Safety Profile of Mesenchymal Stromal Cell Therapy

Evidence reviewed through: July 2026

What "Safe" Means Here — and What It Doesn't

Safety is not a single property. It depends on which cells, how they were processed, how they are delivered, into whom, and for how long anyone looked afterwards. A regulatory review of stem cell risk makes this explicit: the risk profile depends on the type of cells, their differentiation status and proliferation capacity, the route of administration, the intended location, in-vitro culture and other manipulation steps, the irreversibility of treatment, and long-term survival of engrafted cells (Herberts, Kwa and Hermsen, 2011).

That same review's overall conclusion is worth quoting in full, because it is more conditional than it is usually made to sound: the majority of small clinical trials with MSCs "has not reported major health concerns, suggesting that MSC therapies could be relatively safe. However, in some clinical trials serious adverse events have been reported, which emphasizes the need for additional knowledge, particularly with regard to biological mechanisms and long term safety."

That is the honest starting position: reassuring, conditional, and incomplete.

The Trial Evidence

Three systematic reviews from the same Canadian research group form the backbone of MSC safety evidence. All three examined intravascular administration (intravenous or intra-arterial) in adults. None of them covers intra-articular injection into a joint.

SafeCell (2012) — the foundational review

Thirty-six studies, 1,012 participants, across ischaemic stroke, Crohn's disease, cardiomyopathy, myocardial infarction, graft-versus-host disease, and healthy volunteers. Eight were randomised controlled trials enrolling 321 participants. Follow-up across studies ranged from 0.5 to 60 months.

What it found:

• Meta-analysis of the RCTs did not detect an association between MSCs and acute infusional toxicity, organ system complications, infection, death, or malignancy.
• It did find a significant association between MSCs and transient fever — odds ratio 16.82 (95% CI 5.33–53.10) across four RCTs. All four used autologous cells.
• In uncontrolled trials, fever was reported in 24 of 62 participants (39%); acute infusional toxicity in 12 of 159 (8%).

(Lalu et al., 2012)

How to read "did not detect an association." That phrase is weaker, and different, from "no serious adverse events occurred." Serious events, including deaths, did occur in these trials; they were not statistically attributable to the cells, in populations where such events are common from the underlying disease itself. An absence of detected association across eight small randomised trials is also, in part, a statement about statistical power.

The authors' own caution deserves carrying: they noted that because "use of MSCs may be associated with neoplastic growth long term, it is difficult to understand why approximately 50% of studies did not report follow up duration for adverse events."

The 2020 update

Fifty-five studies, 2,696 patients, randomised trials only, intravascular delivery in adults.

• Increased risk of fever: relative risk 2.48 (95% CI 1.27–4.86), with no heterogeneity between studies.
• No association with non-fever acute infusional toxicity (RR 1.16), infection (RR 0.99), thrombotic/embolic events (RR 1.14), or malignancy (RR 0.93).
• Mortality was lower in MSC groups (RR 0.78, 95% CI 0.65–0.94) — a confidence interval that excludes 1. Note this is across seriously ill populations being studied for conditions like critical illness and cardiac disease, not elective wellness patients.
• No trial was stopped early for safety reasons.

(Thompson et al., 2020)

The 2026 umbilical cord update — the most relevant to cord-derived therapy

Because umbilical cord–derived cells differ biologically from bone marrow cells, the same group published a review confined to them: 42 randomised trials, 2,280 adults, intravascular delivery, searched to August 2025.

• Increased risk of fever: relative risk 2.34 (95% CI 1.02–5.35).
• No signal of harm for non-fever acute infusional toxicity, infection, death, malignancy or ectopic tissue formation, thrombotic/thromboembolic events, or mycotic aneurysms.
• No trial stopped early for safety.

Their conclusion, stated precisely: "UC-MSC therapy appears safe for study in heterogeneous patient populations" (Hum et al., 2026).

The phrase "safe for study" is the authors' own. It supports continued clinical investigation. It is not a statement that the therapy is proven safe as a marketed treatment, and it should not be read as one.

For joint injection specifically

The intravascular reviews above do not cover injection into a joint. The most current synthesis of intra-articular cell therapy in knee osteoarthritis — 28 randomised trials — reports a different and less favourable adverse event profile: significantly higher rates of injection-site pain (RR 2.04), joint swelling (RR 3.39), and other adverse events (RR 1.26), with serious complications uncommon and not statistically increased. The authors describe the profile as "generally manageable but non-negligible" (Awad et al., 2026).

In umbilical-cord-specific osteoarthritis trials — 8 studies, 688 participants — no serious treatment-related adverse events were reported and adverse events were "mild, transient, and manageable" (Lohrasbi et al., 2025).

The Fever Signal

Fever is the one adverse event that shows up consistently, significantly, and across every major analysis: OR 16.82 in the 2012 review, RR 2.48 in the 2020 update, RR 2.34 in the 2026 umbilical cord review.

It is mild and transient. It is also a genuine, reproducible, statistically significant effect of the treatment. Patients should expect a meaningfully increased chance of fever after infusion and should know in advance how to distinguish it from infection.

Why Product and Route Change the Answer

This is the most important section on this page, and the one most often skipped.

Cell source

A review of intravascular MSC products makes the point directly: meta-analysis has shown that bone marrow MSC infusion is safe, but "it is not clear if this also applies to diverse new cell products derived from other sources, such as adipose and perinatal tissues." Different MSC products carry varying levels of highly procoagulant tissue factor and may trigger an instant blood-mediated inflammatory reaction (IBMIR) — a clotting and inflammatory response when cells meet blood. The authors call for hemocompatibility testing and new clinical guidelines (Moll et al., 2019).

Umbilical cord tissue is a perinatal source, so that caution applies directly to cord-derived products. The review is often quoted for its first finding — that bone marrow MSC infusion is safe — without the warning that immediately follows it.

Autologous vs allogeneic

• Autologous — your own cells. No donor-matching question, but requires a harvest procedure (bone marrow aspiration or liposuction), which carries its own risks.
• Allogeneic — donor cells, including umbilical cord. No harvest procedure, available off the shelf, but introduces donor screening, manufacturing and immune questions (see below).

No definitive clinical advantage of autologous over allogeneic MSCs has been demonstrated (Ankrum, Ong and Karp, 2014).

Minimally manipulated vs culture-expanded

This distinction is regulatory as well as biological. International guidance adopts the standard that stem cell products are drugs or advanced therapy medicinal products if they have been substantially manipulated or are used for non-homologous purposes — a standard aligned with the US FDA, the European Medicines Agency, and Australia's TGA (Lovell-Badge et al., 2021).

Cells expanded in culture over weeks are substantially manipulated. Under that standard they are drug products requiring approval, not a same-day surgical procedure.

Route of administration

Route changes both what the cells do and what can go wrong:

• Intravenous — most cells are filtered by the lungs on first pass. In an anaesthetised rat model comparing cell types, the majority of MSCs were trapped in the lungs after intravenous infusion, "with a therapeutically questionable number of cells reaching the arterial system acutely" (Fischer et al., 2009). This is a preclinical animal study, not human safety data, but it explains why intravenous dosing behaves differently from local injection.
• Intra-articular — higher rates of local pain and swelling, as above.
• Other routes — intrathecal, intravitreal and intra-arterial administration have produced the most serious documented harms (below).

Immunogenicity: Immune Evasive, Not Immune Privileged

MSCs are still widely described as "immunoprivileged," with the implications that no HLA matching is required and that rejection has never been observed. That reflects an older understanding of the biology, and it overstates what the current literature supports.

The current position, from a review in Nature Biotechnology titled precisely "Mesenchymal stem cells: immune evasive, not immune privileged": MSCs have long been reported as hypoimmunogenic, "however, 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 not known. The authors also raise patient sensitisation toward donor antigens as a consideration (Ankrum, Ong and Karp, 2014).

A subsequent review is titled, equally plainly, "Immunoprivileged no more." It notes that the prevailing dogma has been that allogeneic MSCs are immune privileged, but studies controlling for MHC matching "have reported both cell-mediated and humoral immune responses to MHC-mismatched MSCs," and that the clinical implications remain unknown (Berglund et al., 2017).

What this means practically:

• Immune reactions to donor cells are documented, not hypothetical.
• Their clinical importance is genuinely uncertain — this is not a known danger, but it is no longer a settled non-issue.
• It is most relevant to repeat allogeneic dosing, where sensitisation is the theoretical concern.
• In one randomised trial of intravenous allogeneic umbilical cord MSCs in heart failure, none of the patients tested developed antibodies to the cells — but only 7 patients were tested (Bartolucci et al., 2017).

Tumour Risk: What Is Known and What Is Not

This is the concern patients raise most, and it deserves a careful answer rather than a reassuring one.

What supports low risk:

• An autopsy study examined tissue from 18 patients who had received HLA-mismatched MSCs, with 108 tissue samples from 15 patients analysed by PCR. No ectopic tissue formation and no malignant tumours of donor origin were found. Donor DNA was detectable in some tissues — lungs, lymph nodes, intestine — at levels of 1/100 to under 1/1,000, and detection declined with time since infusion. The authors conclude MSCs act by a "hit and run" mechanism, and that "the lack of sustained engraftment limits the long-term risks" (von Bahr et al., 2012).
• The large intravascular meta-analyses found no association with malignancy (Thompson et al., 2020; Hum et al., 2026).
• In a preclinical study, high and repeated doses of allogeneic human umbilical cord MSCs produced no tumour formation in immunocompromised mice over 26 weeks (Chin et al., 2024). This is an animal study of one manufacturer's product and does not establish human long-term safety.

What remains genuinely unresolved:

• The autopsy series is 18 patients. That is meaningful evidence, not definitive evidence.
• The SafeCell authors explicitly flagged that MSC use "may be associated with neoplastic growth long term" while criticising that half the trials did not report follow-up duration for adverse events (Lalu et al., 2012).
• Follow-up in most trials is months to a few years. Tumour latency can exceed that.

The accurate statement is: no MSC-derived tumours have been documented in the published clinical literature, and the biological rationale for low risk is reasonable — but the follow-up duration and sample sizes are not yet sufficient to exclude rare or late events. That is an open question, and it should be presented as one rather than as a settled reassurance.

Documented Harms Outside Clinical Trials

Trial safety data describes carefully manufactured products given under protocol with monitoring. It does not describe the commercial marketplace. The published record of harm from that marketplace is real and specific, and any honest safety page must include it.

Contaminated umbilical cord products

A US Centers for Disease Control and Prevention investigation examined bacterial infections in 20 patients who received umbilical cord blood–derived products marketed as stem cell treatment, across 8 states, between August 2017 and September 2018. The patients had sought treatment for pain, osteoarthritis, rheumatoid arthritis and injury.

• All but one required hospitalisation.
• The most frequent organisms were common enteric bacteria — E. coli (14 patients) and Enterobacter cloacae (7 patients).
• Of unopened, undistributed vials sampled, 65% (22 of 34) were contaminated with at least one of 16 bacterial species.
• Whole-genome sequencing matched patient isolates to product isolates across states, indicating a common source of extensive contamination, likely during processing of cord blood into product.

The investigators' conclusion: "Unapproved stem cell products can expose patients to serious risks without proven benefit" (Hartnett et al., 2021; see also the earlier field report, Centers for Disease Control and Prevention, 2018).

This is the most directly relevant published harm to anyone considering an umbilical cord–derived product, and it is why the manufacturing questions in the checklist below matter more than any marketing claim.

Serious harms by route

• Intravitreal (into the eye): three patients developed severe bilateral vision loss after intravitreal injection of autologous adipose-derived "stem cells" at a US clinic. Visual acuity before injection ranged from 20/30 to 20/200; one year afterwards it ranged from 20/200 to no light perception (Kuriyan et al., 2017).
• Intrathecal / stem cell tourism: a glioproliferative lesion of the spinal cord was reported as a complication following stem cell treatment obtained abroad (Berkowitz et al., 2016).
• Infection following clinic treatment: Mycobacterium abscessus infection was reported following treatment at an unregulated stem cell clinic (Olmedo-Reneaum et al., 2019).

These cases involve different cell types and routes from intravenous or intra-articular MSC therapy, and they should not be presented as typical. They are included because they are the documented severe-harm record, and because they share a common feature: treatment outside a regulated trial framework, with products of uncertain characterisation.

Regulatory Status

Being straightforward about regulatory status is part of safety, because it determines what oversight exists.

• International guidance (ISSCR) issues an updated recommendation to "forcefully caution against the premature commercialization of unproven stem cell-based interventions," and adopts the standard that substantially manipulated or non-homologous stem cell products are drugs or advanced therapy medicinal products, aligned with the US FDA, the EMA and Australia's TGA (Lovell-Badge et al., 2021).
• The same guidance notes that patient registries are not adequate substitutes for randomised controlled trials for demonstrating safety and efficacy, and that registries used as a form of advertising are "at best misleading and goes against a duty of care for patients."
• The CDC investigation above notes that clinics marketing stem cell products have increased "despite warnings from the US Food and Drug Administration that stem cell products for these and other indications have not been proven safe or effective" (Hartnett et al., 2021).

Sterling's position, stated plainly: cells used through our partner network are processed in Thailand under Thai FDA oversight, and patients travel for treatment. These treatments are not FDA-approved in the United States, and this page does not claim otherwise. Travelling for treatment is lawful and common; it also means the regulatory protections of your home country do not apply. That is a real trade-off and you should weigh it deliberately.

Reported Side Effects

Grouped by how consistently they appear in the published trial literature, rather than by a percentage attached to each item.

Consistently reported and statistically confirmed

• Transient fever — the one adverse event significantly associated with MSC infusion in all three major intravascular meta-analyses. Typically resolves within 24–48 hours.

Commonly reported in trials

• Acute infusional reactions (reported in 12 of 159 participants, 8%, in uncontrolled trials in the 2012 review)
• Fatigue and headache
• For joint injection: injection-site pain (RR 2.04) and joint swelling (RR 3.39) versus control (Awad et al., 2026)

Reported but not statistically associated with MSC treatment in meta-analysis

• Infection, thrombotic or embolic events, organ system complications, death, malignancy (Thompson et al., 2020; Hum et al., 2026)

Documented outside trials

• Serious bacterial infection from contaminated product (Hartnett et al., 2021)
• Severe route-specific injury (Kuriyan et al., 2017; Berkowitz et al., 2016)

We have deliberately not assigned a percentage to each item. The published trials report events in specific populations over specific windows; reproducing those as universal rates would misrepresent what was actually measured.

Manufacturing and Screening

The following describe operational practice, not research findings, and are not evidence of a safety outcome:

• Donor health history, and infectious disease screening including HIV-1/2, hepatitis B and C, syphilis and HTLV-I/II
• GMP-certified clean room processing with environmental monitoring
• Sterility testing at multiple stages, including culture-based testing
• Endotoxin testing and mycoplasma screening
• Viability confirmation and certificate of analysis per batch
• Traceability from donor to patient

These steps exist precisely because the failure mode is real: the CDC investigation found 65% of unopened vials from one producer contaminated at the processing stage. Rigorous manufacturing reduces that risk. No process eliminates it, and no clinic — including ours — should tell you otherwise.

Who Should Not Have This Treatment

Certain conditions make cell therapy inappropriate or require careful specialist evaluation. These are matters for a clinician who knows your history, and the list below is educational rather than a set of thresholds to apply to yourself:

• Active cancer or recent cancer history. The safety of MSCs in active malignancy is not established, and MSC secretion of growth factors raises a theoretical concern. This is the most important contraindication on the list.
• Active systemic infection.
• Bleeding disorders or anticoagulation, given the tissue factor and clotting considerations described above.
• Unstable cardiovascular disease.
• Pregnancy — safety data is limited.
• Significant immunosuppression, poorly controlled diabetes, or recent major surgery — each requires individual assessment.

We do not publish specific numeric thresholds. Deciding whether a particular laboratory value or medication dose precludes treatment is a clinical judgement about an individual, not something a web page should assert.

What We Do Not Know

Stated plainly, because a safety page that only lists reassurances is not a safety page:

• Very long-term outcomes. Meaningful human data beyond roughly 10–15 years does not exist for most applications. Roughly half of the early trials did not report follow-up duration for adverse events at all.
• Whether trial safety transfers to commercial practice. Trial products are characterised and monitored; marketed products vary enormously.
• Whether bone-marrow safety data applies to perinatal and adipose products. Specialists have explicitly said it may not.
• The clinical significance of immune responses to donor cells, particularly with repeated allogeneic dosing.
• Whether rare or late adverse events exist that current sample sizes and follow-up cannot detect.
• Optimal dosing, which remains unresolved across conditions.

Separately, and relevant to any risk–benefit judgement: most clinical-stage MSC therapies have been unable to meet primary efficacy endpoints, and the therapeutic functions of MSCs in humans are "not as robust as demonstrated in preclinical studies" (Levy et al., 2020). Weighing risk requires an honest view of benefit, and the benefit picture is weaker than marketing suggests.

Questions Worth Asking Any Provider

• What is the cell source — bone marrow, adipose, or umbilical cord — and what safety data exists for that source and that route?
• Was the product culture-expanded? Under what regulatory framework was it manufactured?
• Will I see the certificate of analysis — cell count, viability, sterility, endotoxin and mycoplasma results — for my specific batch?
• What is the sterility testing method, and are results available before administration?
• If I receive allogeneic cells more than once, is donor sensitisation monitored?
• What is the plan if a complication develops after I return home?
• What adverse events has this provider actually recorded, and over what follow-up period?

A provider who answers these specifically is giving you better information than one who offers a percentage without a denominator.

Take the Next Step

Every patient deserves a thorough understanding of risks and benefits before any medical decision. Our medical team is available to review your history, discuss risk factors specific to you, answer questions, provide the underlying literature, and coordinate with your own physicians.

[Schedule Safety Consultation]

This document is for educational purposes only and does not constitute medical advice. It describes published research, not outcomes at any particular clinic. Individual results vary. All treatments should be discussed with qualified healthcare providers. This reflects the published evidence as of July 2026.