What we know
That is the entire published human record for Muse cells. It is worth seeing in full, because the individual results are often quoted without the context that makes them interpretable.
How to read a small study
Was there a comparison group? If everyone in a study got the treatment, improvement tells you little — many conditions improve on their own, and people who enrol in trials tend to do better anyway. Only a study where some patients get a dummy treatment can separate the drug from everything else.
What was the study actually asking? Early trials usually ask “is this safe?” rather than “does this work?” Anything else they measure is a side observation — a hint worth following up, not a finding. Quoting those side observations as proof is the single most common way early data gets oversold.
The studies
| Condition | People | What happened |
|---|---|---|
| Stroke | 35 | The only study with a comparison group. 40% of treated patients regained independence at 12 weeks against 10% on the dummy treatment — but the study was designed to test safety, and that 30-point gap was not statistically significant, meaning it could plausibly be chance. A promising signal that needs a proper trial, not a proven result. |
| Spinal cord injury | 10 | Everyone received treatment; no comparison group. Motor scores improved against each patient's own starting point, in a condition where meaningful spontaneous recovery is common. A second measure of daily functioning did not reach significance. The authors state plainly that a control group is needed. |
| Newborn oxygen deprivation | 9 | A dose-safety study in babies. Treatment was tolerated, and two thirds had normal development at follow-up. It was not designed to measure benefit. |
| ALS (motor neurone disease) | 5 | Safe over 12 months. Three of five patients declined more slowly, but the pre-planned comparison was not significant, and patients continued to get worse overall. Several inflammation markers moved in the wrong direction. |
| Epidermolysis bullosa | 5 | A rare blistering skin condition. The company announced the study met its target. The published paper on the same study shows a wound-size result whose margin of error crosses zero, with wounds back to their starting size by week 12 and itching increased afterwards. Where a press release and a paper disagree, the paper is what counts. |
| Heart attack | 3 | Three patients, no comparison group. Heart pumping function improved by about 11 points over three months — but that is also the normal recovery pattern after a treated heart attack. The authors called for a proper controlled trial. It was never run. |
A seventh study you will not see quoted
A trial in COVID-19 lung failure was registered and begun in 2021. Its results have never been published. Studies that go unreported are more often disappointing than not, which is precisely why the silence is worth noting.
Safety
Across every published study: no tumours, no rejection episodes, and no need for anti-rejection drugs even though patients received cells from unmatched donors. For a cell therapy, that is a genuinely good result and it has been consistent.
Three honest limits on it. Sixty-seven patients is far too few to detect a side effect that happens rarely. The record belongs to one specific manufactured product, and does not automatically transfer to a differently made preparation. And “no problems caused by the treatment” is not the same as “no problems” — the spinal cord study recorded two serious adverse events, judged unrelated to the cells.
The programme that produced all of this evidence has stopped
Every study above used one product, made by one Japanese company. In February 2023 that company discontinued its development, citing commercial timelines rather than any safety concern. The confirmatory trials that were supposed to follow the stroke and heart results were never conducted, and no other developer picked the product up.
This matters when you read that Muse cell research is advancing. The underlying biology is genuinely active — new work published in 2026 confirmed these cells across other species. But the clinical programme that generated the human results people quote is not moving forward.
Where things stand legally
No Muse cell product holds marketing approval from any regulator in any country — not the FDA in the United States, not the EMA in Europe, not Japan's PMDA, where the research began.
In the United States, these preparations are investigational. Some states have passed laws permitting clinics to offer stem cell treatments, but state law does not override federal authority — a clinic can be operating lawfully under its state's rules while offering something the FDA has not approved.
None of this means the treatment is unsafe; the published safety record is good. It means that if you receive it commercially, you are outside the system that would normally have required large trials, verified manufacturing and ongoing monitoring before the product reached you. That is a reasonable thing for an informed adult to choose. It is not a reasonable thing to have concealed from you.
Sources
Everything on this site traces to published research or a named laboratory report, rather than to company materials. The main sources:
Kuroda & Dezawa, PNAS 2010 (original description) · Dezawa et al., Nature Protocols 2013 (isolation method) · Uchida, Niizuma et al., Stroke 2017 (nerve repair in animals) · Yamada et al., Circulation Research 2018 (the homing mechanism) · Leng et al., Cell Transplantation 2019 (purity across sorting and expansion) · Noda et al., Circulation Journal 2020 (heart attack, 3 patients) · Fujita et al., JEADV 2021 (epidermolysis bullosa) · Niizuma et al., J Cereb Blood Flow Metab 2023 (the stroke trial) · Yamashita et al., Cell Transplantation 2023 (ALS) · Koda et al., Stem Cell Research & Therapy 2024 (spinal cord injury) · Kushida et al., Cell Mol Life Sci 2024 (tumour testing) · Wakao et al. 2026 (cross-species background rates) · Dominici et al., Cytotherapy 2006 and Horwitz et al. 2005 (how MSCs are defined) · Thompson et al., EClinicalMedicine 2020 (MSC safety across 55 trials) · ClinicalTrials.gov, August 2026 (trial counts) · Mitsubishi Chemical Group release, 14 February 2023 (discontinuation).
Purity benchmarks and the 8.43% batch figure come from an independent flow cytometry report by CICESE, Ensenada, Mexico, July 2026.
