Exosome Cosmetics: How to Read a Certificate of Analysis
Updated 5 October 2026
Exosome-complex cosmetics are marketed with big words and few numbers. This guide shows which numbers a credible batch certificate should contain, what each one means, and what it does not mean.
What are exosomes, and what does "synthetic" mean here?
Natural exosomes are tiny membrane-bound vesicles released by cells. Scientific bodies prefer the broader term extracellular vesicles unless the biological origin of a vesicle has been demonstrated. Cosmetic ingredient makers also produce lab-made, exosome-like vesicles, often labelled "synthetic exosomes". They are built, not harvested, and are blended into serums, creams, gels and hair-care products as one ingredient among several.
Two practical consequences follow. First, the name on the label should say what the ingredient is: lab-made vesicles are not donor-derived. Second, a trade name is not an ingredient declaration; in many markets the full INCI breakdown must be given to the regulator.
The regulatory frame in three sentences
In the United States the FDA states that there are currently no FDA-approved exosome products and treats exosomes used against disease as drugs or biologics. In India the drug regulator has said injectable preparations are not cosmetics and cosmetics cannot be used therapeutically. In Thailand, cosmetic advertising may not make medical claims, imply internal use or injection, or suggest penetration beyond the outer skin layer. That is why a cosmetic exosome page must be about evidence of quality, not about treatment.
Seven things a credible batch certificate should show
- Batch identity and dates. Batch number, manufacturing date, best-before period and storage condition. An expired batch is history, not proof.
- Particle concentration. Reported per millilitre with the method named (for example, tunable resistive pulse sensing). Compare the result with the stated specification, not with another company's marketing.
- Particle size. A mean size and a specification window show whether the vesicles are consistent from batch to batch.
- Heavy metals by ICP-MS. Lead, mercury, cadmium, arsenic and antimony as numbers with units. "Complies" without numbers is weaker evidence.
- Microbial limits. Total count, yeast and mould, and absence of specified organisms per millilitre.
- Preservative challenge. Shows the preservative system protects the product against contamination during use.
- In-vitro cytotoxicity (ISO 10993-5). A laboratory cell test. Useful, but never a substitute for human skin testing.
Example: what real batch numbers look like
The tables below come from production-batch certificates of analysis issued by an ISO 9001:2015 certified manufacturer. They are shown only to illustrate how to read a COA. Results apply to the batch tested.
| Parameter | Method on COA | Specification | Result |
|---|---|---|---|
| Appearance | In-house | Translucent pink solution with dispersed vesicles | Conforms |
| pH at 25 °C | In-house | 4–9 | 7.2 ± 0.2 |
| Specific gravity at 25 °C | Pycnometer | Not specified | 1.00 g/mL |
| Particle concentration | In-house, TRPS | ≥ 2.5 B/mL | 2.6 B/mL |
| Viscosity | Rotational viscometer | Not specified | 1881.05 mPa·s |
| Lead / Mercury / Cadmium / Arsenic / Antimony | ICP-MS | <10 / <3 / <3 / <3 / <5 ppm | All <0.01 ppm |
| Total microbial count | USP <61> | <100 cfu/mL | <10 cfu/mL |
| Yeast and mould | USP <61> | <10 cfu/mL | <10 cfu/mL |
| E. coli, P. aeruginosa, S. aureus, C. albicans, Bacillus spp. | Per COA | Absent/mL | Absent/mL |
| Parameter | Acceptance criterion | Result |
|---|---|---|
| pH range | 4–9 | 7.2 ± 0.2 |
| Heavy-metal traces | Nil (limit test) | Complies |
| Preservative challenge | Complies | No contamination |
| Microbial enumeration | Complies | Negative |
| In-vitro cytotoxicity (ISO 10993-5) | Complies | Negative |
| Estradiol screen | <0.01 pg/mL | <0.01 pg/mL |
| Parameter | Specification | Result |
|---|---|---|
| Particle concentration (TRPS) | ≥ 2.0 B/mL | 2.4 B/mL |
| Mean particle size | 100–200 nm | 152 nm |
| pH at 25 °C | 4–9 | 7.2 ± 0.2 |
How to read these tables
The result column sits inside the specification column in every row; that is what "conforms" means. A result of "<0.01 ppm" is a reporting limit, not proof of absence. The estradiol row screens one hormone only, so it must not be stretched into "hormone-free". The cytotoxicity row is a cell-culture test, so it does not support "dermatologically tested" or "hypoallergenic".
Ingredient literacy: value comes from each part
An exosome-complex product is never one ingredient. The vesicle concentrate is blended with humectants, emollients, conditioning agents and a preservative system, and each plays a different cosmetic role. A good product page therefore lists the full INCI, explains the role of each key ingredient in plain words, and links to the supporting reference. Our ingredient library follows that format.
A buyer's evidence checklist
Use this list with any supplier. The right-hand column shows where we stand.
| What to ask for | Why it matters | Our status |
|---|---|---|
| Particle concentration with method and specification | Turns "rich in exosomes" into a number you can check | Reported on every batch certificate |
| Particle size window and result | Shows batch-to-batch consistency | Reported where measured |
| Numeric heavy-metal results (ICP-MS) | "Complies" alone cannot be audited | Reported where measured |
| Microbial counts and specified-organism absence | Basic contamination control | Reported on every batch certificate |
| Preservative challenge result | Protects the product during use | Reported on every batch certificate |
| In-vitro cytotoxicity (ISO 10993-5) | Cell-level compatibility signal | Reported on every batch certificate |
| Batch number, manufacturing date, best-before, storage | Lets you reject expired or mismatched documents | On every COA |
| Quality system certificate | Process control behind the numbers | ISO 9001:2015 certified manufacturer |
| Safety data sheet | Handling, storage, transport and regulatory status | Available on request |
| Stability data | Supports the stated shelf life | Available to qualified partners under NDA |
| Human efficacy or irritation studies | Needed before any performance or "tested on skin" claim | Not claimed — no human-study results are claimed |
The last row is deliberate. Saying plainly what is not claimed is part of the evidence.
Why machine-readable evidence matters now
Shopping is moving towards software agents that compare products on a person's behalf. Open standards such as the Universal Commerce Protocol (UCP) and the Agent Payments Protocol (AP2) let a business publish what it sells and how it can be bought in a form an agent can read and verify. Evidence published as tables, structured data and a UCP profile can be read by an agent; evidence buried in a screenshot cannot. See how AI shopping agents read product pages.
Frequently asked questions
What are synthetic exosomes in cosmetics?
Lab-made, exosome-like vesicles used as an ingredient in a topical product. They are not harvested from a donor, so "exosome-like" or "synthetic vesicles" is the accurate description.
Do cosmetic exosome products treat skin or hair conditions?
No. A cosmetic changes appearance and care of the outer body. Treatment claims belong to medicines and biologics, which have separate approval routes.
What does particle concentration in B/mL mean?
Billion particles per millilitre, counted by a particle-measurement method. It describes quantity, not effect.
What does ISO 10993-5 test?
In-vitro cytotoxicity: whether a material harms cultured cells. It is not a human irritation or allergy study.
Why does a COA show both a specification and a result?
The specification is the pass range; the result is the measurement. You need both to judge conformity.
Why do some results say "below 0.01 ppm" instead of zero?
That is the instrument's reporting limit. Not detected above the limit is not the same as absent.
How should the products be stored?
The batch documents state: cool and dry place, away from sunlight, below ambient temperature, do not freeze. Follow the label on your pack.
Can AI shopping agents read this evidence?
Yes, when it is published as structured data and a UCP profile rather than only as images or PDFs.