What has been tested in humans, what was found, and what has never been studied at all
One intravenous mesenchymal stem cell product slowed whole-brain volume loss by 48.4% in a randomized phase 2a trial of 49 patients published in Nature Medicine. Its cognitive endpoint’s confidence interval crossed zero. That single trial is the entire controlled evidence base for regenerative therapy in Alzheimer’s disease — and it is not what commercial clinics are selling.
The appeal of this category is easy to understand. After a genotype result, the interventions actually available feel indirect — exercise, sleep, blood pressure, lipids. A therapy that promises to repair rather than slow is compelling in a way that a statin is not. Clinics understand this, and market accordingly.
What follows is the published human evidence for stem cells, umbilical-derived products, exosomes, and peptides in Alzheimer’s disease — what has been tested, what was found, and what has never been studied at all.
Evidence at a glance
| Intervention | Best available human evidence | Status |
|---|---|---|
| IV allogeneic bone-marrow MSCs (laromestrocel) |
Randomized, double-blind, placebo-controlled phase 2a, n=49, 10 US centers | PIPELINE |
| Intraventricular umbilical cord blood MSCs | Open-label phase I, n=9, Ommaya reservoir; safety only | EARLY / NO EFFICACY DATA |
| Intranasal MSC-derived exosomes | Three-arm dose-finding phase I/II, no placebo group | EARLY / UNCONTROLLED |
| Cerebrolysin (porcine peptide preparation) | Meta-analysis of 6 RCTs; global impression improved, cognition unresolved | NOT APPROVED IN US |
| Intranasal insulin | Phase 2/3, 27 sites, n=289; primary endpoint not met | NEGATIVE |
| Semaglutide (GLP-1 peptide) | Two phase 3 trials, n=3,808, 40 countries; primary endpoint not met | NEGATIVE |
| Clinic peptides (BPC-157, thymosin β4, epitalon, dihexa) |
None. No published human trial in Alzheimer’s disease. | NO EVIDENCE |
Intravenous mesenchymal stem cells: the one real signal
Laromestrocel is an allogeneic mesenchymal stem cell preparation derived from the bone marrow of young healthy donors and delivered intravenously. The CLEAR MIND trial (NCT05233774) randomized 49 participants with mild Alzheimer’s disease across ten US centers into four arms: placebo ×4 monthly infusions (n=12); 25 million cells once followed by three placebo infusions (n=13); 25 million cells monthly ×4 (n=13); and 100 million cells monthly ×4 (n=11).
The trial met its primary safety endpoint. Rates of treatment-emergent serious adverse events within four weeks of infusion were comparable across arms (0% to 9.1%), with no infusion-related reactions, no hypersensitivity, and no amyloid-related imaging abnormalities. At 39 weeks, whole-brain volume loss was slowed by 48.4% across all treatment groups combined (P=0.005) and left hippocampal volume loss by 61.9% (P=0.021). Change in bilateral hippocampal atrophy correlated with change in MMSE score across the full study population (R=0.41, P=0.0075).
The prespecified secondary endpoint was a composite Alzheimer’s disease score. The published between-group difference for the arm that met it was 0.38, with a 95% confidence interval of −0.06 to 0.82. That interval crosses zero. The endpoint was declared met under the trial’s prespecified criterion, but the estimate is compatible with no cognitive benefit.
The brain-volume result also requires interpretive caution. Preserved volume is intuitively appealing, but the anti-amyloid immunotherapies that slowed clinical decline in phase 3 trials produced excess volume loss in treated participants — a phenomenon now described in the literature as amyloid-removal-related pseudo-atrophy. Volume is therefore not a directionally reliable surrogate for benefit in Alzheimer’s disease.
With 11 to 13 participants per arm, this is a proof-of-concept study. The authors say so explicitly, concluding that larger trials are warranted. That is the correct reading. It is a legitimate result that justifies a properly powered phase 3 — and nothing beyond that. No clinic can offer this product; it exists only inside a regulated development program.
Direct delivery into the brain: tested, and costly
The most-cited rationale for intrathecal or intraventricular administration is that the blood–brain barrier limits what an intravenous infusion can deliver. A Korean group at Samsung Medical Center tested this directly. Nine patients with mild-to-moderate Alzheimer’s disease had an Ommaya reservoir implanted into the right lateral ventricle and received three repeated injections of umbilical cord blood–derived mesenchymal stem cells at four-week intervals.
Every participant developed fever. Seven had headache, five nausea, four vomiting — all resolving within 36 hours. Two participants experienced three serious adverse events attributed to the investigational product, each requiring an additional day of hospitalization. The authors concluded the procedure was feasible and sufficiently safe. This was a phase I trial; efficacy was not an endpoint and none was demonstrated.
The distinction matters when evaluating a clinic offer. A neurosurgically implanted reservoir in an academic center under a registered protocol is a different act from an intrathecal injection administered in an outpatient suite. The published data describe the former. Commercial offerings resemble the latter.
Exosomes and extracellular vesicles
Exosomes are small membrane-bound vesicles released by cells. The therapeutic logic is that most of what a stem cell does at a distance, it does through its secretome — so deliver the secretome and skip the cells. This is biologically coherent and, in Alzheimer’s disease, almost entirely untested in humans.
One published trial exists (NCT04388982). Participants with mild-to-moderate Alzheimer’s disease received intranasal allogeneic adipose-derived MSC exosomes twice weekly for 12 weeks, in one of three dose groups. No adverse events were reported. In the medium-dose arm, ADAS-cog scores improved by 2.33 points and MoCA by 2.38 points at week 12 relative to baseline, with continued ADAS-cog improvement through week 36. There were no significant differences in amyloid or tau deposition across arms.
There was no placebo arm; all comparisons are within-participant against baseline. Repeated cognitive testing produces practice effects in exactly this direction. And the benefit appeared in the middle dose but not the low or high dose — a non-monotonic pattern more consistent with noise than with pharmacology. The trial’s own stated purpose was dose selection for future studies. It selected a dose. It did not demonstrate efficacy.
Peptides: three very different categories
“Peptide therapy” is marketing language that collapses three groups with radically different evidence.
Tested and negative. Intranasal insulin was evaluated in a phase 2/3 trial across 27 sites. In the primary intention-to-treat population of 240 participants, there was no difference from placebo on the ADAS-cog-12 at 12 months (0.0258 points; 95% CI −1.771 to 1.822; P=0.98), nor on any secondary clinical or cerebrospinal fluid outcome. Oral semaglutide, a GLP-1 receptor agonist peptide, was tested in two phase 3 trials enrolling 3,808 participants with amyloid-confirmed early Alzheimer’s disease across 566 sites in 40 countries. Neither trial found benefit on the Clinical Dementia Rating–Sum of Boxes at 104 weeks (estimated differences −0.08 and 0.10; P=0.57 and P=0.46). Both trials were discontinued for negative clinical outcome.
Tested with unresolved results. Cerebrolysin is a porcine brain–derived peptide preparation, widely used in parts of Europe and Asia and not approved in the United States. A meta-analysis of six randomized double-blind placebo-controlled trials found significant improvement in clinical global impression, while explicitly noting that more convincing evidence was still required for cognition and activities of daily living. Two subsequent umbrella reviews reached divergent emphases: one listed Cerebrolysin among agents with beneficial cognitive effects, the other characterized the supporting evidence as limited. The frequently repeated claim that Cochrane assessed Cerebrolysin in Alzheimer’s disease is incorrect — the Cochrane review addresses vascular dementia, where it found benefit on cognition and global function but judged the evidence insufficient to recommend routine use.
Never tested. BPC-157, thymosin β4, epitalon, dihexa, and semax have no published randomized trial in Alzheimer’s disease. This is not weak evidence or preliminary evidence. It is the absence of evidence, and it should be named as such rather than described as “emerging.”
Documented harms
The risks in this category are not hypothetical. A 66-year-old man who received intrathecal infusions at commercial clinics in China, Argentina, and Mexico for residual stroke deficits developed progressive back pain, paraplegia, and urinary incontinence. Biopsy revealed a densely cellular, highly proliferative primitive neoplasm with glial differentiation. Short tandem repeat DNA analysis established that the mass was predominantly non-host cells — it had grown from the infused cells.
A boy with ataxia telangiectasia treated with intracerebellar and intrathecal fetal neural stem cells developed a multifocal brain tumor four years later; molecular analysis showed it derived from at least two donors. A separate lesion type, neuroglial stem cell–derived inflammatory pseudotumor of the lumbosacral cord and nerve roots, has been characterized following intrathecal allogeneic stem cell intervention. A systematic review of reported adverse events concluded that stem cell interventions administered outside proper testing, accepted manufacturing standards, and clinical supervision can pose serious risks.
These are rare events. They are also irreversible, and they occurred in people who believed they were receiving a low-risk biological product. In a condition where the intervention is elective and the alternative is a set of modifiable risk factors with genuine evidence behind them, that asymmetry should carry weight.
Regulatory status in the United States
The only FDA-approved stem cell products are blood-forming (hematopoietic progenitor) cells derived from umbilical cord blood, approved for disorders of blood production and not for other uses. There are no FDA-approved exosome products for any indication. Products marketed as Wharton’s jelly, amniotic fluid, umbilical cord tissue, or stromal vascular fraction for neurologic indications are unapproved. A clinic’s registration with the FDA is not approval, and language implying otherwise should be treated as a warning sign rather than a credential.
Prevention, and the APOE ε4 question
Every trial described above enrolled people who already had symptomatic Alzheimer’s disease — mild dementia or mild cognitive impairment with biomarker confirmation. No regenerative product has been tested in cognitively unimpaired individuals for prevention. There is no trial, no cohort, and no biomarker endpoint study to extrapolate from.
Nor has any trial in this category published APOE-stratified results. For a population in which genotype demonstrably shapes disease biology — APOE ε4 homozygotes show near-complete penetrance of Alzheimer’s pathology with a median symptom onset age of 65.1 years, a pattern that parallels autosomal dominant Alzheimer’s disease and Down syndrome — the absence of genotype-stratified data is a substantive gap, not a technicality. Whether an anti-inflammatory cell therapy would behave differently in an ε4 carrier is a reasonable question. It is also a completely unanswered one.
A clinic that markets a regenerative product to an asymptomatic ε4 carrier is therefore making two extrapolations at once: from symptomatic disease to prevention, and from an unstratified population to a specific genotype. Neither is supported.
If you are evaluating a clinic offer
These questions are diagnostic. A legitimate research program answers all of them without difficulty.
- Is there an IND, and what is the number? Legitimate investigational cell therapy proceeds under an Investigational New Drug application. There is no version of this that is optional.
- Is it registered on ClinicalTrials.gov, and does the registration match what is being offered? Check the enrollment criteria and endpoints yourself.
- Which institutional review board approved the protocol? A named IRB with a functioning contact, not a reference to “ethics oversight.”
- Am I being charged? Participants generally do not pay to receive an investigational product. A five-figure fee is the clearest single indicator that this is not a trial.
- Is there a control group? If everyone receives the product, the study cannot establish that it works.
- What are the release specifications for this lot? Cell identity, viability, sterility, endotoxin, and donor screening should be documented and available.
- Has this specific product been published, in a journal, with results? Conference posters, press releases, and testimonial pages are not publications.
- Does the same clinic treat many unrelated conditions with the same product? A preparation offered for arthritis, autism, long COVID, and dementia is being sold, not studied.
- Who administers it, and what happens if there is a complication? Ask specifically about the plan for a neurologic adverse event, and who provides follow-up.
- Would my neurologist see the record? If the answer involves discretion or avoiding your existing physicians, that is disqualifying on its own.
One intravenous mesenchymal stem cell product has produced a randomized, placebo-controlled, peer-reviewed signal in mild Alzheimer’s disease that warrants a larger trial. It is not commercially available, and its cognitive result remains statistically compatible with no effect.
Everything sold under the regenerative label to people worried about their genotype — intravenous or intrathecal cell infusions, umbilical and Wharton’s jelly preparations, exosome drips, peptide protocols — rests on no controlled human evidence in Alzheimer’s disease, none in prevention, and none in ε4 carriers specifically.
Where genotype-stratified human data do exist, they point somewhere far less dramatic. In a population-based cohort of 4,807 older adults, statin initiation was associated with reduced incident clinical Alzheimer’s dementia among ε4 carriers (HR 0.60, 95% CI 0.49–0.74) but not among non-carriers (HR 0.96, 95% CI 0.82–1.12), interaction P=0.015. That is observational evidence requiring randomized confirmation, and the authors say so. It is still a stronger genotype-specific signal than anything the regenerative category has produced, where the corresponding number is zero.
References
All citations verified in PubMed. DOI links provided.
- Rash BG, Ramdas KN, Agafonova N, et al. Allogeneic mesenchymal stem cell therapy with laromestrocel in mild Alzheimer’s disease: a randomized controlled phase 2a trial. Nat Med. 2025;31(4):1257–1266. doi:10.1038/s41591-025-03559-0
- Kim HJ, Cho KR, Jang H, et al. Intracerebroventricular injection of human umbilical cord blood mesenchymal stem cells in patients with Alzheimer’s disease dementia: a phase I clinical trial. Alzheimers Res Ther. 2021;13(1):154. doi:10.1186/s13195-021-00897-2
- Xie X, Song Q, Dai C, et al. Clinical safety and efficacy of allogenic human adipose mesenchymal stromal cells-derived exosomes in patients with mild to moderate Alzheimer’s disease: a phase I/II clinical trial. Gen Psychiatr. 2023;36(5):e101143. doi:10.1136/gpsych-2023-101143
- Craft S, Raman R, Chow TW, et al. Safety, efficacy, and feasibility of intranasal insulin for the treatment of mild cognitive impairment and Alzheimer disease dementia: a randomized clinical trial. JAMA Neurol. 2020;77(9):1099–1109. doi:10.1001/jamaneurol.2020.1840
- Cummings JL, Atri A, Sano M, et al. Efficacy and safety of oral semaglutide 14 mg (flexible dose) in early-stage symptomatic Alzheimer’s disease (evoke and evoke+): two phase 3, randomised, placebo-controlled trials. Lancet. 2026;407(10544):2167–2179. doi:10.1016/S0140-6736(26)00459-9
- Wei ZH, He QB, Wang H, Su BH, Chen HZ. Meta-analysis: the efficacy of nootropic agent Cerebrolysin in the treatment of Alzheimer’s disease. J Neural Transm (Vienna). 2007;114(5):629–634. doi:10.1007/s00702-007-0630-y
- Chen N, Yang M, Guo J, Zhou M, Zhu C, He L. Cerebrolysin for vascular dementia. Cochrane Database Syst Rev. 2013;(1):CD008900. doi:10.1002/14651858.CD008900.pub2
- Fan F, Liu H, Shi X, Ai Y, Liu Q, Cheng Y. The efficacy and safety of Alzheimer’s disease therapies: an updated umbrella review. J Alzheimers Dis. 2022;85(3):1195–1204. doi:10.3233/JAD-215423
- Majidazar R, Rezazadeh-Gavgani E, Sadigh-Eteghad S, Naseri A. Pharmacotherapy of Alzheimer’s disease: an overview of systematic reviews. Eur J Clin Pharmacol. 2022;78(10):1567–1587. doi:10.1007/s00228-022-03363-6
- Belder CRS, Boche D, Nicoll JAR, et al. Brain volume change following anti-amyloid β immunotherapy for Alzheimer’s disease: amyloid-removal-related pseudo-atrophy. Lancet Neurol. 2024;23(10):1025–1034. doi:10.1016/S1474-4422(24)00335-1
- Berkowitz AL, Miller MB, Mir SA, et al. Glioproliferative lesion of the spinal cord as a complication of “stem-cell tourism.” N Engl J Med. 2016;375(2):196–198. doi:10.1056/NEJMc1600188
- Amariglio N, Hirshberg A, Scheithauer BW, et al. Donor-derived brain tumor following neural stem cell transplantation in an ataxia telangiectasia patient. PLoS Med. 2009;6(2):e1000029. doi:10.1371/journal.pmed.1000029
- Sloan EA, Sampognaro PJ, Junn JC, et al. Neuroglial stem cell-derived inflammatory pseudotumor (n-SCIPT): clinicopathologic characterization of a novel lesion of the lumbosacral spinal cord and nerve roots following intrathecal allogeneic stem cell intervention. Acta Neuropathol. 2019;138(6):1103–1106. doi:10.1007/s00401-019-02089-7
- Bauer G, Elsallab M, Abou-El-Enein M. Concise review: a comprehensive analysis of reported adverse events in patients receiving unproven stem cell-based interventions. Stem Cells Transl Med. 2018;7(9):676–685. doi:10.1002/sctm.17-0282
- Fortea J, Pegueroles J, Alcolea D, et al. APOE4 homozygozity represents a distinct genetic form of Alzheimer’s disease. Nat Med. 2024;30(5):1284–1291. doi:10.1038/s41591-024-02931-w
- Rajan KB, Mcaninch EA, Wilson RS, Dhana A, Evans-Lacko S, Evans DA. Statin initiation and risk of incident Alzheimer disease and cognitive decline in genetically susceptible older adults. Neurology. 2024;102(7):e209168. doi:10.1212/WNL.0000000000209168
- US Food and Drug Administration. Consumer alert on regenerative medicine products including stem cells and exosomes. fda.gov (regulatory source; cited for approval status only)
This article is educational and does not constitute medical advice. Decisions about investigational therapies should be made with your treating neurologist or physician, who can weigh your individual clinical circumstances.
