Exosome treatment for hair loss, read by study design
How to tell what kind of study is behind a hair loss claim, what each design can and cannot establish, and why the strongest available design for a treatment is more useful than any adjective attached to it.
Published by Valextino Ltd. Published 2026-09-01. Last reviewed 2026-09-01. Information only. This site is not a clinic and gives no medical advice.
The useful question about evidence for exosome hair treatment is not whether studies exist. It is what kind of study exists, because designs differ in what they are capable of showing. A case series can show that something was tried and describe what followed. Only a randomised, controlled and ideally blinded study can separate the treatment from the natural course of the condition, from the other things a person changed at the same time, and from the expectations of everybody involved.
So the single sentence worth extracting from any claim is: what is the strongest study design that exists for this indication, and has a regulator stated a position. Everything else is commentary.
Why design comes before results
Results are quoted at you. Designs are not. That asymmetry is the whole problem, because a percentage improvement from an uncontrolled series and the same percentage from a randomised controlled trial are not the same number in any meaningful sense, and nothing about the way they are presented tells you which one you are looking at.
Hair loss is a particularly unforgiving subject for weak designs. Androgenetic alopecia progresses slowly and unevenly, telogen effluvium recovers on its own, seasonal shedding varies, and photographs are extraordinarily sensitive to lighting, parting, hair length, wetness and camera angle. A treatment that does nothing at all will still produce a fraction of people who look better in the after photograph. No deception is required for that to happen. It only needs a design that cannot tell the difference.
The ladder of designs, and what each one can establish
These are the designs you will actually encounter, weakest first. The distinction that matters most is between designs where somebody chose which subjects were treated and designs where that was randomised, because that is the point at which the comparison starts meaning something.
| Design | What it can establish | What it cannot |
|---|---|---|
| Single case report | That this happened once, in one person, described in detail | Anything about how often, in whom, or why |
| Case series | That a treatment was given to a group, at a stated dose, with observed outcomes | That the treatment caused the outcome, because there is nothing to compare with |
| Retrospective review of records | Patterns across a larger group, cheaply, and hypotheses worth testing | Comparability, because who was treated was decided by clinicians, not by chance |
| Prospective cohort with a comparison group | Change over time in treated and untreated groups followed the same way | That the two groups were alike at the start, which is where the bias lives |
| Randomised controlled trial | That the difference between groups is attributable to the treatment | Much, if it is small, short, or the outcome was assessed by someone unblinded |
| Blinded randomised controlled trial | The above, with expectation removed from both subject and assessor | Generalisation beyond the people and protocol actually studied |
| Within person or split scalp randomised design | A treated and untreated area in the same person, which removes most between person variation | Whether an effect could cross to the untreated side, which good protocols address |
| Systematic review of the above | What the whole literature shows once selection of studies is itself systematic | More than the studies it contains, and it will say so if it is any good |
Two rungs deserve attention because they are the ones most often skipped over. The case series is the most common form of published evidence in aesthetic medicine, and it is not worthless: it establishes that something has been tried, in whom, at what dose, and what was observed, which is exactly what you need before designing a real trial. What it cannot do is attribute the change to the treatment.
The split scalp or within person design is unusually strong for hair and is under used. Treating one side and leaving the other as a control removes almost all of the variation between people, because the same person, the same hormones, the same medication and the same photographer are on both sides of the comparison. When a within person design exists for an indication, it is usually the most informative thing available.
What was actually measured
The design tells you whether a comparison is valid. The outcome measure tells you whether the thing compared was worth comparing. In hair studies the measures range from countable to entirely impressionistic.
| Measure | How repeatable it is |
|---|---|
| Hair count or density in a marked target area | High. Countable by another person from the same image |
| Standardised photography, fixed distance, lighting and parting | Moderate to high, and only if the standardisation is described |
| Trichoscopy measurements of shaft diameter and density | Moderate to high, dependent on the same site being found again |
| Blinded assessor rating of before and after images | Moderate. Depends entirely on the blinding being real |
| Unblinded investigator global assessment | Low. The person rating the photograph knows which arm it came from |
| Patient satisfaction score | Low as evidence of growth, though it is a legitimate measure of something else |
| Untitled before and after photographs in marketing | None. Not a measure and not evidence |
The rule of thumb is that a measure someone else could repeat is worth more than a measure that requires the original observer. Hair counts in a tattooed target area can be repeated. A satisfaction score cannot be repeated by anybody, and in an unblinded study it partly measures how the appointment felt.
Four questions that survive any sales conversation
You do not need to read the literature to use it. You need four answers, and the reaction to being asked is itself informative.
| Ask | A complete answer |
|---|---|
| What is the strongest study design that exists for this treatment for my type of hair loss | Names a design and where it was published, not a description of results |
| Was it randomised, and was the person assessing the outcome blinded | Yes or no to each. Two noes puts the claim near the bottom of the ladder |
| What was measured, and by whom | A named measure. Hair count in a marked area is a different order of answer from satisfaction |
| Has a UK regulator stated a position on this product or this use | A position, or an acknowledgement that there is not one. Silence is not endorsement |
A complete answer to the first question sounds like a design and a place of publication. An incomplete answer sounds like a description of results with no design attached, an appeal to how many people have had it, or a reference to research in general. There is no version of a strong evidence base that cannot be described in one sentence by the person selling the treatment.
What a thin evidence base does and does not mean
Thin evidence does not mean a treatment does not work. It means nobody has established whether it does, which is a different statement and leaves the risk with the person paying. That is a decision a reasonable adult can make with their eyes open, and this page exists to make the eyes open part possible rather than to make the decision.
What thin evidence does mean is that certain sentences are not available. Nobody can tell you how likely you are to respond, or how long a response lasts, or how it compares with an option that has been tested properly, because those numbers come from designs that have not been run. A clinic that supplies them anyway has not read something you have not read. It has made them up or repeated somebody who did.
It also means the comparison worth making is against the options that do have licensed status and a larger evidence base for the same indication, which is a conversation for a prescriber. Our page on alternatives sets out what those comparisons involve, and the evidence page covers what has and has not been established across indications.
What this page does not cover
This is a page about how to read evidence, not a review of any particular body of evidence. It does not summarise, cite, count or grade individual studies, it names no trial, no author, no journal and no product, and it quotes no result. It does not tell you whether exosome treatment works for your hair loss, which is not a question a general reference can answer. It does not cover the biology of extracellular vesicles, their sourcing, characterisation or manufacture, which belong to Exosomal Therapy. It is not medical advice and it is not a substitute for assessment by a clinician who can examine your scalp.
Common questions
Does exosome therapy work for hair loss
That question cannot be answered from a general reference, and the more useful question is what kind of evidence exists. Ask for the strongest study design that has been run for your type of hair loss, whether it was randomised, whether the assessor was blinded, and what was measured. Those four answers tell you more than any summary of results.
Is a case series evidence
It is evidence of what happened, not evidence that the treatment caused it. Case series are a legitimate and necessary early step, and they are the commonest form of published work in aesthetic medicine. They cannot separate a treatment from the natural course of a condition.
Why does blinding matter so much for hair
Because the outcome is usually judged from photographs, and photographs are sensitive to lighting, parting, length and angle. An assessor who knows which image is the after image will see more improvement than one who does not, without anyone intending it.
What is a split scalp study
A within person design where one side is treated and the other is not, so the comparison happens inside a single person. It removes almost all of the variation between individuals and is unusually informative for hair, which is why it is worth asking whether one exists.
If there are no good studies, does that mean it does not work
No. It means nobody has established whether it does, and the risk of that uncertainty sits with the person paying. What it rules out is any specific claim about how likely, how much or how long, because those numbers can only come from designs that have not been run.
How do I check a claim myself
Ask for the study by name and look at the design section rather than the conclusion. If a study cannot be named, there is nothing to check, and that is the finding.
This page contains no commercial links of any kind. No clinic, practitioner, product or supplier is named, recommended or linked to, and nobody has paid for, influenced or previewed anything on it. External links go only to UK regulators and professional bodies and carry a nofollow attribute. Published by Valextino Ltd under our editorial policy.