4-Butylresorcinol: twenty people and forty-five, and both beat their own vehicle

This is a 2010 Korean trial. 20 people with melasma took part, in a randomised, double-blind, vehicle-controlled, split-face design.

One side of the face received 0.1% 4-butylresorcinol cream and the other the same cream with only this ingredient removed, twice daily for 8 weeks. Melanin index was measured with a mexameter, alongside photography.

Result: melanin index on the treated side was significantly lower than the control side at both week 4 (P = .006) and week 8 (P < .0005). All 20 completed, and adverse reactions were all mild and transient.

Two things to see in this.

First, there were 20 people. A single small trial cannot rule out chance.

Second, the split-face design carries its strength and its weakness together. Comparing the two sides of one person cancels out differences between individuals. But melasma is often not symmetrical, and if one side was darker to begin with it also has more room to fall.

This is a 2016 Korean trial. Here the ingredient was carried not in a cream but in a dissolving microneedle patch.

45 people took part in the efficacy assessment, against a control of the same patch with only the 4-butylresorcinol removed. Double-blind and placebo-controlled, over 8 weeks, measuring melanin index and ITA (a way of expressing skin colour as an angle).

The result was reported as more than twice as effective as the control patch.

Safety was assessed separately: a primary irritation test in 31 subjects and a cumulative irritation and sensitisation test in 50 women. It was judged safe.

What 'twice' is twice of is hard to pin down from the abstract alone. And a cream and a patch deliver the ingredient differently, so this result does not transfer directly to something you apply.

Both trials ran 8 weeks. What happens over longer use, and after stopping, has not been established.

In Korea, nine ingredients are listed in the functional cosmetics review rule (Annex 4) as accepted for a 'helps whiten skin' claim without submitting efficacy data of your own: broussonetia extract, arbutin, ethyl ascorbyl ether, licorice extract (oil-soluble), ascorbyl glucoside, magnesium ascorbyl phosphate, niacinamide, alpha-bisabolol and ascorbyl tetraisopalmitate.

4-Butylresorcinol is not on that list — the same position as thiamidol and cysteamine. It does not mean it does not work; it means a product making the claim has to go through individual review.