Niacinamide Is in a Third of Our Catalog. Does the Percentage on the Label Actually Matter?

We recounted the catalog on Thursday and got a number we had already published wrong.

On August 7 we wrote that niacinamide was in 16 of our products. That was true of the 50 products we had parsed at the time. We have now parsed all 54, and the real figure is 19, or 35.2 percent of the shelf. Still about a third. We would rather correct a count than let one drift quietly.

The figure that bothered us more is a different one. Of those 19 products, the number printing a niacinamide percentage anywhere on the label is zero.

Nineteen products, and the position swings from 2nd to 25th

We pulled every ingredient list in the store and found niacinamide at position 2 in the lacelltix Pepta Tox Ampoule and at position 25 in the FRUDIA Re:proust lip and eye remover. Nothing lists it first. Four products put it in the top five, eight land it between sixth and tenth, three sit in the teens, and four are past position 21.

Look at the FATION patch row for a second. Niacinamide is seventh out of seven ingredients, which sounds like a rounding error, except a hydrocolloid patch is mostly polymer by weight and there is very little room on that list for anything. Then look at Sunday To Monday: position 10 out of 87. Tenth place in an 87-ingredient formula is a much stronger signal than seventh place in a list of seven. The rank means nothing until you know the denominator, and even then it only gets you an ordering.

Twelve of the 19 are face products, five are body, two are scalp.

What Korea actually put in writing

There is exactly one place a government has attached a number to niacinamide in skincare, and it is Korean. The MFDS notice on reviewing functional cosmetics carries an annex listing the ingredients that get to skip the efficacy dossier, and niacinamide is number 7 on the whitening list at 2 to 5 percent. It shares that list with only eight others: broussonetia extract at 2 percent, arbutin at 2 to 5, ethyl ascorbyl ether at 1 to 2, oil-soluble licorice extract at 0.05, ascorbyl glucoside at 2, magnesium ascorbyl phosphate at 3, alpha-bisabolol at 0.5, ascorbyl tetraisopalmitate at 2.

The waiver comes with conditions almost nobody quotes. It applies to five dosage forms only: lotions, liquids, creams, solid forms and sheet masks. An ampoule or an essence outside those five does not get the shortcut. The permitted claim wording is fixed at "helps whiten the skin" and nothing stronger.

One absence in the same annex is worth more than the presence. Korea's anti-wrinkle waiver list runs retinol at 2,500 IU/g, retinyl palmitate at 10,000 IU/g, adenosine at 0.04 percent and polyethoxylated retinamide at 0.05 to 0.2 percent. Niacinamide is not on it. In Korean regulatory terms it is a brightening ingredient, full stop, which sits awkwardly next to the wrinkle claims it collects in marketing copy everywhere else.

The trials tested percentages, not each other

We went looking for the study that compares 2 percent against 5 against 10 in the same faces. We did not find one. What exists instead is a scatter of trials that each picked a number and stayed there.

Hakozaki and colleagues published the pigmentation work in 2002: niacinamide inhibited melanosome transfer from melanocyte to keratinocyte by 35 to 68 percent in coculture, with no effect on tyrosinase and no effect on melanin synthesis itself. The clinical side ran two arms, 18 subjects on 5 percent and 120 on 2 percent, and the 2 percent arm was paired with a sunscreen. All ten authors worked for Procter & Gamble.

Draelos tested 2 percent on sebum in 2006 and the result is more interesting than the way it usually gets cited. The Japanese cohort, 100 subjects split evenly against placebo over four weeks, showed a significantly lowered sebum excretion rate. The US cohort, 30 subjects split-face over six weeks, showed a drop in casual sebum level but no significant change in sebum excretion rate at all. The paper's own conclusion is that 2 percent "may be effective." Half of it missed.

Navarrete-Solís ran the only comparison we found against an active drug: 4 percent niacinamide on one side of the face, 4 percent hydroquinone on the other, 27 women with melasma, eight weeks. Melasma severity fell 62 percent on the niacinamide side and 70 percent on the hydroquinone side. Final lightness measured by colorimetry did not differ between sides, p = 0.78. Side effects ran 18 percent against 29 percent. Every patient also reapplied SPF 50+ every three hours for eight weeks, and there was no untreated arm, so a good share of that 62 percent belongs to the sunscreen and nobody can say how much.

For barrier, Soma put 2 percent against white petrolatum on the forearms of 28 people with atopic dry skin. Water loss through the skin fell on the niacinamide side and did not fall on the petrolatum side. For wrinkles, Bissett ran 5 percent split-face on 50 women for 12 weeks and reported improvement across fine lines, spots, blotchiness and sallowness. We could not get past the paywall on either full text, so we have the designs and none of the effect sizes, and we are not going to invent them.

Read that list again as a table of numbers rather than a list of findings. Sebum: 2 percent only. Barrier: 2 percent only. Melasma: 4 percent. Wrinkles: 5 percent only. Pigmentation: 2 and 5, in different studies, on different people, a decade apart.

Where the ingredient list stops being evidence

We have now used the same trick three times in three weeks. We read concentration off ingredient-list position for salicylic acid, for sodium DNA at 37th of 48, and for menadione at 20th of 30. It is a fair inference and we are going to keep making it. But this is the fourth article in a row to lean on it, so here is the ceiling.

Korea's cosmetic labelling rules say ingredients go in descending order of amount used, and then add the escape hatch in the same sentence: anything used at 1 percent or less, plus fragrance and colourants, may be printed in any order at all. Below that line the list is not sorted. It is a set.

Two more clauses in the same annex do quiet damage. Blended raw materials must be broken out into their individual component names, so a niacinamide-containing premix gets decomposed and niacinamide's printed position moves without the formula changing. And a company that can show a legitimate commercial interest may petition to print "other ingredients" instead of the actual names. The list is not guaranteed to be complete. The minimum legal type size, incidentally, is 5 point.

The EU wall sits in almost the same place with one wrinkle worth knowing. Regulation 1223/2009 orders ingredients "in descending order of weight of the ingredients at the time they are added," and lets anything under 1 percent go in any order. At the time they are added. Water driven off during processing, or an active delivered as a dilute pre-dispersion, both distort the finished-product ratio the printed order appears to describe.

What our screening found

The practical takeaway

If a product prints a niacinamide percentage, believe it and use it.

Almost none do, so the real question is what to do with the 19 that stay quiet.

Rank tells you two useful things and no third thing. If niacinamide appears before the first obvious under-1-percent marker on the list, a preservative like phenoxyethanol, a chelator like disodium EDTA, a gum or a fragrance, then it is very likely above 1 percent. If it appears after all of those, it is very likely below. Above that marker the list is genuinely ordered and you can trust the ordering. Past it, the list is a bag.

The second thing is a warning rather than a reading. A short list flatters position and a long list buries it, which is why the FATION patch at 7 of 7 and Sunday To Monday at 10 of 87 cannot be compared at all. We put the full position table for all 51 benchmarked products through the same treatment when we scored ourselves against Amazon, and the same denominator problem showed up there.

What you should not do is chase the number upward. The trial record tops out at 5 percent for wrinkles and 4 for melasma, and the higher-percentage serums on the market are running ahead of published evidence, not behind it. Niacinamide at 10 percent is also more likely to sting, which the safety review noticed at concentrations above the ones it actually reviewed. Our awards page lists what passed our screen; none of it was chosen for a percentage nobody printed.