How FDA tests whether a generic skin patch sticks

On Aug. 3, 2026, FDA announced in the Federal Register¹ that it had finalized guidance on a question that sounds small: whether a generic skin patch stays on as well as the brand. The agency first proposed an approach in a June 2016 draft⁷. Ten years and two revisions later, the final guidance² scores each patch from 0 (at least 90 percent stuck to the skin) to 4 (completely off), and asks a generic to show, statistically, that its average score is no more than 0.15 points worse than the brand's.

FDA spelled out the stakes in the guidance's background section. "When a TDS loses its adhesion during wear, the amount of drug delivered to the patient may be reduced," FDA wrote, using its shorthand for transdermal and topical delivery systems. A patch that falls off completely also raises "the risk of unintentional exposure of the drug product to an unintended recipient (e.g., a household member who may be a child)."

Patches are on FDA's list of complex products in the GDUFA III commitment letter¹⁴, and several still have no generic. FDA's June 2026 off-patent list¹⁵ includes extended-release films, a dosage form FDA's guidance counts among patches, of nitroglycerin, selegiline, granisetron and two estradiol combinations, with no patents, no exclusivity and no approved generic. This piece explains how FDA's adhesion and skin-reaction studies work and what is still being worked out. GxRadar's piece on the off-patent list explains that list.

A quarter century of patch guidance

FDA's written advice on generic patches predates the generic drug user fee program by 12 years.

Before 2016, adhesion advice lived in individual product-specific guidances. The 2016 draft said it would "supersede the recommendations related to adhesion studies provided in individual product-specific guidances" published before it, according to its Federal Register notice. Each patch's own guidance still sets the strength to test and other details.

How an adhesion study works

FDA's final guidance recommends a randomized crossover study in which each volunteer wears the generic (test) patch in one period and the brand (reference) in the other, at the same strength, usually on the same body site on opposite sides. Volunteers are typically healthy adults, the same population used for the patch's blood-level study.

Each study has to mimic real wear. Volunteers move freely and, for patches worn a day or longer, shower routinely without protecting the patch. Overlays, pressing a lifting edge back down and lotions on the site are ruled out. Patches are applied whole, because "altering the size or shape of the TDS may alter its adhesion characteristics."

Scoring happens at evenly spaced times across the brand's labeled wear period. FDA's minimums depend on how long the patch is worn:

Each reading uses FDA's five-point scale.

For the analysis, a patch never gets credit for recovering. The worst score seen so far carries forward until a worse one appears, so readings of 1, 2, 1, 3 count as 1, 2, 2, 3, and a patch that falls off scores 4 for the rest of the study. Averaging those scores gives each patch a mean adhesion score. The generic passes if the study shows, at a 0.05 significance level, that its mean score exceeds the brand's by less than 0.15. FDA asks for enough volunteers to give the study at least 80 percent power, and photographs at every reading to back up the observer's estimate. The final version also invites developers to propose alternative scales, as long as each reading is also reported on the five-point scale.

Irritation and sensitization studies

Patches sit on the skin for hours or days under a backing that traps moisture, and a strongly adhesive patch can strip away part of the skin's outer layer when it comes off. FDA's revised draft irritation guidance⁴, announced in a separate Aug. 3, 2026 notice⁶, says such reactions "can be unpleasant to the patient and may affect patient compliance, skin permeability, and/or adhesion." Even a low rate, FDA adds, "could affect thousands of individuals."

Irritation studies are deliberately harsh. For irritation alone, FDA recommends at least 40 evaluable volunteers wearing test and reference patches for 21 days, with each patch removed and replaced on the same skin site at the brand's labeled interval. When sensitization (an allergic skin reaction) also has to be checked, the minimum rises to 200 volunteers. After the 21-day phase and a rest of 14 to 17 days, patches go on fresh skin for 48 hours, and reactions are read at 30 minutes, 24, 48 and 72 hours after removal. Volunteers with a possible allergic reaction are rechallenged four to eight weeks later.

FDA's 2026 revision clarifies when the larger study may not be needed. FDA recommends the combined study when the generic contains a known sensitizing inactive ingredient absent from the brand, or an ingredient at an amount not previously used in a similar approved product. FDA says a sensitization assessment is "unnecessary or unethical" when the active ingredient is a known skin sensitizer or the patch's components and composition match the brand.

FDA's own research, part of the user-fee program described in GxRadar's piece on FDA's generic drug science program, points the same way. A retrospective review of a decade of approved generic patch applications, described in the FY 2025 GDUFA Science and Research Report¹⁶, found that irritation and sensitization study failures "did not correlate with specific inactive ingredients" in the products studied.

What is still open

The adhesion test still depends on human observers and volunteers. FDA's fiscal 2027 research priorities¹⁷ call for criteria for in vitro tests, such as "predictive adhesion performance of transdermal delivery systems," that could "eliminate the need for certain in vivo studies." Work under a grant to the Center for Research on Complex Generics, guided by an expert committee on adhesion testing, has completed initial evaluations of skin-mimicking test surfaces, the FY 2025 report says, and FDA held a Sept. 10, 2026 workshop¹⁸ on modernizing bioequivalence approaches for topical and transdermal products.

Supply and quality data show patches are not trouble-free. On FDA's shortage list as of Nov. 3, 2026, according to GxRadar's analysis of openFDA shortage data¹⁹, methylphenidate extended-release film is a current shortage and a rivastigmine film is listed as to be discontinued; both are on GxRadar's shortage tracker. GxRadar's count of openFDA enforcement records²⁰ from June 2012 through Oct. 28, 2026 finds 125 recall records whose product descriptions name a transdermal product, brand and generic combined: 2 Class I, 110 Class II and 13 Class III. The series is on GxRadar's recalls page.

Questions about how a particular patch should be worn, or what to do if one lifts, belong with a pharmacist or prescriber. The irritation guidance remains a draft, open to comment under docket FDA-2018-D-3546.