Melatonin: only two of four comparisons were significant, and the two were different sites in different groups

A 2004 trial in Br J Dermatol from Friedrich Schiller University, Jena, Germany. Double-blind, randomised, placebo-controlled.

The paper's background says: melatonin has been reported to have a beneficial effect on hair growth in animals, but the effect of melatonin on hair growth in humans has not been investigated so far. In other words, the first trial to ask this in people.

40 women with diffuse alopecia or androgenetic alopecia took part.

• a 0.1% melatonin solution or a placebo solution applied to the scalp once daily for six months
• trichograms to assess anagen and telogen hair rate
• blood samples throughout the study to monitor effects on physiological melatonin levels

The results. Here are the four comparisons as reported.

• occipital hair in women with androgenetic hair loss — a significantly increased anagen rate compared with placebo (n = 12, P = 0.012)
• frontal hair in the diffuse alopecia group — a significant increase (n = 28, P = 0.046)
• the occipital samples of the diffuse group and the frontal counts of the androgenetic group also showed an increase of anagen hair, but the differences were not significant
• plasma melatonin rose under treatment but did not exceed the physiological night peak

Conclusion: to the authors' knowledge, this pilot study is the first to show that topically applied melatonin might influence hair growth in humans in vivo. The mode of action is not known, but the effect might result from an induction of anagen phase.

Now how to read it.

First, that there are four comparisons is the heart of this entry. Two groups by two sites. Run several comparisons and some can reach significance by chance. And the abstract does not say that this possibility was corrected for.

Second, look at the shape of the two significant results. In androgenetic it was the occipital; in diffuse, the frontal. They cross. Had both sites risen in one group it would be far easier to read. As it stands, it is hard to say what to expect at which site in whom.

Third, the numbers. It is 40 people, but one significant result came from 12. With P = 0.012. A small p from a small group often disappears on a repeat.

Fourth, the other p is 0.046. Close to 0.05. On the laser-assisted drug delivery page this site carried across a paper's own word for p = 0.049: borderline. Read this with the same eye.

Fifth, the design does have good points.

• it is double-blind and placebo-controlled
• it ran six months. Hair cycles are long, so short trials are hard to judge on; this one is not short
• the endpoint is a trichogram. Plucked hairs counted under a microscope for the anagen-to-telogen ratio. Not self-assessment.
• blood was checked, and it reported not exceeding the physiological night peak

Sixth, what a trichogram measures. The proportion of anagen hairs. Not how many more hairs there are. A rising anagen rate is a signal in that direction, but it is not the same kind of value as the 'this many more hairs inside a defined area' reported on the minoxidil and finasteride pages. The same badge can rest on different measurements.

Seventh, twenty years have passed. A 2024 review collected clinical studies of topical melatonin, but we did not find a trial that asked this question again with the same design and confirmed it. The first trial in humans still stands almost alone.

Eighth, only women took part. Results for men cannot be taken from this trial.

So it sits at 'early evidence'. The design is proper, but there is one trial, four comparisons, a small group behind one significant result, and thin replication.