Structure/Function Claims: Where Supplements Become Drugs

The Sentence That Costs Companies Their Product Line

I've read enough FDA warning letters to know exactly where most of them start going wrong: not in the manufacturing suite, but in the marketing folder. A single testimonial gets posted on a website or handed to a sales rep as a "customer story." It references a named clinical study. It claims the product helped someone avoid a diagnosed disease. Legal never saw it. Regulatory never saw it. Now the label review committee — if one exists — has a problem that no amount of good manufacturing practice can fix after the fact.

FDA's January 22, 2026 warning letter to Diamond-Herpanacine of PA, Inc. (case number 713854-01222026, issued by the agency's Human Foods Program) is a useful case study precisely because it isn't unusual. The lesson isn't about one company's website copy. It's about a structural gap that exists across a huge share of supplement operations: nobody owns the line between "this supports eye health" and "this treats macular degeneration." That line isn't a marketing nuance. It's a legal boundary defined in the Federal Food, Drug, and Cosmetic Act, and crossing it changes what your product legally is.

How a Supplement Becomes an "Unapproved New Drug"

Under Section 201(g)(1)(B) of the FD&C Act, a "drug" is defined — in part — as any article "intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease." That definition doesn't care what's printed on your label as the product category. It cares about intended use.

Intended use is established by the totality of the evidence: labeling, website claims, social media posts, sales materials, and testimonials the company chooses to publish or distribute. A product can be formulated, labeled, and sold as a dietary supplement under 21 U.S.C. § 321(ff) and still be treated by FDA as an unapproved new drug the moment its marketing claims disease treatment or prevention.

When that happens, two violations stack automatically:

• Unapproved new drug (FD&C Act § 505(a)). The product hasn't gone through NDA or ANDA approval, but its marketing now claims a drug's effect.
• Misbranding (FD&C Act § 502(f)(1)). A product marketed as a drug needs adequate directions for use and dosing for the specific disease — information a supplement label was never built to carry.

You don't get to be "a little bit of a drug." Once the claim crosses the line, the whole product falls under drug law.

Referencing a named research program is a common trigger for this. AREDS — the Age-Related Eye Disease Studies, run by the National Eye Institute — is the one that shows up most often in eye-health supplement marketing. AREDS and its follow-up, AREDS2, studied specific nutrient formulations for their effect on progression of age-related macular degeneration, a diagnosed disease.

If a company's marketing implies its own product replicates or delivers those studied benefits, FDA reads that as a disease claim, not a nutritional support claim. This is especially true when the language comes through a testimonial saying something like "I stopped worrying about my diagnosis." The Supplement Facts panel doesn't matter at that point. The claim already decided what the product is.

The Line Between Structure/Function Claims and Disease Claims

The Dietary Supplement Health and Education Act (DSHEA) of 1994, codified at 21 U.S.C. § 343(r)(6), lets supplement makers describe how a nutrient or ingredient affects the structure or function of the body, without FDA pre-approval. Three conditions have to be met at the same time:

• No disease claim. The statement can't say the product treats, prevents, cures, mitigates, or diagnoses a disease.
• Substantiation on file. The manufacturer must have evidence the claim is truthful and not misleading, before the claim runs.
• 30-day FDA notification. The manufacturer must notify FDA of the claim within 30 days of first marketing it, under 21 CFR 101.93(g).

That third condition is the one most small and mid-sized supplement companies forget entirely. It isn't optional, and FDA can and does check whether a notification letter exists for a claim it's investigating.

The required disclaimer under 21 CFR 101.93(c) — "This statement has not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease" — has to appear on the label in a specified type size and placement. But the disclaimer doesn't cure a disease claim. I've seen companies treat it as a legal shield that lets them say almost anything above it. It isn't one. If the claim itself describes disease treatment, the disclaimer sitting below it doesn't change what the claim says.

Here's the comparison that belongs on every label reviewer's desk:

Notice the difference between the top row and the third row isn't the ingredient. It's the verb. In the testimonial row, it isn't even a verb the company wrote directly — it's a customer's words that the company chose to publish. FDA treats a company-selected, company-published testimonial as the company's own claim. That's a distinction plenty of marketing teams miss.

Why Testimonials Are the Highest-Risk Claim Source

Formal label copy usually goes through some kind of review before it ships, even if that review is informal. Testimonials rarely do.

They arrive as unsolicited praise. They feel authentic. They get published because they're persuasive — which is exactly why they're dangerous. A customer describing their own experience isn't bound by FDA's rules. The company publishing that testimonial is. The moment a business selects, edits, or features a testimonial that references a diagnosis, a named study, or a disease outcome, that testimonial becomes labeling under FD&C Act § 201(m). It's then evaluated by the same disease-claim standard as anything printed on the bottle.

The practical failure mode looks like this: a customer emails a glowing story that name-drops a study they read about. AREDS is common in eye-health supplements, but the same pattern shows up with cardiovascular, joint, and cognitive-health products citing other named research. Marketing publishes it because it's compelling copy, and nobody flags that quoting a named clinical trial in the same breath as a personal disease outcome reframes the product as something that delivers the trial's studied effect.

Multiply that across a website, a few thousand social posts, and years of accumulated customer-review pages, and a company can build an enormous compliance liability without ever intending to make a drug claim on its actual label.

What a Working Label Claims Review System Actually Requires

I've built these systems for clients across supplements, cosmetics, and OTC drugs, and the ones that hold up share a few non-negotiable elements. None of them are exotic. All of them require someone with actual authority to say no.

• A designated claims reviewer with veto power. Not a committee that meets quarterly — a specific person or role that reviews every piece of outward-facing content before it publishes, including testimonials, social posts, influencer scripts, and trade show signage. If marketing can publish without that sign-off, the system doesn't exist in any meaningful sense.
• A substantiation file per claim, not per product. Each structure/function claim needs its own file showing the scientific basis, dated and retained, because FDA can request substantiation during an inspection or in response to a warning letter. "We have studies on the ingredient" isn't the same as "we have a substantiation file mapped to this exact claim language."
• A standing list of banned words and constructions. "Prevents," "treats," "cures," "reverses," and references to named diseases (macular degeneration, diabetes, arthritis) or named disease-specific studies (AREDS, the Women's Health Initiative, any NIH-funded trial tied to a diagnosed condition) go on a checklist every content creator sees before anything ships. This is a five-minute training, repeated often enough that new hires and outside agencies don't rediscover the line the hard way.
• A testimonial intake and approval workflow. Every testimonial gets the same review as label copy before publication, with no exception for "it's just a customer quote." If a testimonial mentions a diagnosis, a named study, or a disease outcome, it either gets edited to remove the disease reference or it doesn't get published.
• A 30-day notification tracker. Every new structure/function claim gets logged with its first-marketed date and a corresponding FDA notification letter under 21 CFR 101.93(g), filed and dated. This is a spreadsheet, not a system, but almost nobody keeps it, which is exactly why FDA finds the gap so often.
• cGMP documentation under 21 CFR Part 111 for the underlying manufacturing. A warning letter almost never arrives citing only labeling; once FDA is in your facility investigating a claims issue, quality system gaps compound the exposure.

None of this requires new software or outside certification. It requires someone in the organization with the standing to stop a testimonial from going live, and a paper trail that shows the company took its own claims seriously before FDA had to ask.

If a Letter Has Already Arrived

The same claims-review gap that creates the violation also shapes how badly the response goes. FDA expects a written reply within 15 working days that maps point-by-point to each cited claim — with dates and evidence of correction, not a general promise to "review our practices." A response that fixes the specific sentence FDA flagged but leaves the review gap in place invites a repeat citation, because the process failure that produced the claim is still sitting there, ready to produce the next one. If you're working through that response now, our guide on how to respond to an FDA Form 483 walks through the timeline and structure in more depth than fits here.

Building the Habit, Not Just the Checklist

The companies I've seen avoid this entire category of warning letter don't have more lawyers than everyone else. They have one habit: every piece of content that leaves the building — label, website, social post, testimonial, trade show banner — passes through the same three questions before it publishes. Does this claim describe structure or function, or does it describe disease treatment? Do we have substantiation on file for the exact language used? If it's a testimonial, do the customer's own words cross into disease territory even though we didn't write them?

That habit costs almost nothing to build, and it's cheaper than any warning letter response. It only works if someone owns it before the content ships, not after FDA asks about it.

For companies working through a claims audit or preparing for potential FDA scrutiny, our FDA compliance services walk through label review, substantiation files, and inspection readiness together, since these issues rarely show up in isolation.

FAQ

What's the legal difference between a structure/function claim and a disease claim? A structure/function claim describes how a nutrient affects the body's normal structure or function (e.g., "supports eye health") and is permitted under 21 U.S.C. § 343(r)(6) without pre-approval, provided it's substantiated and FDA is notified within 30 days under 21 CFR 101.93(g). A disease claim states or implies the product diagnoses, treats, cures, mitigates, or prevents a disease, which makes the product an unapproved new drug under FD&C Act § 201(g)(1)(B) and § 505(a).

Can a customer testimonial turn a supplement into an unapproved drug? Yes. Under FD&C Act § 201(m), a company-selected and published testimonial is treated as the company's own labeling. If that testimonial references a diagnosed disease or implies the product delivered a disease-treatment outcome, FDA evaluates it under the same disease-claim standard as printed label copy.

Does the FDA disclaimer on supplement labels protect against disease-claim violations? No. The disclaimer required under 21 CFR 101.93(c) is a mandatory notice, not a legal shield. If the underlying claim describes disease treatment or prevention, the disclaimer doesn't change the legal character of the claim.

How long does a company have to respond to an FDA warning letter? FDA expects a written response within 15 working days of the letter's date, detailing the specific corrective actions taken or planned, including evidence that the company has addressed the systemic cause of the violation, not just the specific language cited.

What is the 30-day notification requirement under 21 CFR 101.93(g)? Any company making a structure/function claim on a dietary supplement must submit a notification to FDA describing the claim within 30 days of first marketing the product with that claim. Many companies never file this notification, which is itself a separate compliance gap FDA can identify during an investigation.

Last updated: 2026-08-12