On February 23, 2026, during Rare Disease Week, the FDA released a 23-page draft guidance document with a title so dense most people scrolled past it. "Considerations for the Use of the Plausible Mechanism Framework to Develop Individualized Therapies that Target Specific Genetic Conditions with Known Biological Cause." Buried in that bureaucratic language is a policy shift that could reshape how treatments reach the 30 million Americans living with a rare disease.
To understand why this matters, start with what gene editing actually is. Many rare diseases are caused by a single error in a person's DNA, a misplaced letter in the genetic code that prevents a critical protein from being made correctly. Gene editing tools like CRISPR can find that specific error and fix it, the way you'd correct a typo in a document. Antisense oligonucleotides, another class of therapy covered by this framework, work differently. They don't rewrite the DNA itself. Instead, they intercept the faulty instructions before the cell reads them, blocking the bad message or redirecting the cell to read around the error. Both approaches go after the root cause of a disease rather than managing symptoms.
The problem has never been the science. It's the approval process. To get FDA approval, drug developers traditionally need to run large clinical trials with hundreds or thousands of participants, then compare treated patients against a control group. That works when you're testing a blood pressure medication with millions of potential patients. It falls apart when the disease you're treating affects 50 people on earth, or when each patient's mutation is slightly different.
The plausible mechanism framework changes that equation. If a therapy targets a known genetic mutation, and scientists can demonstrate a plausible biological mechanism for why it should work, the FDA will no longer demand that kind of large-scale trial. Instead, developers can seek full approval based on a clear scientific rationale, solid natural history data showing what happens without treatment, and clinical evidence from a small number of patients proving the therapy reached its target and improved outcomes.
For decades, the same evidentiary bar applied whether you were testing a cholesterol drug with 10 million potential patients or a gene therapy for a condition that affects 12 people worldwide. That math never worked. Even with orphan drug designations that offer tax credits and market exclusivity, pharmaceutical companies couldn't justify the cost of a traditional clinical trial for a disease with so few patients. And even if they tried, they couldn't enroll enough participants to generate statistically significant results. The plausible mechanism framework is the FDA's formal acknowledgment that its own clinical trial approval standards were quietly killing rare disease treatments before they ever had a chance.
What the framework actually changes
The draft guidance, issued jointly by the Center for Biologics Evaluation and Research (CBER) and the Center for Drug Evaluation and Research (CDER), lays out a path to full FDA approval for individualized therapies. Not accelerated approval. Not emergency use. Full, traditional approval, built on a different kind of evidence.
The requirements center on 4 criteria. First, the developer must identify the specific genetic, cellular, or molecular abnormality causing the disease. Second, they must demonstrate that the therapy directly targets the root cause or a proximate biological pathway. Third, they need well-characterized natural history data showing what happens to untreated patients over time. Fourth, they must confirm successful target engagement, meaning they can prove the therapy actually reached and modified the intended target.
The guidance specifically names genome editing (CRISPR and base editing) and RNA-based therapies like antisense oligonucleotides as the primary candidates. It also leaves the door open for other targeted therapeutics that directly address the underlying cause of disease, though it stops short of defining exactly where that boundary sits.
One line in the guidance stands out above the rest. A disease with 100 causing mutations will no longer require 100 clinical trials. Once a platform therapy proves it works for one mutation through a plausible mechanism, the FDA can extend that approval to additional variants of the same disease without requiring a separate trial for each one. For conditions like Duchenne muscular dystrophy, where over 7,000 different mutations in the dystrophin gene can cause the disease, this is the difference between a theoretical treatment and a real one.
Baby KJ and the 6-month therapy
The framework didn't emerge from theory. It emerged from a baby named KJ.
KJ was born with severe carbamoyl phosphate synthetase 1 (CPS1) deficiency, a metabolic disorder so rare it has no established prevalence figure. His body couldn't process nitrogen from protein, which meant ammonia built up in his blood to toxic levels. Without intervention, the prognosis was liver transplant or death. He spent the first months of his life in the hospital at Children's Hospital of Philadelphia (CHOP), surviving on a severely restricted diet.
A team at CHOP and Penn Medicine, led by Dr. Rebecca Ahrens-Nicklas and Dr. Kiran Musunuru, designed a personalized CRISPR base-editing therapy specifically for KJ's mutation. They manufactured it using lipid nanoparticles that delivered the gene editor to his liver. The entire process, from identifying the mutation to administering the first dose, took 6 months.
KJ received his first infusion in February 2025. By April, he'd had 3 doses with no serious side effects. He tolerated increased dietary protein. He needed less nitrogen scavenger medication. In February 2026, on the 1-year anniversary of his treatment, CHOP reported that KJ was walking and talking, hitting developmental milestones that his disease should have made impossible.
“Within six months, their team designed and manufactured a base editing therapy delivered via lipid nanoparticles to the liver in order to correct KJ's faulty enzyme. KJ has achieved meaningful clinical improvements, such as walking and talking, as he continues to grow and thrive.”
Children's Hospital of Philadelphia, February 2026
KJ's case is the FDA's proof of concept for the plausible mechanism framework. His therapy didn't go through a randomized controlled trial with hundreds of participants. It went through a rigorous process of identifying the exact genetic defect, designing a therapy with a clear biological rationale for why it should work, confirming the therapy reached its target, and measuring clinical outcomes against well-documented natural history of the disease. That's exactly the evidentiary model the new guidance describes.
The precedent that made this possible
KJ wasn't the first. In 2018, a team at Boston Children's Hospital led by Dr. Timothy Yu designed milasen, a custom antisense oligonucleotide for a 6-year-old girl named Mila Makovec who had Batten disease, a fatal neurodegenerative condition. The drug was designed, manufactured, and approved for an n-of-1 trial by the FDA in 10 months.
Before treatment, Mila had 30 seizures a day, each lasting about 2 minutes. After treatment, she had 5 to 12 seizures daily, lasting seconds. The therapy didn't cure her disease, but it slowed the progression in ways her family and doctors could measure. Mila passed away in February 2021, at age 10. She lived longer than her prognosis predicted.
Milasen showed that individualized genetic therapies could work. KJ's case, 7 years later, showed they could work faster, with newer tools, and with outcomes that looked less like slowing decline and more like restoring function. Together, they built the clinical foundation that the FDA's framework now rests on.
Scaling from 1 patient to thousands
The obvious question after KJ's story is whether this can scale. A therapy designed for a single baby with a single mutation is remarkable science. It's not a healthcare system.
The CHOP and Penn Medicine team is already working on the answer. They've reached agreement with the FDA on designs for "umbrella" clinical trials that enroll patients with different genetic variants under a single protocol. Instead of running a separate trial for every mutation, all versions of the gene editing therapy would be considered a single drug. The first umbrella trial, planned for 2026, will enroll patients with any of 7 different urea cycle disorders caused by variants in any of 7 genes, all treatable with the same type of base editor used for KJ.
The numbers the FDA is discussing are striking. Successful treatment of as few as 5 to 10 participants might be sufficient for approval of the gene editing platform. Compare that to the hundreds or thousands of participants in a typical Phase 3 trial. For diseases where the global patient population might number in the dozens, this is the difference between "someday" and "this decade."
Antisense oligonucleotides already in the clinic
Gene editing gets the headlines, but the plausible mechanism framework equally applies to antisense oligonucleotides (ASOs), and ASOs have a longer track record in approved therapies. These short synthetic strands of nucleic acid bind to specific RNA sequences and modify how genes are expressed. They can silence a toxic protein, restore a missing one, or skip over a damaged section of genetic code.
Nusinersen (Spinraza), approved in 2016 for spinal muscular atrophy, was the first ASO to demonstrate that targeting the genetic root of a rare disease could fundamentally change its trajectory. Infants who would have died before age 2 were sitting, some were walking. Tofersen (Qalsody), approved in 2023 for SOD1-associated ALS, showed the approach could work in adults with a neurodegenerative disease. Inotersen, approved in 2018 for hereditary transthyretin amyloidosis, added another data point.
Each of these approvals involved traditional clinical trials. What the plausible mechanism framework offers is a path for the next generation of ASOs, the ones targeting mutations so rare that a traditional trial is logistically impossible. A research team at any academic medical center could identify a patient's specific mutation, design an ASO to address it, and pursue FDA approval through documented mechanism of action and natural history comparison rather than a randomized trial they could never enroll.
The single pivotal trial shift
The plausible mechanism framework wasn't the only policy change the FDA announced in February. On February 19, 4 days before the framework guidance, FDA Commissioner Marty Makary and CBER Director Vinay Prasad published a paper in the New England Journal of Medicine declaring that a single adequate and well-controlled clinical trial, combined with confirmatory evidence, would now serve as the default standard for all drug approvals.
For decades, the FDA's informal default was 2 independent pivotal trials. Makary and Prasad argued that advances in biology, trial methodology, and statistical science made the second trial redundant in most cases. They framed the change as common sense. In 2026, they wrote, "there are powerful alternative ways to feel assured that our products help people live longer or better than requiring manufacturers to test them yet again."
For rare disease drug development, dropping from 2 trials to 1 is a massive practical change. If your entire patient population is 200 people, enrolling enough participants for 1 trial is already difficult. Enrolling enough for 2 is often impossible. The single-trial default, combined with the plausible mechanism framework for the rarest conditions, creates a regulatory environment that's fundamentally more accessible than what existed 6 months ago.
Legitimate concerns
Not everyone is celebrating. The Health Affairs journal published a 2-part analysis titled "The Promise and Perils of FDA's New Plausible Mechanism Pathway." STAT News reported concerns from bioethicists and former FDA officials about the potential for lowered safety standards. The criticism falls into a few categories.
Post-market monitoring is the biggest open question. The draft guidance discusses post-market evidence generation in terms of sponsor "commitments" rather than enforceable requirements. Critics argue that the FDA has stronger tools available, like mandatory post-market studies, and should use them. When you're approving a therapy based on 5 to 10 patients, long-term safety data doesn't exist at the time of approval. The monitoring plan needs to be airtight.
Regulatory consistency is another concern. NORD, in their official statement, was measured. They called the framework "encouraging" while emphasizing that "its impact will depend on consistent, transparent implementation that drug developers and patients can trust." That language reflects a real worry. The FDA under Makary and Prasad made several decisions about rare disease drug applications in 2025 that contradicted their own prior guidance, creating confusion among drugmakers about what the agency actually expected. A framework is only as good as its application.
Patient advocate Judy Stecker raised a different kind of concern. Individualized therapies help one child and one mutation at a time. For families whose child has a different variant of the same disease, watching another family's child get treated while theirs waits can be agonizing. The umbrella trial model is the right answer to this problem, but it's still early. The gap between the first treated patient and the 100th could be years.
What this means if you're tracking a trial
If you're a patient or caregiver following a clinical trial on Trial Friend, here's what to watch for. The framework applies most directly to gene editing and ASO therapies for ultra-rare genetic conditions. If the trial you're tracking involves a CRISPR-based therapy, a base editing approach, or an antisense oligonucleotide targeting a specific mutation, the approval pathway just got shorter.
Look for trials that mention "n-of-1" or "umbrella" designs. These are the trial structures built for the plausible mechanism pathway. The CHOP urea cycle disorder trial launching in 2026 is the first major umbrella trial under this framework. Others will follow across different disease families.
The diseases most likely to benefit first are monogenic conditions, diseases caused by a single gene defect, where the biology is well understood and natural history data already exists. Many of the conditions tracked on Trial Friend fall into this category. Duchenne muscular dystrophy, sickle cell disease, spinal muscular atrophy, cystic fibrosis, and dozens of metabolic disorders all have the kind of genetic clarity the framework requires.
The FDA rare disease approval process shifted in February 2026. The treatments that could use this new ground are already being built. The question now is whether the FDA can implement the framework as consistently as the science demands, and whether the cost of manufacturing individualized therapies can come down fast enough to make the promise real for more than a handful of families at a time.