Cosmetics Claims Are Borrowing Drug-Level Confidence Without Drug-Level Proof

A cosmetics brand can print "clinically proven" on a bottle after running a single in-home consumer perception test with no control group. A drug company making a comparable claim needs, at minimum, one adequate and well-controlled trial plus confirmatory evidence under the FDA's 2023 draft guidance on substantial evidence of effectiveness. That gap between what cosmetics claims actually require and what shoppers assume they require sits at the center of a growing wave of enforcement actions, and it traces back to a legal distinction packaging never discloses: cosmetics claims only have to clear the Federal Trade Commission's "competent and reliable scientific evidence" standard, a flexible test with no fixed study-count requirement, as laid out in the FTC's Health Products Compliance Guidance.

Cosmetics and Toiletries described the underlying tension plainly: companies want the credibility of pharmaceutical-style claims without the pre-market evidentiary burden that pharmaceutical companies actually carry. The article noted that FDA warning letters now demand substantiation for every claim in an advertisement, including comparative claims, not just the headline promise. That shift matters because comparative language, phrases like "more effective than" or "outperforms," is exactly where cheap consumer-perception surveys tend to stand in for the controlled studies a comparative claim would need to survive scrutiny.

The National Advertising Division, the ad industry's self-regulatory body, spent 2025 cataloguing what that substitution looks like in practice, according to Happi's year-end roundup of NAD decisions. NAD rejected SolaWave's substantiation for its skincare wand after finding the underlying in-home test had no washout period and no blinding, the two controls that separate a measured result from a placebo effect. In a challenge brought by Johnson & Johnson's Neutrogena, NAD recommended L'Oreal drop "#1 Dermatologist Recommended" language for CeraVe. The same monitoring cycle pushed back on Amyris's Biossance brand over ingredient claims and on NuOrganic over eyelash-growth promises, each case turning on the same question: does the underlying study design match the strength of the claim built on top of it.

FDA's own enforcement has moved in parallel. In an April 2026 warning letter, the agency cited Purolea Cosmetics Lab for marketing products with language like "Shingles Relief" and "Genital Herpes Relief," claims that cross from cosmetic into disease-treatment territory regardless of what the label calls the product. That crossing point, known in FDA regulation as the "intended use" test under the Food, Drug, and Cosmetic Act, is also why the word "cosmeceutical" carries no legal standing with the agency at all. Dermatology Times has described it as industry shorthand that lets companies gesture at drug-like performance through euphemisms such as "supports cellular renewal" while staying just outside the FDA's drug definition.

Underneath both the NAD challenges and the FDA letters sits a narrower technical problem: the lab tools generating cosmetic efficacy data are not built to answer the questions brands ask of them. A review in the International Journal of Cosmetic Science by Filaire and colleagues found that in vitro skin models still lack immune cells, vascularization, and full barrier function, which makes them reliable for toxicology and safety screening but a poor proxy for the kind of human efficacy benefit a label claim implies. Closing that gap requires in vivo human testing, the accepted gold standard, but it is slow and expensive, which is precisely the commercial pressure that pushes brands back toward cheaper in vitro data or consumer-perception scores that were never designed to carry a claim's full evidentiary weight.

Theresa Callaghan of Callaghan Consulting International, writing elsewhere in Cosmetics and Toiletries, named the skills gap behind that pattern directly: many claims professionals lack the ability to read across in vitro and ex vivo data into a defensible human study design. That is a different failure than fraud. It is a translation problem, between what a petri dish or a reconstructed skin model can show and what a sentence on a bottle promises a person's face will do, and it is the specific terrain the FTC and FDA standards disagree about how strictly to police.