Tranexamic acid: the oral route is established, the topical one is not

A 2024 network meta-analysis compared the routes of tranexamic acid side by side, pooling PubMed, EMBASE, Cochrane and Web of Science, with melasma severity measured as MASI.

Against placebo, topical tranexamic acid showed a statistically significant difference from week 8 onwards. Intradermal and microneedling routes started at the same point; only the oral route separated as early as week 4.

One sentence in the conclusion matters: in the long run, the effect of tranexamic acid alone has nothing to do with the route of administration. The oral route is faster to start, and over time they converge.

A different meta-analysis from the same year pooled 28 randomised controlled trials. Its question is different: not placebo, but whether the drug adds anything on top of existing treatment.

The numbers as published:

• Oral: 8 weeks SMD 1.61 (95% CI 0.44–2.79, P = .007); 12 weeks SMD 2.39 (95% CI 1.42–3.35, P < .00001). Significant.
• Topical: 8 weeks SMD -0.05 (95% CI -1.08–0.97, P = .92); 12 weeks SMD 0.66 (95% CI -0.10–1.42, P = .09). Not significant at either point.
• Intradermal: 8 weeks SMD 1.21 (P = .14); 12 weeks SMD -0.55 (P = .54). Also not significant.

The authors conclude plainly that the oral formulation can be used alongside adjuvant treatment, and that data are still required for the topical and intradermal routes. They also flag substantial heterogeneity among the included studies.

This looks like it contradicts the item above, but it does not — the comparator differs. Better than applying nothing; not demonstrated on top of what you already use. That is the most accurate sentence available about this ingredient right now.

A 2024 double-blind, split-face randomised trial in Indonesia. Twenty people with mixed-type melasma applied 3% tranexamic acid cream to one side and 4% hydroquinone cream to the other for eight weeks.

mMASI fell significantly from baseline in both groups at weeks 4 and 8, and patient global assessment showed no significant difference between the two.

Reading that as 'as good as hydroquinone' goes too far. Twenty people cannot demonstrate the absence of a difference — with a small sample, a real difference can easily fail to reach significance. Concluding equivalence requires a trial sized to show it, and twenty is not that size.

Also, hydroquinone is not permitted in cosmetics in Korea. The comparator in this trial is not a product you can buy there.

Korea's functional-cosmetics review rule (Annex 4) lists nine whitening actives that may be reported as functional without submitting data: broussonetia extract, arbutin, ethyl ascorbyl ether, oil-soluble liquorice extract, ascorbyl glucoside, magnesium ascorbyl phosphate, niacinamide, alpha-bisabolol and ascorbyl tetraisopalmitate.

Tranexamic acid is not among them.

To be precise about what that means: it is not a ban. It means that to carry a whitening functional claim with this ingredient, a company must submit data and go through review. Ingredients on the notified list skip that step by meeting a set concentration.

So 'contains tranexamic acid' on the front of a bottle may have nothing to do with functional review. Checking means looking up whether the product itself was reported as a functional cosmetic.