Where do stem cells come from?
When someone offers you "stem cell therapy," the first question worth asking is which tissue the cells came from. Three sources supply nearly all mesenchymal stem cell treatment: bone marrow, fat, and umbilical cord tissue. They are cousins, not clones, and the differences show up in what you receive, how it was made, and what the trials tested.
This article is the comparison, source by source. The mechanism behind all three, how MSCs signal rather than rebuild, is covered in our MSC therapy explainer, and the manufacturing side has its own article on how MSC products are made.
Bone marrow
The original source, and the one with the deepest research history. Bone marrow MSCs are drawn from the hip in a procedure done under local anesthetic, and in laboratory comparisons they consistently show one distinctive strength: when three laboratories grew all three cell types under one harmonised protocol, bone marrow cells drove blood-vessel formation most strongly.
The trade-offs are real. Harvesting is a procedure with its own discomfort and small risks; the cells come from the patient, so they are as old and as metabolically stressed as the patient; and MSC content in marrow is strikingly low, with cells meeting the MSC definition making up roughly 0.08% of the bone marrow mononuclear fraction.
Adipose (fat)
Fat is the most abundant source and the easiest to harvest in volume, through liposuction. In that same harmonised three-laboratory comparison, adipose-derived cells were the strongest immunosuppressors, which matters when the goal is calming an overactive immune response rather than building vessels.
The trade-offs mirror bone marrow: it is still a procedure, still autologous, and donor health follows the cells. Obesity is associated with senescence and a pro-inflammatory shift in adipose-derived cells, and long-standing diabetes with intrinsic dysfunction of bone-marrow MSCs, both reviewed as reasons an autologous product from a metabolically unwell donor may underperform. The irony is not subtle: the patients most likely to want the therapy are sometimes the least likely to supply strong cells for it.
Umbilical cord
Cord tissue flips the model. The cells come from donated tissue after a healthy full-term birth, so the donor is young and screened, the patient has no harvesting procedure at all, and one donation can be expanded into many banked doses. That combination is why cord-derived lines are the standard offering in clinics worldwide, and it is covered in depth in our umbilical cord MSC article.
The trade-off is that you are receiving someone else's cells. In practice, MSCs display little of what the immune system uses to spot foreign tissue, and randomized trials have given unrelated recipients cord cells without matching or immunosuppression. The honest caveat lives in repeat dosing, where an early pediatric study saw donor-directed antibodies develop without clinical consequence so far.
The bigger divide: same-day versus culture-expanded
Here is the split that matters more than tissue source, and the one most people never hear about.
Same-day preparations are made during the visit: bone marrow aspirate concentrate is marrow drawn and spun, and stromal vascular fraction is what remains after fat is digested and spun. These are legitimate procedures containing regenerative cells, and they are mixtures in which actual MSCs are a small minority. A 2025 side-by-side characterisation of the two found roughly 1,000 MSCs per millilitre in each, with different soluble-factor profiles and different effects on cartilage cells in the laboratory.
Culture-expanded products are grown in a laboratory over weeks until a small sample becomes a full dose of verified cells, which is what allows identity checks, potency testing, and controlled cryopreservation. Every MSC product used in a registered clinical trial went through that. And critically, expansion is available from all three tissues: cultured bone marrow, cultured adipose, and cultured cord cells all exist and all appear in trials. Source and preparation are two separate questions, and confusing them is the most common mistake in this whole conversation.
What the head-to-head trial found
One trial tested the sources against each other directly. The MILES trial compared single injections of bone marrow aspirate concentrate, adipose stromal vascular fraction, and culture-expanded umbilical cord MSCs against a corticosteroid injection in knee osteoarthritis. In 440 analyzed patients at twelve months, none of the three cell preparations outperformed the steroid.
Read that carefully, because it is easy to over- or under-read. It tested one injection of each, at one dose, in one condition, against an active comparator that itself relieves pain. It did not test repeat dosing, which has beaten single dosing in two small randomized trials, and only one of its three arms used culture-expanded cells. What MILES genuinely establishes is that a single shot of any of these preparations did not beat a steroid shot in a knee.
How to think about choosing
In practice, the choice is usually made by what is available and legal where you are treated, and by the condition. Broadly: autologous same-day preparations dominate US orthopedic clinics, where regulation restricts expansion; culture-expanded cord products dominate international regenerative practice, and are what the immune and systemic protocols use.
Practitioners tend to reason by tissue and target, bone marrow toward orthopedic and vascular repair, adipose toward immune modulation, cord toward systemic and immune-driven conditions, and that reasoning outruns the head-to-head evidence, because almost none of it exists. What is documented is that the sources differ in the laboratory, that expansion changes the product profoundly, and that donor health matters when the donor is you.
The pros and the cons
What's clear
- All three sources yield real MSCs that meet the standard laboratory definition.
- The sources differ measurably: bone marrow strongest for vessel formation, adipose strongest for immune suppression.
- Cord tissue avoids a harvesting procedure and supplies young, screened, bankable cells.
- Any of the three can be culture-expanded into a tested product.
What's uncertain
- Head-to-head clinical comparisons barely exist; MILES is the notable exception and tested single injections.
- Laboratory differences between sources may or may not translate into different patient outcomes.
- Same-day preparations are minority-MSC mixtures with no release testing.
- Donor age, diabetes, and obesity degrade autologous cell quality in ways hard to predict per person.
Worth asking
- Which tissue, and is this same-day or culture-expanded? These are separate questions.
- If autologous: how does my age and metabolic health affect what you will get from me?
- If donated: what donor screening, what bank, what potency and post-thaw viability numbers?
- What has actually been tested in a trial for my condition, with this preparation?
What the key studies tested
Our framing rule for evidence: a study tests one preparation, in one condition, on one endpoint. Where the evidence sits:
Preclinical & practice
Most source-selection reasoning, and the laboratory comparisons behind it.
Early trials
Repeat-dosing comparisons; source-specific trials in individual conditions.
Late-stage trials
MILES: three preparations against corticosteroid in 440 knee-osteoarthritis patients.
Approved uses
One US-approved MSC product, from bone marrow, for children's graft-versus-host disease.
Every study behind this article is filterable in our research library on the Science page.
Questions people ask
Which source is best?
No trial supports a general winner, and anyone who tells you otherwise is ahead of the evidence. Laboratory comparisons show real differences (bone marrow leads on vessel formation, adipose on immune suppression), cord tissue wins decisively on practicality and donor quality, and the one head-to-head clinical trial found no difference between preparations in a knee. Condition, availability, and preparation method matter more than source alone.
Are my own cells safer than donated ones?
Intuitively yes, in practice it is more complicated. Autologous cells avoid any donor-immune question, and they carry your age and metabolic health with them: diabetes and obesity are both associated with degraded MSC function. Donated cord cells come from screened, healthy, young tissue and have been given to unrelated recipients in trials without matching or immunosuppression. Both routes have real safety records; neither is automatically safer.
What is the difference between BMAC, SVF, and "stem cells"?
BMAC is concentrated bone marrow, SVF is the cell fraction from digested fat, and both are same-day mixtures where MSCs are a small minority, roughly a thousand per millilitre in a 2025 side-by-side study. Culture-expanded MSCs are a grown, counted, tested product. All three get marketed as "stem cell therapy," and they are meaningfully different medicines.
Does the MILES trial mean stem cells do not work for knees?
It means a single injection of those three preparations did not beat a single corticosteroid injection at twelve months in 440 patients. That is a real, important result about that regimen. Repeat dosing beat single dosing in two smaller randomized trials, and only one MILES arm used culture-expanded cells. The honest reading is regimen-specific rather than a verdict on the whole approach.
What to take away
If you remember five things from this article, make them these:
- Bone marrow, fat, and cord tissue all yield real MSCs, with measurable laboratory differences and almost no head-to-head clinical data.
- The bigger divide is preparation: same-day BMAC and SVF are minority-MSC mixtures; culture-expanded products are grown and tested.
- Expansion is possible from all three sources, so source and preparation are separate questions.
- With autologous cells, your age and metabolic health ride along; donated cord cells come from screened young tissue.
- MILES tested single injections of all three against a steroid and found no winner, which is a fact about that regimen, not the field.
The evidence
Selected references, each verified against primary sources (PubMed and ClinicalTrials.gov). Explore the full, filterable research library on our Science page.
This article is for educational purposes only and is not medical advice, a diagnosis, or a treatment recommendation. MSC source comparison is discussed in the context of the published research; inclusion of a study does not imply a guaranteed outcome. Many of these compounds are investigational and not approved for the uses described in all jurisdictions. Any treatment decision should be made with a qualified physician. Individual results vary.