No published trial has compared 10% niacinamide with 5% on the same faces. The human studies behind niacinamide's reputation used 2%, 4% or 5%, and Korean law ties the whitening claim to 2–5%. Above that range you are buying a number, not evidence.
The clients who arrive with the highest percentage on the bottle are usually the ones whose skin is in the worst shape. The sequence repeats: 5% stopped feeling like it worked, so they moved to 10%, layered over an acid toner and a vitamin C. The number went up as the skin went down.
What is niacinamide, and what is it supposed to do?
Niacinamide is the amide form of vitamin B3, used in leave-on skincare to even out diffuse pigment, support the barrier and settle redness — a small, water-soluble molecule that Korea's functional-cosmetic monograph lists as C6H6N2O, MW 122.13. Rather than blocking tyrosinase, it is thought to interfere with the handoff of melanosomes from melanocytes to keratinocytes; in co-culture that transfer fell 25–45%. But that is cell culture, and a 2021 review in Antioxidants is blunt: there is no convincing evidence for a specific molecular target.
How much niacinamide was actually in the studies?
Between 2% and 5%, in every human trial that built the reputation. The pigment work most cited is Hakozaki's pair of split-face studies: 18 Japanese women using 5% twice daily for 8 weeks, and a larger one of 120 Japanese women using 2% inside an SPF 15 moisturiser for 8 weeks. Procter & Gamble's aging-appearance work used 5% in 40 women over 12 weeks; the acne work used 4%.
The only published dose comparison ran 5% against 2% against vehicle in 79 Japanese women for 8 weeks. The 5% side significantly reduced spots; the 2% side did not differ from vehicle. That single trial is the entire published basis for "more is better" — two doses, neither of them 10%. Against a prescription comparator the effect is modest: in 27 people with melasma, 4% niacinamide and 4% hydroquinone over 8 weeks gave good-to-excellent improvement in 44% and 55%, with fewer side effects on the niacinamide side.
Why does Korean law stop at 5%?
Because 2–5% is the range Korea's regulator will attach a whitening claim to. Under the Regulation on Examination of Functional Cosmetics there are nine notified whitening ingredients with fixed content ranges — arbutin 2–5%, bisabolol 0.5% — and niacinamide sits at 2–5%. Stay inside the range and the product qualifies as a 미백 기능성화장품 without filing its own efficacy dossier.
That is a better label rule than the number on the box. A bottle advertising 10% is not making a stronger regulatory claim — it is making no functional claim at that figure at all, and industry reviews put cosmetic use as wide as 2–20%. So in a Seoul pharmacy or on an Olive Young shelf: find the 기능성화장품 (미백) mark on the back panel, and choose a plain 2–5% lotion or emulsion used across the whole face twice daily, as the trials did, over a high-percentage serum dabbed on spots. The largest dataset of all tested 2% inside a sunscreen. Melasma is a 피부과 conversation. Formats, not brands.
Does it get in, and does the eye area behave differently?
It gets in; nobody has published how much reaches the cells you are aiming at. At 122 Da it sits far below the ~500 Da rule of thumb for crossing the stratum corneum, so permeability is not the bottleneck. But both absorption figures in circulation are weak: an in vitro study on abdominal skin using an acetone carrier — not a cosmetic vehicle — found 28.8% of the dose in the receptor fluid at 24 hours, and the figure of roughly 10% from a 2% product is unpublished company data.
Thinner sites are assumed to absorb more, and that assumption is less measured than you would think. A review of regional percutaneous absorption found only eight usable human in vivo studies, concluding the data "seem to suggest increased PA in the forehead and genital skin," while the datasets are "scarce and inadequate for drawing complete conclusions." There are no eyelid or neck numbers for niacinamide at all. Going lighter around the eyes is caution, not a measured dose — and much of what people want fixed there is vascular shadow, not melanin, which a pigment-pathway ingredient cannot touch.
Know your own skin before you buy anything
An ingredient earns a place in your routine only if your skin, as it is this week, can take on a new variable — a different question from whether it works, and the one almost nobody asks.
Think of someone seriously unwell deciding to get healthy by training harder. The instinct is admirable, the timing ruinous: a body in that state converts effort not into recovery but into further damage. What it needs first is rest. A stinging, tight, reactive barrier is in the same position, and even a well-evidenced active is still load.
So when the barrier is down, the move is subtraction. Strip back to a gentle cleanse, one bland moisturiser and sunscreen, and hold there until the skin stops complaining — weeks, not days. Then bring back one thing, at the low end of its range, and give it the eight to twelve weeks the trials gave it. Niacinamide at 2–5% is a sensible first thing to reintroduce, and a poor thing to pile onto a struggling face. None of this replaces medical diagnosis: persistent redness, melasma, or a reaction that is worsening belongs with a dermatologist.
The case against some of this
Three of my own claims above are weaker than I made them sound.
1. "Niacinamide supports the barrier." The human evidence is thin. The ceramide mechanism — up to five-fold, dose-dependent — is from cultured keratinocytes, not ceramides measured in anyone's skin. The human figure usually quoted, a 27% drop in transepidermal water loss from 2% on dry lower legs over four weeks, appears in a conference review by scientists at the company selling it, with no sample size stated. Narrowed: 2–5% stays in my barrier-supportive routines, but I should not have implied it rebuilds your ceramides.
2. "5% is the useful ceiling." This rests on one 79-person trial plus a regulatory range set for claim approval, not a dose-response curve. An integrative review of randomised melasma trials found only four studies, could not report a combined participant total, and noted three of the four did not specify morning, night or both — concluding there is "very little evidence published in the last decade using pure niacinamide." Narrowed: 2–5% is the range with trials behind it and the range Korean law recognises, not a biological optimum. I do not know 10% is worse, only that nobody tested it.
3. "It fades pigment." Effect sizes are modest and do not hold. In the 5%-versus-2% trial the 5% benefit was no longer statistically significant by week 42 of the regression phase. For acne, the 2020 Cochrane review covers four nicotinamide trials (roughly 374 participants), rates certainty of evidence low, judges risk of bias "unclear" in all three clindamycin trials, and notes no trial measured participants' own global assessment or quality of life. Narrowed: expect modest, maintenance-dependent improvement in diffuse pigment over two to three months — not melasma treatment, not an acne therapy.
Is 10% niacinamide better than 5%?
No published trial compares them, so nobody knows. The one dose comparison that exists ran 5% against 2% and found 5% better. Everything above 5% is extrapolation, and Korea's whitening claim stops at 5%.
Does niacinamide shrink pores or reduce oil?
The oil evidence is tissue-level, not clinical: in facial biopsies cultured for four days, 25 mM niacinamide cut sebaceous lipogenesis by 42% — excised tissue in a dish. Pore-appearance claims rest largely on image-analysis endpoints in industry-run studies.
How long before I would see anything, and will it last?
The trials ran 8 to 12 weeks of twice-daily use before measuring, so judge it at two to three months, not two weeks. Treat it as maintenance: in the one trial that followed people after they stopped, the pigment benefit had faded to non-significance by week 42.
At Laurel Studio the percentage on your bottle is nearly the last thing we look at. We start from what your barrier can currently tolerate — a different answer for a visitor three days off a long flight than for someone who lives here. If you would rather not work that out by trial and error, a 1:1 consultation is the shortest route.
Sources
- Matts PJ, Oblong JE, Bissett DL. A review of the range of effects of niacinamide in human skin. IFSCC Magazine, 2002. Full paper read. The authors were employed by Procter & Gamble. Source of the 2% / 27% TEWL figure, the keratinocyte ceramide data, the 25 mM facial-biopsy lipogenesis data, the acetone-vehicle absorption figure and the unpublished internal penetration estimate. PDF
- Boo YC. Mechanistic basis and clinical evidence for the applications of nicotinamide (niacinamide) to control skin aging and pigmentation. Antioxidants, 2021;10(8):1315. Full text read. Source of the Hakozaki trial details (n=18, n=120), the 5%-vs-2% dose-response trial (n=79), the melasma comparison with hydroquinone (n=27), and the statement that no specific molecular target is established. Open access
- Lev-Tov H, Maibach HI. Regional variations in percutaneous absorption. Journal of Drugs in Dermatology, 2012;11(10):e48–e51. Full text read. Source of the regional-absorption quotations.
- Republic of Korea, Ministry of Food and Drug Safety. Regulation on Examination of Functional Cosmetics (기능성화장품 심사에 관한 규정), MFDS Notice No. 2025-88, effective 16 December 2025. Regulation page and whitening monograph annex confirmed directly (C6H6N2O, MW 122.13). The annex file listing the 2–5% content range could not be opened; that range and the comparator ingredient figures come from a regulatory-compliance database summarising the annex, so treat them as second-hand rather than read from the primary text.
- Niacinamide-containing facial moisturizer improves skin barrier and benefits subjects with rosacea. Cutis, 2005;76(2):135–41. Abstract only. 50 subjects, randomised investigator-blinded, 4 weeks; the abstract states neither the concentration nor numerical results.
Sources for the case-against section
- Liu H et al. Topical azelaic acid, salicylic acid, nicotinamide, sulphur, zinc and fruit acid (alpha-hydroxy acid) for acne. Cochrane Database of Systematic Reviews, 2020. Full text read. Four nicotinamide trials; certainty of evidence rated low; risk of bias unclear in all three clindamycin trials; no nicotinamide trial measured participants' global self-assessment or quality of life; 26 of 49 included studies judged high risk of bias in at least one domain. PMC
- Pedroso AG, Furtado GRD, Barbosa KL. Niacinamide for the treatment of melasma: an integrative review of randomized clinical trials. Research, Society and Development, 2022;11(11):e198111133581. Full text read. Four included studies; no combined participant total reported; three of four did not specify application schedule; "very little evidence published in the last decade using pure niacinamide."
- The 42-week regression result and the 2%-fails-versus-vehicle finding are from the dose-response trial as reported in Boo (2021), cited above.
- The sample-size discrepancy for the rosacea study (48 vs 50 subjects) was identified by comparing Matts et al. (2002) with the Cutis (2005) abstract, both cited above.
