How to write a clinical trial or FDA milestone press release

A release about an investigational product may report results but may not represent the product as safe or effective: 21 CFR 312.7(a) prohibits promoting a drug that is still under investigation. In practice the copy states the endpoint, the effect size, the comparator, the population and the p-value, and adds no adjective to any of them. Read the trial's public registry entry before you write, because it is already published and the release has to agree with it.

This is the most regulated announcement type in ordinary commercial distribution, and the one where a distributor's opinion is worth least. What follows is what the published rules say, quoted, with links so you can read them yourself.

This is not legal, regulatory, medical or financial advice. If you are announcing trial data, a filing or an approval, your regulatory and legal advisers write this release. A distribution vendor does not.

The rule that shapes every sentence

An investigational product is not a product yet, and US regulation is explicit about what may be said. 21 CFR 312.7(a) provides:

A sponsor or investigator, or any person acting on behalf of a sponsor or investigator, shall not represent in a promotional context that an investigational new drug is safe or effective for the purposes for which it is under investigation or otherwise promote the drug.

The same section says this is "not intended to restrict the full exchange of scientific information concerning the drug, including dissemination of scientific findings in scientific or lay media" — the intent is to restrict promotional claims of safety or effectiveness and "to preclude commercialization of the drug before it is approved for commercial distribution".

That distinction is the whole craft of this category. Reporting a result is permitted. Characterising the result as evidence the product works, in a promotional register, is the thing the rule addresses. In practice it means the copy states outcomes and lets them stand: the endpoint, the effect size, the comparator, the population, the p-value. What it does not do is add an adjective.

The registry is already public, so write against it

There is a second constraint particular to trials, and it is the one most releases forget: the trial's registry entry usually exists before your announcement, and anyone can read it.

Under 42 CFR 11.24, registration information for an applicable clinical trial must be submitted "not later than ... 21 calendar days after the first human subject is enrolled". Under 42 CFR 11.44, results information must generally be submitted "no later than 1 year after the primary completion date", with defined exceptions where an approval application is pending.

The practical consequence: a journalist covering your topline release can open the registry record and compare your stated primary endpoint against the registered one. A release that leads on a secondary endpoint without saying so, or that describes a changed endpoint as the primary, is checkable in about ninety seconds. So name the endpoint you registered, say plainly whether the result is on the primary or a secondary, and give the enrolled number rather than the number analysed if they differ — or give both.

Which wires will not take this at all

The category is gated, and the two wires that publish clear rules go in opposite directions on the boundary.

Newsfile states plainly:

Medical/Health Content — We do not accept content about unreviewed, unregulated or unsubstantiated health supplements, medicine, surgical treatments or pharmaceuticals that can prevent or cure COVID19 or any other illness or condition.

EIN Presswire bans supplements outright irrespective of legality, and treats pharmacies as conditional on prescription status and a US or UK site. PRUnderground's wording is "Supplements of any kind, regardless of legality or FDA status". None of these rules tracks whether a product is lawful — they track the wire's exposure to health claims, which is a different thing and catches many companies by surprise. The full map of which industries wires refuse and what a compliance review can cost you are the two companion pieces worth reading before you buy.

Sponsors running regulated trials are not usually the ones caught by those bans. The companies caught are the adjacent ones — diagnostics, wellness devices, nutraceuticals — whose announcement sounds clinical and whose product is not.

What counts as news

Only a handful of things, and they are all events with dates:

• A regulatory submission accepted for review, or a decision date set.
• An approval, clearance, authorisation or designation granted, named exactly.
• A trial reaching a defined stage: first patient dosed, enrolment complete, database lock, topline results.
• An independent committee recommendation, with the committee named.

Designations are the ones most often overstated. A designation is a procedural status, not an assessment of efficacy, and writing it as an endorsement is precisely the promotional characterisation the rules address. Name it exactly, state what it confers procedurally, and stop.

The shape that clears editorial

Numbers are the substance here, and length is billed: PR Newswire's card charges $325 per additional 100 words above 400 on its US1 National circuit, so a full results release routinely runs into overage. That is the correct thing to spend money on in this category.

A headline and lede you could file

Meridine Therapeutics reports topline results from the Phase 2 ARDENT trial of MTX-114 in moderate plaque psoriasis

>

CAMBRIDGE, United Kingdom, 6 October 2026 — Meridine Therapeutics today reported topline results from ARDENT (NCT0XXXXXXX), a randomised, double-blind, placebo-controlled Phase 2 trial of MTX-114 in 212 adults with moderate plaque psoriasis. The trial met its primary endpoint: 44 per cent of patients receiving MTX-114 achieved a 75 per cent reduction in PASI score at week 16, against 19 per cent on placebo (p=0.003).

An identifier, a design, an n, the registered primary endpoint, both arms, and a p-value. No adjective anywhere. A reader can check every element of it against the registry.

What the announcement can and cannot do

It can put a precise, dated, citable account of the result or the filing into the public record, which matters for investigators, patient groups, partners and — for listed companies — for orderly disclosure. It can be the document that stops other people characterising your data for you.

It cannot make a result look better than it is to the audience that matters; clinicians read the numbers and ignore the framing. It cannot promote an investigational product, and the regulation is explicit about that. It cannot substitute for the publication or the conference presentation, which is where the data will actually be scrutinised. And it will not pass ranking credit — press links are overwhelmingly nofollow, which is normal rather than a defect and printed outlet by outlet where we have read it — so if the release is being bought for SEO, it is being bought for the wrong reason.

The check before you send

• Regulatory and legal counsel have signed the final text, not a draft.
• No safety or efficacy characterisation of an investigational product.
• Endpoint named exactly as registered, with primary and secondary distinguished.
• n, effect size, comparator and statistical result all present.
• Adverse events and discontinuations summarised.
• Designations described procedurally, not as endorsements.
• The wire's written policy on health content obtained before payment.