Exosomes for Hair Loss: Reading the Study Designs

In short

The design of a study sets a ceiling on the claim it can support. Most published work on exosome preparations for hair loss sits at the lower rungs of that ladder: laboratory work on cells, then animal work, then small human case series. None of those designs can establish that a treatment works in people.

The design that would settle the question in androgenetic alopecia is a randomised controlled trial in which everyone receives the same microneedling and only the applied preparation differs, with assessors who do not know which side is which, hair counts rather than impressions, and follow up long enough to cross the hair cycle. Ask which of those features the cited study had. The answer is more informative than the result.

Clinics rarely misquote a study. What happens far more often is that a study is quoted accurately and its design is left out, and the design is the part that decides how much weight the result can carry. This page is a reading skill rather than a verdict. Work through it once and you will be able to place almost any citation you are shown, on this subject or any other.

Why the design sets the ceiling

Every study answers a narrower question than the one a reader wants answered. A laboratory experiment answers whether something happens to cells in a dish. An animal experiment answers whether something happens in that animal, under conditions chosen by the experimenter. A case series answers what happened to a group of people who all received the treatment. Only a controlled comparison answers the question a patient is actually asking, which is whether they will be better off having the treatment than not having it.

That is why the ladder matters. Moving up a rung does not make a result more impressive. It removes a specific rival explanation. Randomisation removes the possibility that the treated group was different to begin with. A control group removes the possibility that time, seasonality or a change in grooming produced the improvement. Blinding removes the possibility that the assessor saw what they expected to see. Objective counting removes the possibility that a photograph was taken in kinder light. Nothing about a study is impressive in the abstract. Each feature is there to close one door.

The ladder, from weakest to strongest

The table below is deliberately blunt. It is not a scoring system, and a study low on the ladder is not worthless. It is a statement of what each design can and cannot rule out.

Study designs and the claims each can support
Design	What it can show	What it cannot rule out
Cell culture work	That a preparation changes the behaviour of cells in a dish	Everything about a living scalp: delivery, dose, survival, immune response, duration
Animal study	That an effect occurs in that species under controlled conditions	That human hair biology, human dosing and human follow up behave the same way
Single case report	That one outcome occurred once	Chance, selection, natural fluctuation and the effect of anything else done at the same time
Uncontrolled case series	What happened to a group who all had the treatment	What would have happened to the same people without it
Controlled trial, no blinding	A comparison against something	Assessor expectation and participant expectation shaping the outcome measure
Split scalp randomised trial	A within person comparison with each person as their own control	Carry over between sides, and effects that are systemic rather than local
Blinded randomised controlled trial	Whether the treatment outperformed the comparator in that population	Whether the result transfers to a different preparation, dose or protocol

What a case series can and cannot tell you

The uncontrolled case series is the workhorse of the aesthetic literature, and it is the design most often described in a consultation as a study showing that the treatment works. It cannot show that. A case series records what happened to people who had the treatment. It has no comparison group, so there is no way to know what would have happened without it.

Three things routinely produce apparent improvement in an uncontrolled series. The first is regression to the mean: people tend to seek treatment when a problem feels worst, and problems that fluctuate tend to be less bad later regardless of what is done. The second is co-intervention: people who start a treatment often start other things at the same time, including licensed topical or oral treatment, better sleep, iron correction or a change in styling. The third is measurement drift, which is what happens when the follow up photograph is taken with different lighting, a different parting or a wet rather than dry scalp.

None of this makes a case series dishonourable. Early series are how a question gets asked. The error is treating the answer to the question as though it had already been given.

The attribution problem, which is specific to this treatment

Exosome preparations for the scalp are almost always applied after microneedling, because particles of that size do not cross intact skin in useful quantity. Microneedling has an effect of its own. So a study in which one group receives microneedling plus a preparation and the other group receives nothing at all cannot separate the two, no matter how well it is otherwise conducted.

The design that closes that door is needling plus preparation against needling plus an inert vehicle, ideally on two halves of the same scalp. It is not an exotic request. It is standard practice in dermatology trials, and its absence from a body of literature is a finding in itself. Our evidence page sets out the recurring weaknesses in the human literature on this subject, and the skin page explains why the same attribution problem shows up in facial treatment.

Outcome measures, from weakest to strongest

What was measured matters as much as who was compared. Satisfaction scores are the weakest outcome available, because they measure how a person felt about a process they paid for and invested hope in. Investigator global assessment is stronger but still a judgement. Standardised photography with fixed distance, lighting and head position is better again. Hair count and hair calibre measured over a tattooed or otherwise fixed target area, read by someone blinded to the treatment, is the outcome that carries real weight, because it counts objects rather than impressions.

When a claim is quoted to you, ask which of those was used. A study that reports a large change in satisfaction and a small change in hair count has told you something quite specific, and it is not the thing the headline says.

Follow up, and the hair cycle

Hair grows in cycles, and any treatment that acts by pushing follicles from rest into growth will appear to do something within weeks that has little to do with sustained benefit. Follow up measured in weeks cannot distinguish a shift in cycle timing from a change in the underlying condition. This is why dermatology has long treated three to six months as the earliest point at which a hair treatment can be assessed at all, and why a twelve week endpoint should be read as a preliminary observation rather than a result. Our hair loss page sets the licensed options against the unlicensed ones on exactly this timescale.

Why results do not transfer between preparations

There is one further problem that is specific to this field. Two products described by the same word can differ in source cell, species, culture conditions, isolation method, characterisation standard, concentration, storage and handling. The reporting frameworks published by the International Society for Extracellular Vesicles exist precisely because a preparation cannot be interpreted without that information. Where a study does not describe its material to that standard, its result cannot be carried across to a different product in a clinic, even if both are called the same thing.

A five minute reading rule

When a clinic shows you evidence, work through these in order and stop at the first no.

Is it in humans, with a scalp, rather than in cells or in mice?
Was there a comparison group, and did that group also receive the microneedling?
Were people allocated at random, or by choice?
Did the person assessing the outcome know which treatment was given?
Was the outcome counted, or rated, or asked about?
Did follow up run to at least three to six months?
Is the preparation described well enough that you could tell whether the clinic uses the same one?

Very few citations in this area survive to the end of that list. That is the point of the exercise. It is not a reason to refuse the treatment. It is a reason to price it as something being tried rather than something established, and to expect the person selling it to describe it that way.

What this page does not cover

This is a guide to reading designs, not a systematic review, and it does not grade individual papers or name products. It does not cover the biology of the vesicle itself, the manufacture of these preparations or the laboratory methods used to characterise them, which belong to a different discipline and are properly answered by the extracellular vesicle literature rather than by a patient guide. It does not cover hair loss that is not androgenetic, which needs a diagnosis before any treatment question is meaningful. And it is not medical advice.

Sources

Common questions

Does a study in mice mean nothing?

It means something narrow. Animal work shows that an effect can occur in a living system and helps decide whether a human study is worth running. It cannot establish a human dose, a human response or a human safety profile, and hair biology differs between species.

Is a case series useless?

No. A case series is how a question gets raised and how obvious harms get spotted early. It simply cannot answer whether the treatment caused the change, because there is nobody to compare the treated group with.

What is a split scalp study?

A design in which one side of a person's scalp receives one treatment and the other side receives the comparator, so each participant acts as their own control. It removes differences between people, which is the largest source of noise in hair research.

Why does blinding matter for something visible?

Because visible things are rated by people, and people rate what they expect. If the assessor knows which side was treated, the rating carries their expectation as well as the result. Blinding is the only practical way to separate the two.

How long should follow up be for a hair treatment?

Three to six months at minimum, because the hair cycle produces changes on that timescale by itself. Anything reported at four or eight weeks is an early observation, not an outcome.

Can I ask a clinic for the studies?

Yes, and the request is reasonable. Ask for the citation rather than a summary, then apply the seven checks above. A clinic that answers with the design as well as the finding is telling you something useful about how it reads its own field.

No commercial links

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 Northbank Media under our editorial policy.

If you want to speak to someone

We are not a clinic and we give no medical advice. If you would like to, we can pass your details to a registered UK practitioner who will contact you directly. Nothing is booked on your behalf and there is no obligation. Read the regulation page first.

Make an enquiry

Exosomes for hair loss: what each study design can actually show