CellGenic_logo
Provider Login
My Account
Become a Provider
CellGenic_logo

Exosome Safety: What the Published Literature Reports, and What It Does Not

M.Sc. Patricia Cruz Martinez Medically Reviewed By — Written By — Updated on September 16, 2026

On this page

Exosome Safety: What the Published Literature Reports, and What It Does Not
Physicians evaluating exosome-based research material tend to start with mechanism and end with sourcing. Risk rarely gets the same rigor. Part of that is a supply problem: “exosomes side effects” returns almost nothing written for a clinical audience. Part of it is that what does exist tends to land at one of two extremes: dismissal or alarm. Neither holds up against the actual record.

That record, drawn from registered trials, federal contamination investigations, and preclinical mechanism studies, is more specific than either extreme suggests. Reported adverse events in controlled studies are low and mostly mild. The serious harms that have actually reached hospitals trace to product quality and manufacturing control, not to some fundamental property of extracellular vesicles. Two questions, immunogenicity and tumorigenicity, remain genuinely open in the literature, not resolved in either direction [1–8]. Here’s what the evidence currently supports, and where it runs out.

What the Trial-Reported Adverse-Event Data Actually Shows

Across registered interventional studies, reported adverse events involving mesenchymal stromal cell–derived extracellular vesicles (MSC-EVs) are infrequent and generally mild. A 2025 systematic review and meta-analysis in Regenerative Engineering and Translational Medicine pooled 24 interventional trials (536 patients total) across conditions including COVID-19 ARDS, Crohn’s perianal fistula, osteoarthritis, and skin aging. It found a pooled serious adverse event (SAE) rate of 1.0% (95% CI, 0–2%) and an overall adverse event (AE) rate of 2.0% (95% CI, 1–5%), with no meaningful difference by EV source, engineering status, or administration route. [1] 

The dermatology literature tells a similar story in isolation. A systematic review of MSC-derived exosomes for cutaneous scars, aging, and hyperpigmentation found that reported adverse events were consistently mild, localized, and transient: erythema, edema, procedural discomfort, dryness. No infections, scarring, or systemic reactions turned up across the included studies. [2] A separate systematic review and meta-analysis of placental MSC-exosome use in wound healing and scar therapy reported a safety analysis spanning more than 738 patients over six years, with no allergic reactions, hypersensitivity responses, or adverse events attributable to exosome administration across topical, subdermal, subcutaneous, intravenous, and intra-articular routes.

That’s a favorable early signal. It isn’t a settled safety profile, and the reason for the distinction runs through the next three sections.

Where the Real Risk Signal Originates: The Contamination Pathway

The serious, hospitalization-level harms actually documented in the literature didn’t come from the biology of extracellular vesicles. They came from manufacturing and sterility failures in unapproved products distributed to clinics. That’s a supply-chain problem, not a property of the material class.

The clearest public record of this is the 2018 outbreak investigated by the CDC, involving umbilical cord blood–derived products processed by Genetech, Inc. and distributed by Liveyon, LLC. Clinics in Texas, Florida, and Arizona administered the product by injection or infusion for orthopedic and pain indications. [4,5] Twelve patients across those three states developed bloodstream infections, joint infections, and epidural abscesses.[4]  All twelve required hospitalization. None died. Laboratory testing of unopened vials from the implicated clinics, not just samples taken at the point of administration, found bacterial contamination already present in the sealed product (CDC MMWR, Perkins et al., 2018). The distributor issued a voluntary recall once the pattern was identified.

A follow-up national investigation broadened the case count substantially. A 2021 case series in JAMA Network Open documented 20 patients across 8 states who developed bacterial infections after receiving umbilical cord blood–derived products marketed as stem cell treatments (Hartnett et al., 2021). Whole-genome sequencing of patient and product isolates traced the contamination to a common source across products from multiple donors, pointing to a systemic manufacturing failure rather than mishandling at individual clinics.[5] 

The same pattern has since surfaced in products specifically marketed as containing exosomes. The CDC has separately reported that patients in Nebraska who received exosome-marketed products experienced serious adverse events, including bacterial infections, prompting a joint review with state health authorities. [6,7] 

Neither outbreak was caused by an intrinsic toxicity of extracellular vesicles. Both trace to what happened, or failed to happen, during manufacturing and chain-of-custody before the product ever reached a syringe. That’s why per-lot mycoplasma, sterility, identity, and potency testing, backed by a Certificate of Analysis a provider can actually trace to the batch in hand, functions as safety infrastructure rather than paperwork. For guidance on reading one line by line, see how to verify a Certificate of Analysis against the batch you’ve received. Storage and reconstitution practices between manufacture and injection matter for the same reason; a cold chain broken once can undo testing done correctly.

Immunogenicity: An Open Question, Not a Settled One

MSCs themselves are frequently described as hypoimmunogenic, given their limited expression of major histocompatibility complex (MHC) class II molecules and absence of key costimulatory molecules. That’s reasonably well established for the cells. It’s less settled for their exosomes.
A 2017 review in Frontiers in Immunology laid out the open question directly. Extracellular vesicles derived from dendritic cells are known to carry transferable MHC molecules capable of provoking immune responses in recipients, which raises the possibility that allogeneic MSC-derived EVs could carry similar immunogenic cargo. The review’s own conclusion was candid about where the evidence stood: whether MSC-derived exosomes reliably contain MHC molecules, whether those molecules transfer functionally to recipient cells, and whether any of this produces a clinically meaningful immune response were, at the time, each unresolved (Lohan et al., Frontiers in Immunology, 2017).
Nothing in the more recent clinical safety data contradicts that caution. It doesn’t resolve it either. Low AE/SAE rates in short-follow-up trials are consistent with genuinely low immunogenicity. They’re just as consistent with an immune signal that hasn’t had the time or the sample size to surface yet. The current evidence can’t distinguish between those two possibilities. Extrapolating a cell’s immune profile onto its secreted vesicles remains an assumption, and donor-to-donor and preparation-to-preparation variability makes that assumption harder to lean on, not easier.

Tumorigenicity and Oncogenic Cargo: Context Determines the Signal

The tumorigenicity question isn’t a single finding. It’s a body of work showing genuinely divergent results depending on cargo, tissue source, and disease context.
At the mechanistic level, MSC-derived exosomes have been shown in preclinical models to carry microRNAs capable of promoting tumor-relevant phenotypes in recipient cells. Exosomal transfer of miR-221 has been reported to enhance tumor aggressiveness in gastric cancer models. Glioma-associated MSC exosomes enriched for miR-1587 have been shown to increase proliferation and clonogenicity in glioma stem-like cells by downregulating a tumor-suppressor target in the recipient cells (reviewed in the Journal of Cancer, PMC6134817). A more recent in vivo study found that exosomes from rat bone-marrow MSCs promoted proliferation, invasion, and lung metastasis of colorectal cancer stem cells in a mouse xenograft model, while inhibiting those cells’ apoptosis (PMC12552078).
This body of work does not describe MSC-exosomes as tumorigenic in a general sense, though. The same literature documents tumor-suppressive effects from MSC-exosomes in other contexts. The prevailing framing in recent reviews is that the direction of effect depends heavily on cellular origin, molecular cargo, and the specific tumor microenvironment a given exosome population interacts with, rather than any fixed pro- or anti-tumor property of the vesicle class itself. None of this evidence comes from registered human trials measuring cancer incidence as an outcome. It’s preclinical and mechanistic, describing what exosomal cargo can do to specific recipient cell populations under specific experimental conditions. It is not a signal of elevated cancer risk observed in treated patients.
The honest read: oncogenic cargo transfer is a documented mechanism worth monitoring in long-term follow-up and worth accounting for in sourcing and characterization. It stops short of a demonstrated clinical outcome in the human safety data collected so far.

The Honest Limits of the Evidence Base

Every safety claim above has to be read against the shape of the evidence producing it, and that shape has three consistent limitations.
Trials are small. The meta-analysis cited above pooled 536 patients across 24 trials, an average of roughly 22 patients per trial. Individual studies in the dermatology and wound-healing literature frequently enroll fewer than 40 participants. A rare but serious adverse event, at a rate too low to appear in a 20- or 40-patient study, simply wouldn’t be visible in this literature yet. That’s a design limit, not a reassurance.
Follow-up is short. Most trials report safety outcomes over weeks to a few months. Even the longer-running data, the six-year, 738-patient wound-healing safety analysis cited earlier, represents pooled cross-sectional follow-up across accumulated cases rather than multi-year tracking of individual patients. A delayed immunogenic response, or a tumorigenic signal that takes years to manifest, sits largely outside the observation window this literature currently offers.
Products are heterogeneous. “Exosomes” in the published literature span different source tissues (bone marrow, adipose, umbilical cord, placenta), different isolation methods (ultracentrifugation, tangential flow filtration, precipitation-based kits), and inconsistent characterization against reporting standards. Two products both labeled “exosomes” in two different trials aren’t necessarily comparable products, which limits how confidently safety findings from one preparation generalize to another. A closer look at where exosome products currently stand under applicable regulatory frameworks helps explain why that inconsistency persists across the market.
The field’s own trial registry data backs up how early-stage this still is. A 2026 analysis of the global clinical trial landscape for human-derived extracellular vesicles found that of roughly 148 registered interventional trials, only 3 had reached Phase III (Wei et al., Clinical and Experimental Medicine, 2026). The evidence base this article draws on is, by the field’s own numbers, still predominantly early-phase.
None of that makes the existing safety signal wrong. It’s a genuinely good early signal. It just hasn’t had the sample size or the follow-up window to catch something rare or slow to develop, and it should be weighed the way any Phase I/II dataset gets weighed in other domains: informative, not conclusive.

Common Questions

Both documented contamination cases, the umbilical cord blood-derived products investigated by the CDC and the CDC-reported Nebraska exosome cases, point to manufacturing and sterility control as the failure point, not to something unique to exosomes as a material. The risk is best understood as a supply-chain and quality-control question that applies to any minimally regulated biologic product, with exosome-specific products being one documented instance of it.
No. Low AE/SAE rates in short-follow-up, small-sample trials are consistent with low immunogenicity, but they're also consistent with an immune signal too rare or too delayed to be captured by the current evidence base. The published literature itself treats this as an open question rather than a resolved one.
No. The literature shows a mixed picture, with tumor-promoting and tumor-suppressive effects both documented depending on exosome source, cargo, and the specific tumor microenvironment involved. None of the tumorigenicity data comes from registered trials measuring cancer outcomes in treated patients. It's mechanistic and preclinical, and belongs in the monitoring-priority column rather than the demonstrated-clinical-finding column
The CDC investigations traced contamination back to before the product ever reached the clinic. That points to per-lot mycoplasma, sterility, identity, and potency testing as the relevant checks, along with a Certificate of Analysis specific to the batch in hand rather than a generic product-level document.
The adverse-event rates and the contamination case histories are drawn from human clinical trials and CDC epidemiological investigations. The tumorigenicity mechanism data (oncogenic microRNA transfer, xenograft metastasis findings) is preclinical, drawn from cell culture and animal models, and hasn't been demonstrated as a clinical outcome in human trial participants.

References

    1. Efficacy and Safety of Mesenchymal Stromal Cell-Derived Extracellular Vesicles: A Systematic Review and Meta-Analysis. Regenerative Engineering and Translational Medicine. 2025. doi:10.1007/s40883-025-00532-8. Publisher pagelink.springer

    2. The Efficacy of MSC-Derived Exosome-Based Therapies in Treating Scars, Aging and Hyperpigmentation: A Systematic Review of Human Clinical Outcomes. 2025;8(4):268. PMCpmc.ncbi.nlm.nih

    3. Human Placental Mesenchymal Stem Cell-Derived Exosomes for Wound Healing and Scar Therapy: A Systematic Review and Meta-Analysis. 2026. PubMedpubmed.ncbi.nlm.nih

    4. Perkins KM, Cardozo L, Bower WA, et al. Invasive Bacterial Infections in Patients Who Received Stem Cell Products Derived from Umbilical Cord Blood—United States, 2018. Morbidity and Mortality Weekly Report. 2018;67:1395–1400. CDC MMWR

    5. Hartnett KP, Powell KM, Rankin D, et al. Investigation of Bacterial Infections Among Patients Treated With Umbilical Cord Blood-Derived Products Marketed as Stem Cell Therapies. JAMA Network Open. 2021;4(10):e2128615. doi:10.1001/jamanetworkopen.2021.28615. JAMA Network Openjamanetwork

    6. Centers for Disease Control and Prevention. Stem Cell and Exosome Products. Available at: CDC.cdc

    7. Lohan P, Treacy O, Griffin MD, Ritter T, Ryan AE. Anti-Donor Immune Responses Elicited by Allogeneic Mesenchymal Stem Cells and Their Extracellular Vesicles: Are We Still Learning? Frontiers in Immunology. 2017;8:1626. doi:10.3389/fimmu.2017.01626. Full textfrontiersin

    8. Figueroa JM, Phillips LM, Shahar T, et al. Exosomes From Glioma-Associated Mesenchymal Stem Cells Increase the Tumorigenicity of Glioma Stem-Like Cells via Transfer of miR-1587. Cancer Research. 2017;77(21):5808–5819. doi:10.1158/0008-5472.CAN-16-2986.

    9. Mesenchymal Stem Cell-Derived Exosomes in Cancer Progression and Metastasis. 2018. PMCpmc.ncbi.nlm.nih

    10. Wang M, et al. Dual Role of MSC-Derived Exosomes in Tumor Development. 2020. PMCpmc.ncbi.nlm.nih

REGULATORY DISCLAIMER · CELLS

These products are intended for laboratory research use only. They are not drugs, foods, cosmetics, or medical treatments and must not be used for any form of human or animal administration. All information provided is for educational and scientific reference only and the products should be handled exclusively by licensed, qualified professionals. Misbranding, misuse, or mislabeling of these products as therapeutic or consumable substances is strictly prohibited by law.

M.Sc. Patricia Cruz Martinez Medically Reviewed By — Written By — Updated on September 16, 2026

Read next

Shopping cart0
There are no products in the cart!
Continue shopping
Cart Gated

Cart is provider-gated. Log in to view.

Provider Login →