Rosemary oil or minoxidil: what the hair-loss studies show

What should you check before switching?

If you are already using minoxidil, first identify what you want to change. Is it helping but difficult to fit into the day? Is the scalp irritated? Is the problem cost, the feel of the preparation or uncertainty about whether it is working? Those questions can lead to different discussions with a pharmacist or clinician.

For example, someone who finds a solution unpleasant to apply may want to discuss formulation and suitability. Someone whose hair loss is progressing despite consistent use needs to review the diagnosis and response. Neither problem is answered by the statement that a natural oil matched minoxidil in one trial.

Keep the continuing-treatment issue in view. If stopping minoxidil allows its benefit to fade, a subsequent change during a rosemary trial could reflect both processes: the loss of one treatment's effect and whatever the new preparation is doing. Starting and stopping several products at once makes that harder to interpret. Discuss a planned change before stopping prescribed treatment.

The rosemary comparison lasted six months and found no significant improvement at three months. That gives little support to judging the ingredient from a few days of use. It does not mean you should continue through an adverse reaction for six months; troublesome scalp symptoms need attention rather than a longer experiment.

The NHS hair-loss guidance recommends seeing a GP if you are worried and finding out the cause before visiting a commercial hair clinic. Pattern hair loss is only one possibility. Widespread shedding following a trigger is a different question, covered in our guide to telogen effluvium and when blood tests may help.

Bring a short timeline to that appointment: when the change started, whether it is diffuse or patchy, any scalp symptoms, recent illness and the treatments already used. That helps make the next decision about your hair loss, rather than about whichever ingredient is currently most visible online.

Rosemary remains an option with limited, preparation-specific human evidence. Minoxidil has a stronger established treatment basis for appropriate forms of pattern hair loss. If you are choosing between them, the uncertainty in the rosemary evidence should be part of the decision rather than edited out of the product claim.

How well has minoxidil worked in controlled research?

The UK product information for Regaine for Men 5% solution identifies it as a licensed treatment for male pattern hair loss in men aged 18 to 65. It is available without prescription. That particular licence and formulation should not be assumed to cover a different person, product or cause of thinning.

The product information reports a 48-week trial in which non-vellus hair counts increased by an average of 18.6 hairs per square centimetre with 5% minoxidil, 12.7 with 2% minoxidil and 3.9 with vehicle. Non-vellus hairs are the larger hairs distinguished from fine vellus hairs in the count. The 5% result was significantly greater than the 2% result, and both exceeded vehicle.

The label says visible growth may take two months or more and continued use is needed to maintain it; regrown hair may disappear three to four months after stopping. This is a continuing treatment decision, rather than a short course whose result can be assumed to last indefinitely.

Those figures establish a useful benchmark within that trial. They cannot be compared directly with a rosemary count from another study if the populations, duration or area being counted differ. The fair question is what happens when treatments are compared under the same conditions, with enough precision to judge the difference.

What did the rosemary-versus-minoxidil trial find?

The original 2015 paper, on pages 15 to 21 of the journal issue, enrolled 100 men aged 18 to 49 with androgenetic, or pattern, hair loss. It compared a standardised rosemary lotion with 2% minoxidil for six months. The rosemary preparation contained at least 3.7 milligrams of 1,8-cineole per millilitre. Participants used one millilitre twice daily.

Two dermatologists assessing the results were unaware of treatment allocation. The report describes matching the containers, but does not explain how rosemary's smell was masked. It should not be presented as securely blinded for everyone involved.

Hair counts did not increase significantly at three months. At six months, mean counts rose from 122.8 to 129.6 with rosemary and from 138.4 to 140.7 with minoxidil. The report does not specify an area denominator for these counts, so they should not be labelled hairs per square centimetre. The between-group difference was not statistically significant.

These were modest changes in the measured field, not a measurement of whole-scalp restoration. The trial was in men, used a particular lotion and had no vehicle-only group. It leaves substantial uncertainty about the extra benefit attributable to rosemary.

Why does no significant difference not mean equally effective?

A study can fail to detect a difference for several reasons. The treatments may have similar effects, or the study may be too small or variable to distinguish them. To support a claim that one is acceptably close to the other, researchers need a design that defines how much difference would count as too much and estimates that difference precisely enough.

A simple invented example makes the problem clearer. Suppose a treatment group's average hair count rises from 20 to 30, while another group's rises from 20 to 29. Both groups improved. That alone does not tell us whether the one-hair difference is meaningful, whether chance could explain it or whether either treatment outperformed doing nothing. We need the spread of results and the appropriate comparisons.

The rosemary trial's result therefore allows a narrower statement: it observed gains after six months and did not detect a difference from 2% minoxidil. Turning that into a promise of equal effectiveness removes the uncertainty that matters when choosing between them.

The missing 5% comparison also matters. We have evidence that 5% and 2% minoxidil performed differently in another controlled trial. We cannot use the weaker comparator as a bridge to a conclusion about 5% without testing that comparison. This is not proof that rosemary has no effect; it is a reason to keep the claimed effect within what was studied.

Does the newer rosemary-mixture trial strengthen the case?

A 2025 trial randomised 90 people with thinning or shedding hair to two rosemary-containing blends or coconut oil for 90 days. Its analysis included 82 people who completed the study. It was described as double-blind, although the methods do not clearly explain how differences in smell were concealed. The participants were not all established cases of androgenetic hair loss.

One blend's reported hair-growth rate rose from about 0.22 to 0.34 millimetres per day. The reported 57.73% increase was relative to that group's own starting point. The coconut-oil group also improved, with a reported 21.11% increase. The 57.73% figure therefore does not mean the blend was that much better than the control.

The authors reported that the blends outperformed coconut oil and gave fold comparisons. However, the statistical methods describe paired tests of change within groups, without a clearly reported independent-arm difference and confidence interval. That limits confidence in the advertised superiority figures. The products contained multiple ingredients, so any benefit cannot be assigned to rosemary alone. Soulflower funded the study and an author was an employee.

Counting only people who completed a study also leaves a question about those who did not. If someone withdraws because a product is unpleasant or seems ineffective, their experience matters to the person considering buying it. A completer analysis can miss that part of the decision.

Is the product you are buying the one that was studied?

A study of a standardised lotion does not test neat essential oil, a different concentration in a carrier oil or a rinse-off shampoo. The amount reaching the scalp, how long it remains there and the other ingredients can all differ. The sensible purchase question is therefore specific: has this finished preparation been studied, or is its marketing borrowing a result from another one?

Look for a named human study and check the treatment details against the label. If the evidence concerns a mixture, the seller should identify it as a mixture. If the result is a change from baseline, it should not become a percentage advantage over a comparison treatment. A supplier that cannot make those distinctions has not answered the question of effectiveness.

It is also worth distinguishing a conditioning effect from growth. Hair that looks smoother or is styled differently may appear fuller in a photograph. That does not tell you whether the number of growing hairs increased. In before-and-after pictures, different lighting, hair length and the position of a parting can make this particularly hard to judge.

This is why the actual trial measurements matter more than a dramatic photograph. Hair count, hair-growth rate and a person's rating of appearance answer related questions, but one cannot silently stand in for another. Choose evidence that measures the change you are hoping to see.