FDA Retires a 1995 Guidance on Topical Steroids

FDA finalised “Topical Dermatologic Corticosteroids: In Vivo Bioequivalence” on 14 July 2026, replacing a guidance of the same title that had stood since 2 June 1995. The science it governs — the skin-blanching test that decides whether a generic steroid cream matches the brand — is one of the oldest pharmacodynamic bioequivalence methods FDA uses, and until this week the document describing it predated the modern generic-drug user-fee program entirely.

What the guidance does

A generic drug wins approval by proving it is bioequivalent to the brand it copies. For a pill that means measuring the drug in the blood; for a corticosteroid cream rubbed into skin, blood levels are far too low to measure, so FDA relies on what the drug does rather than where it goes. The vasoconstrictor assay reads the response directly off the skin: a corticosteroid blanches the treated area as it constricts the small blood vessels beneath, and a chromameter quantifies how pale the skin turns. The guidance sets out how to run that test — a pilot study to fix the dose duration that produces a measurable but unsaturated response, then a pivotal study comparing the generic against the reference at that duration.

None of that is new; the method dates to the 1995 guidance and, in its essentials, earlier. What the 2022 draft and now the final do is modernise the recommendations around it — the pilot dose-duration response study, the pivotal vasoconstrictor study, subject selection, and the statistics — and consolidate them into the single document a potential ANDA applicant is expected to follow. FDA issued it under its good guidance practices regulation, 21 CFR 10.115.

What FDA changed

The notice does something the same day’s Federal Register did not always do: it says what moved between the draft and the final. There are two substantive changes and a category of editorial ones.

• Qualification and study subjects can differ. The final clarifies that the subjects used to qualify the chromameter and the operator need not be the same people enrolled in the pilot dose duration-response and pivotal vasoconstrictor bioequivalence studies — a practical point that separates instrument set-up from the trial itself.
• A new exclusion criterion. A history of hypopigmentation — a loss of skin pigment that would confound a test read out as a change in skin colour — was added to the reasons a subject is excluded from the study.
• Editorial clarifications. FDA states it made further editorial changes to improve clarity, without itemising them.

Neither substantive change rewrites the method. Both are the kind of refinement that comes from running the assay for three decades: one tidies how a lab qualifies its equipment, the other removes a source of measurement noise from the endpoint. That is what a finalisation of a mature method usually looks like — not a new test, but a cleaner description of the one already in use.

Why 1995 mattered

The document this replaces was issued the year before the modern generic-drug landscape took shape. In the intervening decades FDA built out its product-specific bioequivalence guidances, stood up the Office of Generic Drugs as it exists today, and negotiated three generic-drug user-fee programs. A method guidance from 1995 kept working because the underlying science held; but a live guidance that predates almost every process wrapped around it is exactly the sort of thing a review division and an applicant would rather not have to reconcile. FDA shared the 14 July Federal Register with its finalised guidance on formal meetings for OTC monograph drugs, another document closing a gap that had been open for years.

The usual caveat governs. The guidance “does not establish any rights for any person and is not binding on FDA or the public”, and an applicant may use an alternative approach that satisfies the statute and regulations. What changes on 14 July is not the law of bioequivalence but the reference an ANDA reviewer will now cite — and, after 31 years, which document that is.