Disease indication · read alongside myotonic dystrophy
The one disease that has actually run the trial DM1 is planning — a molecular primary endpoint against a toxic gene. It missed, the functional measures looked better, and then those missed too.
Disease-modifying: Switch off DUX4, an embryonic transcription factor that should be silent in adult muscle and is toxic when it is not. The cause has been understood since 2010 with unusual precision — and the target is expressed in bursts, in well under one percent of nuclei, which makes proving you have hit it extraordinarily hard.
Nothing alters the course of this disease. Unlike DM1, that is no longer for want of a completed test: one programme reached Phase 3 and failed.
A unifying genetic model published in 2010 explained why a shortened repeat array causes disease only on a permissive chromosome — one that supplies a polyadenylation signal stabilising the DUX4 transcript. Rarely is a cause this precisely specified.
DUX4 appears in brief, stochastic bursts in a tiny minority of nuclei at any moment. The thing to be silenced is barely detectable even when untreated, which is why target engagement cannot be demonstrated the way DM1 demonstrates it.
A single pivotal, as rare disease allows. What is unsettled after the Phase 3 failure is not the size but what it should measure.
The most consequential table on this site for anyone reading the DM1 page. FSHD ran a registrational-intent trial with a molecular primary endpoint against its causal gene — the strategy DM1 is betting on — and the molecular readout and the functional readout pointed in different directions. Then the functional one failed too.
No endpoint is accepted as establishing disease-modifying benefit in Facioscapulohumeral Muscular Dystrophy. Every measure below is either contested or exploratory. A trial can therefore succeed on its own terms and still not support the claim.
| Endpoint | Type | Read at | Standing |
|---|---|---|---|
| DUX4-regulated gene expression signature | Laboratory | 24–48 weeks | Contested |
| Reachable workspace | Performance | 48 weeks | Contested |
| MRI muscle fat fraction | Imaging | 12–24 months | Exploratory |
| MRI STIR hyperintensity | Imaging | Baseline and serial | Exploratory |
| FSHD Composite Outcome Measure (FSHD-COM) | Performance | 12–24 months | Exploratory |
| Quantitative muscle strength | Performance | 12–24 months | Contested |
Every model here forces the toxic gene on. None reproduces how it actually behaves in patients — brief, random bursts in well under one percent of nuclei — and none reproduces the regional, asymmetric pattern the disease is literally named after. High construct validity for the toxin, low fidelity to its behaviour.
| Model | FaceDoes it look like the disease? | ConstructDoes it arise the same way? | PredictiveSilence the gene |
|---|---|---|---|
One programme reached Phase 3 and failed twice over, on two different kinds of endpoint. The older entries are just as instructive: two separate drugs made muscle measurably bigger without making patients measurably better.
| Intervention | Rested on | Primary endpoint | Reached | Outcome |
|---|---|---|---|---|
Losmapimod Fulcrum Therapeutics · p38α/β MAPK inhibition, reducing DUX4 expression | Patient myotubesFLExDUX4 | ReDUX4: DUX4-regulated gene expression in muscle biopsy at 24 weeks. REACH: change in reachable workspace at 48 weeks | Phase 3 ReDUX4 / REACH · 2025 | Failed |
Delpacibart braxlosiran (AOC 1020) Avidity Biosciences · Anti-transferrin-receptor antibody conjugated to siRNA against DUX4 | Patient myotubesAAV-DUX4 | Safety and tolerability, with DUX4-regulated gene expression and functional measures as key endpoints | Phase 1/2 FORTITUDE · 2024 | Ongoing |
ARO-DUX4 Arrowhead Pharmaceuticals · RNA interference against the DUX4 transcript | AAV-DUX4FLExDUX4 | Safety and tolerability, with DUX4-regulated gene expression as a key measure | Phase 1/2 2024 | Ongoing |
ACE-083 Acceleron Pharma · Locally injected follistatin-based molecule increasing muscle mass | Change in total muscle volume by MRI in the injected muscle | Phase 2 2019 | Failed | |
MYO-029 (stamulumab) Wyeth · Myostatin neutralisation to increase muscle mass | Safety, with muscle strength and function as efficacy measures | Phase 2 2008 | Failed | |
Albuterol Investigator-led · β2-adrenergic agonism with anabolic effects on muscle | Change in muscle strength at 12 months | Phase 3 2001 | Failed | |
Prednisone Investigator-led · Broad transcriptional suppression of inflammation | Muscle strength and mass over 12 weeks | Phase 2 1997 | Failed |
Activity of the genes DUX4 switches on, measured in a muscle biopsy as a proxy for DUX4 itself
A molecular surrogate for a causal gene, tested as a primary endpoint and found wanting. DM1's splicing index is the same species of bet, one disease over, and has not yet had its equivalent test.
Rare-disease economics much like DM1's: small trials, a single pivotal, and a population of roughly one in ten thousand. What differs is the likelihood of approval, because unlike DM1 this disease has already run a Phase 3 and lost.
Small, and long enough for a slowly progressive disease to separate the arms. The expensive part is not the headcount but the muscle biopsies and repeat quantitative MRI that both candidate endpoints require — and after the Phase 3 failure, which of those endpoints a pivotal should read is genuinely unsettled.
340 patient-years against 140 for a symptomatic programme. Cost grows more slowly than exposure because the fixed enrolment cost of each participant — screening, baseline imaging — is paid once however long they are then followed. The goal, not the molecule, still sets the budget.
Citedenrolment, duration and study counts, from named trials.Assumptionthe two cost parameters below.Computedeverything else.
| Phase | Design | N × studies | Duration | Patient-years | Modelled cost |
|---|---|---|---|---|---|
| Phase 1 | SAD / MAD, healthy volunteers Reads: Safety, tolerability and pharmacokinetics Conjugated oligonucleotides and systemic kinase inhibitors both carry platform-level safety questions, so first-in-human work is not shortened by the rare-disease setting. Conventional first-in-human designmoderate confidence | 40 | 9 mo | 30 | $1.1M–$1.8M |
| Phase 2 | Randomised, with paired muscle biopsy and MRI Reads: DUX4-regulated gene expression in an MRI-selected muscle, with functional measures alongside Sized after ReDUX4. The biopsy must come from a muscle MRI shows to be actively affected — sample a quiet one and the molecular endpoint has nothing to report whatever the drug is doing. ReDUX4 (Fulcrum) Phase 2b designmoderate confidence | 80 | 1 yr | 80 | $2.6M–$4.2M |
| Phase 3 | Single pivotal, 48 weeks, functional primary Reads: Reachable workspace, with functional composite and MRI fat fraction alongside Sized after REACH, which switched to a functional primary precisely because the molecular one had failed at Phase 2b — and then missed the functional one as well. REACH (Fulcrum) Phase 3 designmoderate confidence | 230 | 1 yr | 230 | $7.4M–$12M |
| Total | 2.8 yr of clinical development, treating phases as sequential | 340 | $11M–$18M |
Model assumptions, not citations — the only free parameters here. The fixed component covers screening, baseline imaging and randomisation, incurred once per participant; the running component covers visits, monitoring and site fees for each year on study. Splitting them matters: a single rate per patient-year would price short symptomatic trials far below what pivotal trials actually cost.
Calibration: priced this way a single pivotal trial in this programme comes to $3.2M–$5.3M, against published pivotal trial costs of $12–33M (median $19M) across all indications. A rare-disease pivotal of roughly a hundred participants sits well below the published all-indication range, exactly as DM1's does. Trial size is not what makes this disease hard. Moore et al., JAMA Internal Medicine 2018.
Programme cost divided by a 12% likelihood of approval from Phase 1 in rare disease, adjusted downward for this indication — what one success costs once the failures are paid for. BIO / Informa / QLS, Clinical Development Success Rates — rare-disease programmes clear Phase 1 to approval well above the all-indication average.
Read this one carefully: Set below the rare-disease average that the DM1 page uses, and deliberately. Those averages are lifted by oncology and by enzyme replacement in diseases with a missing protein to restore; this is neither. It is also the one rare disease on this site that has already run a Phase 3 and lost — twice over, on two different kinds of endpoint — which is direct evidence about this indication rather than an inference from a portfolio.
These are portfolio figures spanning all of R&D and already carry the cost of failure, so they are not comparable with the per-programme clinical costs above — they are the order of magnitude those costs roll up into.
The same conjugate chemistry that rescued DM1 delivery is now aimed at this disease's gene — while the field re-argues, after a failure, what a trial here should measure at all.
The antibody–oligonucleotide conjugates that solved muscle delivery for DM1 are now aimed at DUX4, by more than one sponsor. That is a genuine platform read-across: the hard-won answer to getting a drug into skeletal muscle transfers between diseases even though nothing about the underlying biology does.
Avidity and Arrowhead FSHD programme disclosures
After a Phase 3 that missed a functional primary, following a Phase 2b that missed a molecular one, the field is left without an agreed endpoint and with active disagreement about which failure to learn from. Reachable workspace, the composite functional measure and MRI fat fraction are all being reconsidered — the same endpoint-first problem that Crohn's strictures face from a standing start, arrived at here by having tried.
Post-REACH endpoint discussion in FSHD
DM1's accelerated-approval strategy rests on a molecular surrogate for a causal defect. FSHD ran that experiment first: its molecular surrogate and its functional measures disagreed within a single trial. That does not mean splicing correction will fail to predict function in DM1 — the assays and the biology differ substantially — but it is the nearest available precedent, and it did not go well.
ReDUX4 and REACH results, read against DM1 accelerated-approval strategy