's best long-term number, from 11 boys biopsied after 180 weeks, is 0.93% of normal dystrophin (Exondys 51 label).

ESSENCE Missed in November 2025 and Elevidys Got a Boxed Warning: Duchenne's Confirmation Problem

Accelerated approval is a loan, and in 2025 the loans on 2 very different Duchenne drugs came due in the same season.

The first was ESSENCE, the trial Sarepta owed the FDA for Vyondys 53 and Amondys 45. It enrolled 225 boys aged 6 to 13 with exon 45 or exon 53 deletions and followed them for 96 weeks, measuring how fast they could climb 4 steps. On November 3, 2025, Sarepta reported that boys on drug climbed 0.05 steps per second faster than boys on placebo, a gap small enough to be chance (p = 0.309). The trial missed. A reanalysis that threw out boys whose assessments were disrupted by COVID found a bigger gap that still did not reach significance (Sarepta, Q3 2025 results). Sarepta filed anyway, in April 2026, to convert both drugs to full approval, and the FDA accepted the applications with a decision due February 28, 2027 (decision page).

The second was Elevidys, Sarepta's gene therapy, which had taken the other road to dystrophin: instead of patching the boy's own gene, it delivers a shortened "micro-dystrophin" gene inside a virus, once, by infusion, to boys with a confirmed mutation, whatever the exon. On protein it looked like a different sport. Boys produced micro-dystrophin at 34% to 51% of a healthy person's normal dystrophin level 12 weeks after the infusion, against 0% on placebo (Elevidys label). On function, its 125-boy EMBARK trial missed its primary endpoint, a 17-item motor scale called the North Star Ambulatory Assessment, by 0.65 points with p = 0.24, while 3 timed tests favored the drug (Elevidys label). The FDA gave it accelerated approval for 4- and 5-year-olds in June 2023 and full approval for walking boys 4 and up in June 2024.

Then the year turned. On June 24, 2025, the FDA announced it was investigating 2 deaths from acute liver failure in non-walking patients who had received Elevidys. On July 18 it asked Sarepta to stop shipping the drug entirely, after a third death in a trial of a different Sarepta gene therapy built on the same virus, and said in its press release that the company had refused. Ten days later the agency let shipments resume for boys who could still walk. On November 14, 2025, it added a boxed warning for acute liver failure, limited the label to ambulatory patients, and required a 200-patient safety study (FDA press announcements, June 24, July 18 and November 14, 2025; Elevidys label). The current label also bans the drug outright in boys with deletions in exons 8 or 9, after 2 of 6 such boys developed immune attacks on their muscle.

Why the First Exon Skipping Drugs Barely Reached Muscle and How Antibody Conjugates Fixed Delivery

The idea behind the old drugs was sound; getting them into muscle was the problem. A phosphorodiamidate morpholino oligomer, the chemical form of all 4 approved drugs, is a short strand of synthetic code with no address on it. Infused into a vein, most of it goes to the kidneys and out in the urine within hours, which is why these drugs are given weekly, at high doses, and why 3 of the 4 labels carry a kidney toxicity warning. Only a tiny fraction gets into muscle cells, and the dystrophin numbers above are what a tiny fraction buys.

The 2 drugs now in front of the FDA put an address on the package. Muscle cells are hungry for iron, so they carry a surface protein called transferrin receptor 1 that pulls iron into the cell. Both Dyne Therapeutics and Avidity Biosciences bolted an antibody that grips that receptor, a fragment of one in Dyne's case and a whole antibody in Avidity's, onto the same kind of morpholino strand. The muscle cell sees the antibody, swallows it, and the drug comes along for the ride. Dyne calls its version a FORCE conjugate; Avidity calls its version an antibody oligonucleotide conjugate, or AOC. The chemistry doing the exon skipping is the same class the FDA approved in 2016, and the delivery is the only thing that changed.

Zeleciment Rostudirsen (DYNE-251) Exon 51 Results and the January 21st, 2027 FDA Decision

Dyne's dystrophin result at 6 months

The exon 51 boy, the one with 500 infusions behind him, is who this drug is for. In the registrational part of Dyne's DELIVER trial, 24 boys aged 4 to 16 received 20 mg per kilogram of zeleciment rostudirsen into a vein every 4 weeks and 8 received placebo. At 6 months the treated boys' muscle biopsies showed dystrophin at 2.87% of normal, or 5.46% after adjusting for how much of each biopsy was actually muscle rather than fat and scar, a correction Dyne reports alongside the raw figure. The chance that result was a fluke was under 1 in 10,000 (Dyne, December 8, 2025). Dyne points out that 2.87% is about 10 times the 0.3% eteplirsen produced in its own trial, and that it got there with 1 infusion a month instead of 4.

An earlier group of 8 boys at the same dose had reached 8.72% adjusted, so the registrational number came in lower than the one that had excited investors in 2025 (Dyne, March 16, 2025). On function, the trial was not designed to prove anything; Dyne says so in its own footnote. Two timed tests, rising from the floor and the 10-meter run, favored the drug with nominal significance at 6 months, and the gain in rise-from-floor speed, 0.04 rises per second against pooled placebo, cleared the published threshold for a difference patients can feel (Dyne, May 20, 2026). Across 86 boys followed for up to 36 months, the most common drug-related side effects were fever and headache, with no related serious events in the registrational cohort and 2 boys in the extension who had serious fever or malaise and recovered (Dyne, December 8, 2025).

What happens on January 21, 2027

Dyne filed in May 2026 for accelerated approval on the dystrophin result, and the FDA accepted the application with priority review on July 20, giving it a decision date of January 21, 2027 (Dyne decision page). Dyne is planning a launch in the first quarter of 2027 if the answer is yes. Its confirmatory trial, FORZETTO, started in May 2026 and is recruiting about 90 boys aged 4 to 18 in a 1 to 1 draw against placebo, with the primary readout at week 73 on rise-from-floor speed (Dyne, May 20, 2026; NCT07608432). That trial is the one that will eventually say whether 2.87% is a step. It is also open now, and recruiting, which matters if you have a boy with an exon 51 deletion who is not on anything.

Del-zota Exon 44 Results, 25% of Normal Dystrophin and Novartis's Priority Review

EXPLORE44 and a number nobody had seen in exon skipping

Now consider the exon 44 boy, the one no drug has ever been for. In March 2025 Avidity reported results from EXPLORE44, a trial in 26 boys with exon 44 amenable deletions who received del-zota, then called AOC 1044, at 5 mg per kilogram every 6 weeks or 10 mg per kilogram every 8 weeks. The drug produced about 40% exon skipping, raised dystrophin by about 25 percentage points of normal, and lifted total dystrophin in some boys to 58% of normal. Creatine kinase, the leak marker that is sky high in Duchenne, fell by more than 80% to near normal (Avidity, letter to the Duchenne community, March 17, 2025).

Those are not exon skipping numbers in any prior sense. On the chart's zones, 25 percentage points lifts a boy out of the Duchenne band, through severe Becker, and into the range where the Dutch study found dystrophin level stopped predicting weakness. The dystrophin del-zota makes is also the boy's own shortened full-length protein, not the much smaller micro-dystrophin from gene therapy, and the drug does not use a virus, so there is no one-shot limit and no antibody screen before dosing.

In September 2025 Avidity reported what a year of it looked like. Seventeen boys, 12 still walking and 5 not, had been followed for about a year in the extension study at the 5 mg per kilogram every 6 weeks dose. Compared with a matched group of exon 44 boys from a natural history database who received no drug, the treated boys got faster where the untreated boys got slower. On the 4-stair climb, treated boys improved by 2.1 seconds while the untreated group lost 2.7 seconds. On the 10-meter walk, 0.7 seconds faster against 1.5 seconds slower. The North Star score held steady while the untreated group fell 2.4 points. Half the boys had a normal creatine kinase at 1 year (Avidity, September 10, 2025). The comparison groups were small, 10 treated boys against 22 untreated on most of those measures, and only 6 treated boys on rise-from-floor, and natural history comparisons are not placebo comparisons, a point we come back to below.

Novartis, the September 8th, 2026 filing acceptance, and the missing date

Avidity had told investors it would file by the end of 2025. After an October 2025 meeting with the FDA, that slipped to 2026 (Avidity, Q3 2025 results). Novartis then bought the company, closing the deal on February 27, 2026, and the filing went in on June 25. On September 8, 2026, Novartis told the Duchenne community that the FDA had accepted the application, granted it priority review, and would assess it under the accelerated approval pathway, with a global Phase 3 trial called SAFARI44, run outside the United States, as the confirmatory study (Novartis community letter, September 8, 2026; NCT07587242). Novartis did not disclose a decision date, which is why our del-zota decision page lists the drug as under review without a day. The FDA's priority review goal for a new drug is 8 months from the day the application was submitted, but neither Novartis nor the agency has stated the date for this one, so this story does not guess.

The safety record so far is short and mostly mild. Across the trial and extension, the most common side effects were colds, diarrhea, falls, back pain and headache, and 1 boy left the extension after a hypersensitivity reaction (Avidity, September 10, 2025). Avidity, now part of Novartis, has also opened a managed access program for boys who cannot wait for the decision (NCT07250737).

What the 2016 Exondys 51 Lesson Means for Reading the New Duchenne Numbers

Here is where the story has to turn on itself, because the strongest reason to be careful about 25% is the same reason 0.44% was approved. Both numbers are dystrophin, and dystrophin is a surrogate. The FDA accepted it as "reasonably likely to predict clinical benefit" in 2016, and 9 years later the confirmatory trial for 2 of its sister drugs, built on the same bet, missed. Dyne and Novartis are asking for the same kind of approval, on the same surrogate, with confirmatory trials still running. The surrogate is much higher now, high enough to reach zones where, in Becker patients, it has always meant something. It is still a surrogate.

There are a few other things to hold in mind as the dates approach. The numbers are from small groups: 24 boys on Dyne's drug, 26 on del-zota, 17 followed for a year. Del-zota's functional gains are against a natural history database, not a placebo group, and boys in trials tend to do better than boys in databases for reasons that have nothing to do with the drug. Dyne's 2 headline figures, 2.87% and 5.46%, measure the same biopsies 2 ways, and the first exon skipping drugs were also, in their day, reported with the most flattering available analysis. Avidity's 25% is an average across 2 dose groups that has not been broken out by dose or by patient in public. None of this means the drugs do not work. It means that the patient-level proof, the kind ESSENCE was supposed to deliver and did not, is what FORZETTO and SAFARI44 are for, and both are enrolling now.

Duchenne Clinical Trials Recruiting Now and 3 FDA Dates to Watch

Three dates frame the next 5 months. January 21, 2027 is the FDA's deadline on zeleciment rostudirsen for exon 51. February 28, 2027 is its deadline on whether ESSENCE, despite missing, is enough to make Vyondys 53 and Amondys 45 fully approved. The third date, for del-zota and exon 44, has not been announced; our decision page will update the day it is. All 3 have signup for a decision-day email.

The trials are the part families can act on today. FORZETTO (NCT07608432) is enrolling boys aged 4 to 18 with exon 51 amenable deletions. SAFARI44 (NCT07587242) is enrolling exon 44 boys, outside the United States, with a saline placebo for the first 54 weeks and open-label drug after. Novartis's managed access program (NCT07250737) is the route for exon 44 boys in the United States before a decision. Sarepta's ENDEAVOR cohort for non-walking boys on Elevidys, with a drug called sirolimus given first to blunt the immune response, is dosing in 2026. Capricor's deramiocel, a cell therapy for arm function in non-walking patients, has its own FDA decision on November 22, 2026, and is a story for another day. Our Duchenne trials page pulls every recruiting study from ClinicalTrials.gov and lets you filter by location.

The boy we imagined, born in 2011, turns 16 next year. In the exon 51 version of his life he has spent somewhere between 2 and 3 weeks of his childhood, added up, in an infusion chair, for 0.93% of normal dystrophin, and on January 21st he will find out whether a drug that gets to 2.87% in a monthly infusion is available to him. In the exon 44 version he has had nothing for 10 years and is waiting on a date Novartis has not named, for a drug whose 25% would have sounded like science fiction to the 13 people in that room in April 2016. The number that split the FDA was 0.44%. The number that will decide the next decade is whichever one the confirmatory trials produce, and this time the boys in those trials are enrolling now.