Evidence reviewed through: July 2026
Audience: Healthcare professionals, researchers, and informed patients
On the evidence in this article. Every citation has been individually verified: the DOI resolves,
the paper is the one named, and the paper says what we claim it says. Where the evidence is strong we
report it at full strength with effect sizes and confidence intervals; where it is weak or contested,
that is stated rather than smoothed over.
We cite older work where the older work remains the best evidence on a question. Several of the
foundational mechanism papers here date from 2005–2012 and have not been superseded; replacing them
with newer but weaker reviews would make the page look current at the cost of making it worse.
1. Historical Context and Nomenclature
1.1 Discovery
Alexander Friedenstein identified an adherent, fibroblast-like, colony-forming population in bone marrow capable of osteogenic differentiation — initially termed colony-forming unit fibroblasts (CFU-F). [1]Pittenger and colleagues subsequently demonstrated tri-lineage differentiation from human marrow, establishing multipotency. [2]
1.2 The nomenclature question — and the current society position
Arnold Caplan, who coined the term, later argued for "medicinal signalling cells" on the grounds that the therapeutic mechanism is paracrine signalling rather than differentiation. [4]
The definitive current position is the 2019 ISCT Mesenchymal Stromal Cell committee statement, which supersedes informal usage. It supports continued use of the acronym "MSCs" but recommends that it be:
- supplemented by tissue of origin, to highlight tissue-specific properties;
- taken to mean "mesenchymal stromal cells" unless rigorous evidence for stemness exists, supported by both in vitro and in vivo data; and
- associated with a robust matrix of functional assays demonstrating MSC properties — not generically defined, but informed by the intended therapeutic mode of action. [5]
That second point deserves emphasis on a page like this one. For most clinical preparations, the in vivo stemness evidence does not exist, so "mesenchymal stromal cell" is the accurate term. We have retitled this article accordingly and use "MSC" throughout to mean stromal cells.
2. Defining Characteristics: The ISCT Criteria
The 2006 ISCT minimal criteria remain the field's standard definition, and nothing since has replaced them. [3]
1. Plastic adherence under standard culture conditions.
2. Surface marker expression
Positive (≥95%): CD105, CD73, CD90.
Negative (≤2%): CD45, CD34, CD14 or CD11b, CD79α or CD19, HLA-DR.
3. Tri-lineage differentiation in vitro — osteoblasts (Alizarin Red), adipocytes (Oil Red O), chondrocytes (Alcian Blue/Safranin O).
Recognised limitations
- Heterogeneity. Populations meeting the criteria remain heterogeneous, with variable potency.
- No functional requirement. The criteria do not address immunomodulatory capacity — arguably the most clinically relevant property. The 2019 ISCT statement explicitly calls for functional assays matched to the intended mechanism of action. [5]
- Source-specific differences. Cells from different tissues meet the criteria but differ functionally. [6]
- Culture-induced drift. Extended expansion alters characteristics while criteria are still met.
3. Tissue Sources
3.1 Bone marrow (BM-MSC)
The most extensively studied source, and the one with the largest safety dataset. Comprises roughly 0.001–0.01% of nucleated marrow cells; strong osteogenic capacity; age-dependent decline in number and function. [7]Requires aspiration and expansion.
3.2 Adipose tissue (AD-MSC)
Substantially higher yield per gram than marrow. [8]Requires liposuction — a procedure with its own complication profile. Cell quality is influenced by donor metabolic status.
3.3 Umbilical cord (UC-MSC)
Derived from Wharton's jelly. Higher proliferative capacity than adult-derived MSCs [9]and low HLA expression. [10]No donor procedure; tissue otherwise discarded.
An important counterweight. Meta-analysis has established the safety of bone marrow MSC infusion. Specialists have explicitly cautioned that "it is not clear if this also applies to diverse new cell products derived from other sources, such as adipose and perinatal tissues," because these products display varying levels of highly procoagulant tissue factor. [11]Umbilical cord is a perinatal source. See §7.3.
3.4 Comparative summary
The last two rows are the ones usually omitted from comparisons of this kind. Bone marrow has the strongest safety evidence and the lowest procoagulant activity; perinatal and adipose sources have the highest. [11][12]
4. Mechanisms of Action
Early expectations centred on engraftment and differentiation. The evidence shows this occurs rarely. [13]
4.1 The paracrine paradigm
MSCs function principally by secreting bioactive factors that modulate the local microenvironment. [4]Supporting observations: MSCs rarely persist long-term; benefit occurs without engraftment; conditioned medium reproduces many effects.
Human autopsy data supports this directly — donor DNA was detectable at low levels declining with time since infusion, with no sustained engraftment and no ectopic tissue formation. [14]
4.2 The secretome
MSCs also release extracellular vesicles carrying microRNA and protein cargo. Preclinical work has shown vesicle-mediated transfer of PCNA can rejuvenate aged bone-marrow MSCs and slow age-related degeneration in mice — a mechanistically interesting finding that is animal data, not a demonstrated human anti-ageing effect. [15]
4.3 Immunomodulation
These mechanism papers are older and remain the best primary evidence on their specific questions; nothing since has superseded them.
- T cells — suppression via IDO, PGE2, TGF-β and PD-L1 contact; induction of regulatory T cells. [16]
- B cells — inhibition of proliferation, plasma-cell differentiation and antibody production. [17]
- NK cells — suppression via IDO and PGE2; the interaction is bidirectional, and activated NK cells can lyse MSCs. [18]
- Dendritic cells — impaired maturation and antigen presentation, promoting tolerogenic phenotypes. [19]
- Macrophages — polarisation toward the anti-inflammatory M2 phenotype. [20]
4.4 The "hit and run" model
Administration → entrapment (lungs after IV; local after intra-articular) → licensing by inflammatory signals → secretion → clearance within days to weeks → persisting immune and tissue effects. [14]This explains benefit without long-term engraftment.
5. Immunogenicity and Allogeneic Use
5.1 Immune evasive, not immune privileged
MSCs have long been described as hypoimmunogenic or immune privileged. The current position is more precise: "recent studies describing generation of antibodies against and immune rejection of allogeneic donor MSCs suggest that MSCs may not actually be immune privileged." Whether such rejection affects efficacy is unknown, and patient sensitisation toward donor antigens is a recognised consideration. [21]
A subsequent review reaches the same conclusion from the MHC-matching literature: studies controlling for MHC expression report both cell-mediated and humoral immune responses to MHC-mismatched MSCs, with clinical implications still unknown. [22]
5.2 What is genuinely true about MSC immunobiology
The following remain well supported:
- Low MHC class I and absent MHC class II expression under basal conditions (class II is inducible by IFN-γ)
- Absent co-stimulatory molecules (CD80, CD86)
- Active suppression of alloimmune responses
These properties make allogeneic use feasible without routine immunosuppression, and this is borne out in practice: thousands of patients have received allogeneic MSCs across clinical trials without routine immunosuppressive cover, and the major safety reviews find no excess of immune-mediated harm. [42]⁻ [44]"Immune evasive" describes this accurately; "immune privileged" overstates it.
5.3 What this means practically
- Immune responses to donor cells are documented, not theoretical.
- Their clinical significance is genuinely uncertain.
- The consideration is greatest with repeat allogeneic dosing, where sensitisation is the theoretical concern.
- Monitoring is reasonable; routine immunosuppression is not standard practice.
5.4 What the translational record shows
Allogeneic MSCs are the basis of the field's first approved product (§9), and allogeneic preparations dominate current commercial development because they permit standardised, off-the-shelf manufacture. At the same time, an analysis of the field's translational record notes that "the outcomes of advanced clinical trials have fallen short of expectations raised by encouraging pre-clinical animal data," and identifies pharmacological and manufacturing disparities between preclinical and human studies as a route to better trial design. [23]Both statements are true simultaneously: the allogeneic route is viable and productive, and the efficacy record has been uneven.
6. Clinical Applications
6.1 Knee osteoarthritis
The most studied indication — and the one where the gap between citation and conclusion is widest across this corpus.
The favourable synthesis. A 2023 review evaluated 15 randomised and 11 non-randomised trials of culture-expanded MSCs, finding net positive effects on pain and symptoms (function improved in 12/15 RCTs versus baseline, 11/15 versus control) and on cartilage protection or repair (18/21 studies). The authors are careful: "The relatively small sample size of 610 patients limited the drawing of definitive conclusions," and they describe trends toward benefit at moderate-to-higher doses in selected patient phenotypes, concluding with a roadmap for better-designed trials rather than a recommendation. [25]
The unfavourable synthesis. A systematic review of stem cell injection for knee OA found five RCTs and one non-RCT, judged all trials at high risk of bias, yielding level-3 evidence, and concluded: "In the absence of high-level evidence, we do not recommend stem cell therapy for KOA." [26]
The most recent meta-analysis. Twenty-eight randomised trials: significant improvement in pain (ΔVAS mean difference −1.67) and several function domains, but WOMAC, KOOS quality-of-life and Lequesne scores non-significant, and MRI-based structural scores non-significant. Conclusion: the role is "primarily symptom-modifying rather than structure-modifying." [27]
How much is the cells. Across placebo-controlled trials, contextual (placebo-related) effects accounted for roughly 63% of pain reduction and 61% of functional improvement at 6 months, and about 50% of pain relief at 12 months; MSCs conferred "only a modest incremental benefit." GRADE certainty: low. [28]
Society position. The 2025 ESSKA-ORBIT consensus supports cell therapy for Kellgren-Lawrence grades 1–3 as a second-line option, notes benefits are inferior in KL 4, and rated 22 of its 27 statements as low-level evidence or expert opinion. [29]
Individual trials of note. A phase I/II RCT found BM-MSC plus hyaluronic acid superior to HA alone at long-term follow-up (n=30). [30]A randomised trial of adipose MSCs (n=30, unblinded, conservative-management control) reported pain and functional improvement at 12 months with MRI indicating modification of disease progression. [31]Both are small; the second is not placebo-controlled.
On durability: benefit is consistently demonstrated to about 12 months. [29]The longest placebo-controlled trial followed patients to 60 months, finding improvement at 6, 12, 24 and 36 months and then decline "with scores nearing baseline at 60 months" (29 completers). [32]An uncontrolled 5-year series of 11 patients reported maintained improvement [33]— a difference the contextual-effects finding above largely explains.
6.2 Graft-versus-host disease — the strongest indication
Two 2025 meta-analyses of the same four randomised trials (650 patients) assessed MSCs added to second-line therapy for steroid-refractory acute GVHD:
- Overall response: RR 1.13 (95% CI 1.03–1.23, P = 0.007) [58]; independently RR 1.13 (95% CI 1.04–1.23, P = 0.005) [59]
- Complete response: RR 1.43 (95% CI 1.19–1.70, P < 0.001) [58]; RR 1.60 (95% CI 1.35–1.90, P < 0.00001) [59]
- Reduced chronic GVHD: HR 0.60 (95% CI 0.42–0.86, P = 0.005) and improved failure-free survival: HR 0.72 (95% CI 0.54–0.95, P = 0.022) [58]
- Benefit greatest in grade III–IV disease and gut, skin or multi-organ involvement
- No increase in adverse events (RR 1.00, P = 0.96) or serious adverse events (RR 1.01, P = 0.79) [59]
Overall survival was not improved (RR 1.05, 95% CI 0.90–1.23, P = 0.52), and both analyses call for larger standardised trials with longer follow-up. [58][59]This is the indication where MSC therapy has performed best, and it is the basis of the first US approval (§9).
6.3 Respiratory
COPD. A meta-analysis of 11 studies (371 patients) found a strong tendency toward but not statistical significance in FEV1 improvement (+71 mL, 95% CI −2 to 145, p = 0.056), a significant increase in 6-minute walk distance (52 m, 95% CI 18–87, p < 0.05), and no reduction in hospitalisation risk for acute exacerbations (RR 0.77, p > 0.05). The authors state the evidence "comes from small clinical trials" and call for large, well-designed RCTs. [34]
COVID-19 lung damage. A phase 2 randomised, double-blind, placebo-controlled trial of UC-MSCs in 100 patients with severe COVID-19 did not meet its primary endpoint — change in whole-lung lesion volume differed by −13.31% (95% CI −29.14% to 2.13%, p = 0.080). A component secondary outcome (solid-component lesion volume) was significant (p = 0.043); 6-minute walk difference was not (p = 0.057). Adverse event incidence was similar between groups. [35]
The trial did not assess mortality or oxygenation. Its authors state that "a phase 3 trial is required to evaluate effects on reducing mortality and preventing long-term pulmonary disability." [35]
6.4 Autoimmune disease
Investigated in rheumatoid arthritis, lupus, multiple sclerosis, Crohn's disease and type 1 diabetes. One product — darvadstrocel (Alofisel), expanded allogeneic adipose MSCs — was approved in the EU for complex perianal fistulas in Crohn's disease on the basis of a phase 3 randomised double-blind trial. [36]
Multiple sclerosis merits specific note because route matters: a 2024 meta-analysis of 30 studies found improvement in disability scores with intrathecal delivery (weighted mean difference −0.28; 95% CI −0.53 to −0.03) but no significant change with intravenous delivery, and concluded effectiveness "remains uncertain." [37]
6.5 Cardiovascular
Meta-analyses suggest small effects, with paracrine mechanisms dominating and direct cardiomyocyte differentiation rare. [38]Optimal timing, route and dose remain unresolved.
6.6 Ageing and frailty
The CRATUS trial — phase II RCT, 30 patients randomised to placebo, 100 million or 200 million allogeneic MSCs by IV infusion — reported improved 6-minute walk distance, improved pulmonary function, reduced TNF-α, and no serious adverse events. [39]
A 2024 UC-MSC trial — phase I/II randomised, double-blind, placebo-controlled — reported improved SF-36 physical component scores from the first post-treatment visit (p = 0.042), improved EQ-VAS (p = 0.002 at 6 months), improved timed-up-and-go (p < 0.05), improved grip strength (p = 0.002 at 6 months), and reduced TNF-α and IL-17 (p = 0.034, 0.033). Adverse event incidence did not differ between groups. Note that the 4-metre walk test did not reach significance (p = 0.21) — an omission in our previous summary. [40]
These are encouraging early-phase results in small samples. They are not evidence that MSC therapy slows ageing.
6.7 The pattern worth naming
Across respiratory, cardiac and neurological applications, the dominant result is a null primary endpoint with favourable secondary or subgroup findings. Those are hypothesis-generating. More broadly: most clinical-stage MSC therapies have not met their primary efficacy endpoints, and MSC function in humans has proven "not as robust as demonstrated in preclinical studies." [41]
7. Safety
7.1 What the major reviews found
All three examine intravascular administration in adults.
Two findings in these tables are easy to read past. No trial in any of the three reviews was stopped early for safety reasons. And in the 2020 update, *mortality was significantly lower in MSC groups (RR 0.78, 95% CI 0.65–0.94)* — a confidence interval that excludes 1. That result comes from trials in seriously ill populations rather than elective patients, and it was not a prespecified efficacy endpoint, so it should not be read as evidence that MSCs extend life. But it is a positive signal and it belongs on the record alongside the fever finding.
The SafeCell authors also noted a caution we had omitted: because MSC use "may be associated with neoplastic growth long term," it was difficult to understand why approximately 50% of studies did not report follow-up duration for adverse events. [42]
7.2 Adverse events
Consistently and significantly associated: transient fever, typically resolving within 24–48 hours.
Commonly reported: fatigue, headache, infusion-related reactions. For intra-articular administration: injection-site pain (RR 2.04) and joint swelling (RR 3.39) versus control, with serious complications uncommon. [27]
Reported but not statistically associated: infection, thrombotic events, organ complications, death, malignancy.
7.3 Thrombotic risk and pulmonary embolism
Thrombotic risk after MSC infusion is sometimes described as a mechanical problem — cell clumps lodging in the pulmonary microvasculature, solved by careful preparation. The mechanism is biological rather than mechanical, and it is the most credible serious-harm pathway in the field, so it is set out in full here.
- MSCs express tissue factor — the initiator of the extrinsic coagulation pathway — and phosphatidylserine. Expression varies with cell dose, tissue source, passage number and culture conditions, and the authors state procoagulant activity "may limit their clinical safety." [12]
- Tissue factor is highest in adipose- and umbilical-cord-derived MSCs, and lowest in bone-marrow-derived MSCs. MSCs of diverse origin induced fibrin clots in healthy plasma, partially suppressed by an anti-tissue-factor antibody. [45]
- In humans, intravenous UC-MSC infusion produced a transient rise in D-dimer, while intrathecal infusion did not. [45]
- Tissue factor is not a reliable predictive biomarker — no correlation with D-dimer or other coagulation indicators. [45]
- The authors recommend local administration where feasible, consideration of anticoagulation prophylaxis, and monitoring of coagulation indicators — decisions for treating clinicians, not for patients to act on independently. [45]
- This is the same instant blood-mediated inflammatory reaction described in the product-diversification literature, [11]and it is triggered more strongly by freeze-thawed cells. [46]
Holding this honestly: randomised trials have not shown a significant excess of thrombotic or embolic events. [43][44]But those trials are small and not powered for rare events, the procoagulant mechanism is demonstrated rather than theoretical, and the human biomarker signal is specific to the intravenous route. The accurate statement is that a real procoagulant effect exists, has not produced a detectable excess of clinical events to date, and is the reason product source, dose, route and coagulation monitoring matter.
7.4 Infection and manufacturing
Trial safety describes carefully manufactured products under protocol. A public health investigation of umbilical cord blood–derived products marketed as stem cell treatment identified 20 patients with bacterial infections, of whom all but one required hospitalisation; 65% (22 of 34) of unopened vials sampled were contaminated, with sequencing indicating contamination during processing. [47]
7.5 Tumorigenicity
No MSC-derived tumours have been documented in the published clinical literature. Supporting evidence: the autopsy series above found no ectopic tissue or donor-origin tumours (18 patients) [14]; a 10-year follow-up of 72 patients treated for knee chondral defects found no tumours or deep infections [48]; and the large meta-analyses found no malignancy association.
What this does not establish: sample sizes are small, most follow-up is under five years, and tumour latency can exceed that. No documented cases, sound biological rationale, insufficient follow-up to exclude rare or late events is the defensible position.
7.6 Long-term data
Genuinely limited. Most trial follow-up is under five years; roughly half of early trials did not report adverse-event follow-up duration at all. [42]The longest good-quality human follow-up we could identify is 10 years in 72 patients. [48]
8. Dose
8.1 The evidence does not support "more is better"
It is often assumed that higher cell counts produce better outcomes. The best-controlled dose-ranging trial in knee osteoarthritis found the opposite.
The trial cited in support found the opposite. A randomised placebo-controlled dose-ranging study of pooled allogeneic BM-MSCs in knee OA reported:
- a trend to improvement in the 25-million-cell group across VAS, ICOAP and WOMAC — not statistically significant versus placebo;
- "adverse events were predominant in the higher dose groups (50, 75, and 150 million cells)", most commonly knee pain and swelling;
- whole-organ MRI score showing no difference from baseline or placebo;
- conclusion: "A twenty-five-million-cell dose may be the most effective among the doses tested." [49]
8.2 What the wider dose literature shows
- An analysis of 914 registered trials found minimum effective intravenous doses clustering around 100–150 million cells, *"with lower and higher doses being less effective."* [50]
- A network meta-analysis of 11 randomised knee OA trials found moderate doses produced the greatest long-term improvement, with high doses producing more adverse events. [51]
- Repeated dosing outperformed single dosing at 6 and 12 months in knee OA, but with a higher incidence of adverse events. [52]
- Dose also interacts with procoagulant activity (§7.3), which is dose-dependent. [12]
The honest summary: dose matters, the optimum is unresolved and probably condition-specific, and there is no evidence base for a simple "more cells is better" rule. Specific cell counts, routes and frequencies by condition are protocol decisions for a treating clinician rather than educational content, and are not published here.
9. Regulatory Status
9.1 United States
Culture-expanded MSCs are regulated as biological products; minimal-manipulation and homologous-use exemptions rarely apply. Clinical trials require an IND; marketing requires a BLA.
In December 2024 the FDA approved Ryoncil (remestemcel-L-rknd), an allogeneic bone-marrow MSC therapy, for paediatric steroid-refractory acute graft-versus-host disease — the first MSC product approved in the United States. The ISCT MSC committee described the approval as having "finally ended a long-lasting drought on approved MSC clinical products in the United States," achieved by "addressing concerns regarding clinical efficacy and consistent MSC potency through an iterative process that spanned several years," and expected it to "revive investment and enthusiasm in MSC products." [53]
*What has not changed:* no MSC product is FDA-approved for osteoarthritis, anti-ageing, or general wellness indications. The approval is for one product, in one severe condition, after years of iterative work with the regulator on potency and efficacy.
9.2 Other jurisdictions
International guidance holds that substantially manipulated or non-homologous stem cell products are drugs or advanced therapy medicinal products — aligned with the FDA, EMA and TGA — and cautions "forcefully… against the premature commercialization of unproven stem cell-based interventions." [54]
10. Manufacturing and Quality
10.1 GMP requirements
Validated processes, identity/purity/potency testing, sterility assurance, traceability, and appropriate cleanroom classification.
10.2 Quality attributes
Identity: ISCT surface marker profile, morphology.
Purity: absence of haematopoietic and endothelial cells; sterility (bacterial, fungal, mycoplasma); endotoxin; viral testing.
Potency: the greatest unsolved challenge — no standardised assay exists. The 2019 ISCT statement calls for functional assays matched to the intended mechanism of action. [5]
Viability: verified at release and, for cryopreserved products, post-thaw.
10.3 Cryopreservation
Cryopreservation is often assumed to degrade potency. Where post-thaw viability is preserved, it appears not to: using modified cryopreservation methods, cells retained >95% viability, remained responsive to inflammatory signals, suppressed activated peripheral blood mononuclear cells, and "performed as well as fresh MSC" in an in vivo ischaemia/reperfusion model. [55]
The genuine state of the question is contested:
- For fresh: freeze-thawed MSCs show reduced responsiveness to inflammatory stimuli, impaired anti-inflammatory mediator production, increased IBMIR triggering and complement activation immediately after thawing. [46]
- Against it mattering: a systematic review of 18 comparative preclinical studies — 257 in vivo efficacy experiments across 101 outcome measures — found only 2.3% significantly different, of which two favoured fresh and four favoured cryopreserved. [56]And the study above found no difference when post-thaw viability is preserved. [55]
Honest position: there is a documented post-thaw impairment that appears to be mitigable by manufacturing method, and no demonstrated difference in clinical outcomes.
11. Challenges and Future Directions
Standardisation and heterogeneity. Products from different manufacturers — or different batches — vary in potency. Single-cell characterisation, mechanism-matched potency assays and quality-by-design manufacturing are active areas.
Patient selection. No validated biomarkers predict response. This is arguably the field's central unsolved problem.
Cell-free approaches. Given the paracrine mechanism, MSC-derived exosomes and conditioned medium are under investigation, with potential manufacturing and stability advantages. Clinical evidence remains early.
Engineered MSCs. Enhanced survival, homing or secretion — with additional regulatory complexity and a longer safety runway.
Combination therapies. MSCs with PRP, hyaluronic acid or rehabilitation protocols.
12. Practical Considerations
Questions that distinguish a well-characterised product
- What is the tissue of origin, and what safety data exists for that source by that route?
- What passage number? (In one clinical comparison, one-year survival was 75% with early-passage versus 21% with later-passage cells. [57])
- What potency assay was used, and is it matched to the intended mechanism? [5]
- Will the certificate of analysis — count, viability, sterility, endotoxin, mycoplasma — be available for the specific batch?
- What coagulation considerations apply, and is monitoring performed? (§7.3)
Realistic expectations
Symptom improvement is commonly reported and genuine; a substantial share is contextual. [28]Structural repair is not demonstrated. [27]Benefit is consistently demonstrated to about 12 months. [29]Response cannot currently be predicted for an individual.
13. Conclusion
Mesenchymal stromal cells are the most clinically translated cell type in regenerative medicine, and the mechanistic account of how they work — paracrine signalling and immunomodulation rather than tissue replacement — is well supported and genuinely interesting.
The clinical evidence divides sharply by indication. In steroid-refractory acute graft-versus-host disease, randomised evidence across 650 patients shows significantly improved overall and complete response, reduced chronic GVHD and improved failure-free survival, with no increase in adverse events — and December 2024 brought the first US approval on that basis. Randomised trials support expanded allogeneic MSCs for perianal fistulas in Crohn's disease, approved in the European Union. Early-phase controlled trials in ageing frailty report improved physical performance and reduced inflammatory markers. Across three safety reviews covering more than 6,000 patients, no trial was stopped early for safety.
Elsewhere the case is not yet made. Most advanced trials outside these indications have not met their primary endpoints. In knee osteoarthritis, the largest systematic review recommends against the therapy, the most recent meta-analysis finds it symptom-modifying rather than structure-modifying, and roughly half to two-thirds of symptomatic benefit is contextual. Fever is significantly increased. MSCs are procoagulant in a source- and dose-dependent way. Long-term surveillance adequate to exclude rare harms has not been done.
The consistent pattern is that MSCs perform best in severe, inflammation-driven disease with few alternatives, and least well in elective quality-of-life indications. The biology is sound, the strongest indications are genuinely established, and the marketing has run ahead of the evidence for most of the rest.
Explore condition-specific evidence: Knee Osteoarthritis · Safety Profile · Stem Cells vs PRP