What an n-of-1 trial actually is

The phrase "n-of-1 trial" is doing two jobs in rare-disease writing, and the two are usually conflated. The original sense, codified by Gordon Guyatt and David Sackett at McMaster University in the 1980s, is a clinical research design. The newer sense, popularized after the 2019 publication of the milasen case at Boston Children's Hospital, is a kind of drug development. The two share a name and a single-person focus. They share almost nothing else.

Mila Makovec and milasen, Ipek Kuzu and atipeksen, Valeria and valeriasen, Jaci Hermstad and jacifusen are all the second sense. The children and young adults whose first names show up next to drug names that sound like nothing else are inside the post-milasen redefinition. The Guyatt sense is older and still in active use, but it is not what most readers encountering "n-of-1 trial" in a news story about rare-disease therapy mean.

The original n-of-1 trial

Guyatt, Sackett, and colleagues published "Determining optimal therapy: randomized trials in individual patients" in the New England Journal of Medicine in 1986. A follow-up paper, "The n-of-1 randomized controlled trial: clinical usefulness. Our three-year experience," appeared in the Annals of Internal Medicine in 1990. The papers described an experimental design and a clinical service the team had been running at McMaster Medical Centre, in which a single individual cycled through alternating periods of an active treatment and a placebo, blinded, with washout between cycles.

The design's purpose was to settle a clinical question for one person. Group studies tell you whether a drug works on average across a population. They do not tell you whether the drug works for the specific person sitting in front of you. For chronic conditions where individual response varies a great deal (osteoarthritis pain, ADHD stimulant titration, neuropathic pain medications, migraine prophylactics), the population-average effect is not the answer the clinician needs. The n-of-1 design produced an answer for one person, with the methodological rigor of a randomized blinded trial.

The basic structure is A-B-A-B: treatment period, washout, placebo period, washout, treatment period, and so on. Random assignment of which arm comes first. Outcome measurement at the end of each period. The individual and the clinician are blinded. After three or four cycles, statistical analysis can identify whether the treatment effect for that one person exceeds the noise of their own day-to-day variation.

The Guyatt design has not gone away. The International Collaborative Network for N-of-1 Trials and Single-Case Designs, co-chaired by Jane Nikles and Suzanne McDonald, was established in 2017 and has more than 400 members across 30 countries. Bayesian extensions of the original method are an active area of statistical work. The applications are mostly the same conditions the McMaster group studied: chronic pain, attention disorders, sleep, idiosyncratic responses to common medications.

The Guyatt sense of n-of-1 trial is a method for doing clinical research with one person. The drug being tested is an existing drug. The question is whether it works for this one individual.

The newer sense

The October 2019 New England Journal of Medicine paper by Timothy Yu and colleagues at Boston Children's Hospital, describing the development of milasen for Mila Makovec, used the phrase "patient-customized oligonucleotide therapy" rather than "n-of-1 trial." The community that grew around the case adopted the n-of-1 terminology anyway, because the regulatory submission was structured around one identified individual and because the rhetoric of single-patient evidence had a useful precedent.

What the milasen-class program actually is: a drug designed and manufactured for one mutation, often present in one identified person, dosed under an investigational new drug application held by a sponsor-investigator (typically the academic clinician running the program). The drug is a new molecular entity, synthesized to bind a specific RNA sequence specific to the affected person's genome. The therapeutic mechanism, in the case of antisense oligonucleotides, is splice modulation, exon skipping, or RNA stabilization. The chemistry class is the same chemistry that produced nusinersen, the approved ASO for spinal muscular atrophy that Adrian Krainer at Cold Spring Harbor Laboratory and Frank Bennett at Ionis Pharmaceuticals developed in the mid-2010s.

The newer sense has nothing structurally to do with the Guyatt design. There is no crossover. There is no placebo arm. There is no blinding. The disease is usually progressive and irreversible, and the affected person is usually deteriorating, so the treatment cannot be withdrawn for an A-B-A comparison without ethical problems. Efficacy assessment is open-label and case-by-case, anchored in objective measurements (seizure frequency, neurofilament light chain in cerebrospinal fluid, motor scales, MRI changes) before and after dosing, with the affected person's own pre-treatment trajectory as the comparator.

The two senses share the phrase "n-of-1" and a focus on one person. They do not share the methodology, the regulatory pathway, the development cost, or the question being asked.

How the FDA formalized the newer sense

Between January 2021 and July 2022, the FDA issued four draft guidance documents specifically for sponsor-investigators developing individualized antisense oligonucleotide drugs. The series covers Administrative and Procedural Recommendations (January 2021), Nonclinical Testing (April 2021), Chemistry, Manufacturing, and Controls (December 2021), and Clinical Recommendations (July 2022). The guidance applies to "severely debilitating or life-threatening" genetic conditions caused by a unique variant present in a small number of individuals, typically one or two.

The documents are explicit about how the milasen-class program differs from a conventional IND. The nonclinical package is abbreviated: a single three-month rodent toxicology study can be sufficient if the chemistry class and the route of administration are well precedented. The CMC package is simplified for chemistry classes, such as 2'-O-methoxyethyl phosphorothioate ASOs, where the manufacturing platform has known impurities and known stability characteristics. The clinical recommendations describe how to design dose escalation, how to monitor for class-known adverse events (intrathecal-route hydrocephalus, CSF pleocytosis, transient transaminase elevations), and how to define stopping rules without a control arm.

The guidance never calls the regulatory framework an n-of-1 trial. It calls the program an "individualized antisense oligonucleotide drug product" intended for a "small number of individuals." The four documents together are the closest formal equivalent the FDA has issued to a master protocol for drug-design-for-one. The framework was being assembled while the cases that defined it were being treated.

What a current program looks like

Five named ASO programs beyond milasen have published or publicly described their development. Atipeksen, the ASO Yu's lab developed for Ipek Kuzu's variant of ataxia-telangiectasia, began intrathecal dosing in early 2020 when Ipek was two years and nine months old, escalating from 3.5 mg to 42 mg over ten weeks. She has continued treatment on a roughly quarterly schedule and was six years old at the time of recent published commentary. Valeriasen, developed for an infant named Valeria with a KCNT1 epilepsy variant, suppressed her mutation in vitro and entered clinical use after an eight-to-ten-week toxicology run. Valeria died of disease progression, and her treatment was complicated by hydrocephalus, the same intrathecal-route adverse event the FDA's clinical guidance now flags as a class-known risk. Jacifusen, also called ION363 or ulefnersen, is an ASO developed by Ionis Pharmaceuticals with Neil Shneider at Columbia University for FUS-mutated ALS. The open-label case series of 12 participants treated between June 2019 and June 2023 was published in the Lancet in 2025, with cerebrospinal fluid neurofilament light chain reductions of up to 82.8 percent after six months and apparent slowing of disease progression in some participants. A KIF1A-targeted program is in development at Columbia under Wendy Chung in collaboration with the KIF1A.org family network.

The n-Lorem Foundation, founded by Stanley Crooke, the former chairman and chief executive of Ionis Pharmaceuticals, has formalized the philanthropic infrastructure for the model. As of April 2026, n-Lorem has received more than 440 applications, approved more than 240 nano-rare individuals (the Foundation's term for variants found in one or very few people), and treated 50 individuals across more than 260 doses, with no ASO-related serious adverse events reported across approximately 45 patient-years of cumulative follow-up. The cost of each program is borne by donor funding.

Boston Children's, in published commentary, has placed the development cost of its individualized ASO programs at roughly $1.6 million per program. The median cost of a standard FDA-approved drug, by comparison, is $985.3 million. The cost gap reflects the abbreviated regulatory package, the academic-funded research staff, and the absence of commercial-scale manufacturing.

Where the term will go

"N-of-1 trial" is a phrase in tension with itself. The Guyatt sense is precise and methodological. The milasen sense is rhetorical and practical, and it is the sense most readers encountering the phrase in 2026 mean. The conflation will probably resolve toward the second sense in popular usage and the first sense in clinical-trials methodology textbooks.

A more accurate term for the second sense is one the field has already started using: "n-of-1 drug," "individualized therapy," or, when the same custom drug is given to two or three people with the same private mutation, "n-of-few." None of these terms have settled. What has settled is the underlying activity. A drug can be designed and built and administered for one identified person with one identified mutation, on a timeline of months, under an established regulatory pathway, with a growing safety database that has not recorded an ASO-related serious adverse event across more than 45 patient-years. The vocabulary is catching up to the practice.