FDA OPDP Untitled Letters in 2025 and 2026: What 120 Letters Cited

FDA's Office of Prescription Drug Promotion (OPDP) published 5 untitled letters in 2024 and 63 in 2025, 12.6 times as many. Fifty-two of the 63 carry one date, 9 September 2025. Another 25 followed in 2026 through 26 October.
Across all 120 untitled letters on FDA's page, the most common problem was overstated or unsupported efficacy: 80 letters, 67%. Risk information that was not clear and conspicuous came next (56, 47%), then risk that was omitted or minimized (50, 42%). TV ads drew 58 of the 120 letters.
| Year | Untitled letters | Promotion warning letters | Untitled letters with a close-out |
|---|---|---|---|
| 2017 | 2 | 1 | 0 |
| 2018 | 5 | 2 | 0 |
| 2019 | 7 | 3 | 0 |
| 2020 | 2 | 3 | 0 |
| 2021 | 4 | 2 | 4 |
| 2022 | 3 | 1 | 3 |
| 2023 | 4 | 1 | 4 |
| 2024 | 5 | 0 | 5 |
| 2025 | 63 | 10 | 55 |
| 2026 (to 26 October) | 25 | 2 | 15 |
Source: FDA's untitled letters page and warning letter pages, counted by
scripts/analyze-opdp-letters.py. Data through 26 October 2026, snapshot taken
8 November 2026. Download the letter-level data (CSV)
or the category counts. You can search the
same letters in our OPDP letter search.
The 9 September 2025 batch
On 9 September 2025 FDA announced a crackdown on direct-to-consumer drug advertising. Its press release said: "Today, the FDA is sending thousands of letters warning pharmaceutical companies to remove misleading ads and issuing approximately 100 cease-and-desist letters to companies with deceptive ads."
Most of those letters are not public. What FDA published from that day is 52 untitled letters to 30 companies and 8 warning letters, to Eli Lilly (3), Aytu Biopharma (2), Novo Nordisk, CSL Behring and Alora Pharmaceuticals.
The 52 untitled letters share some traits.
- Most are TV ads. Thirty-six of the 52 concern a TV ad. Forty of the 52 are aimed at consumers.
- FDA wrote one letter per piece. UCB, Boehringer Ingelheim, AstraZeneca and Phathom each received three letters on one product, one for each ad or video. AstraZeneca got five letters that day, for Fasenra and other products.
- They stopped citing the law. Of 35 untitled letters dated before the batch, 31 cite the Federal Food, Drug, and Cosmetic Act by U.S. Code section. Of the 52 in the batch, 1 does. Of the 33 since, 2 do. The letters still say the piece "misbrands" the drug, and they quote the claims and the label, but they rarely name 21 U.S.C. 352 or 21 CFR 202.1 any more.
What FDA cited
The table counts letters, not violations. A letter can carry several categories. The categories are our labels, assigned by a model that read each letter, with an accuracy check described under Caveats.
| What FDA cited | All 120 | Before 2025 (32) | 9 Sep 2025 batch (52) | After the batch (33) |
|---|---|---|---|---|
| Overstated or unsupported efficacy | 80 (67%) | 19 (59%) | 39 (75%) | 19 (58%) |
| Risk not clear and conspicuous | 56 (47%) | 13 (41%) | 25 (48%) | 18 (55%) |
| Omits or minimizes risk | 50 (42%) | 18 (56%) | 14 (27%) | 16 (48%) |
| Unsupported superiority or comparison | 21 (18%) | 9 (28%) | 3 (6%) | 7 (21%) |
| Broader use than approved | 20 (17%) | 8 (25%) | 3 (6%) | 9 (27%) |
| Misleading dosing or convenience | 19 (16%) | 5 (16%) | 6 (12%) | 7 (21%) |
| Missing required disclosure | 6 (5%) | 3 (9%) | 1 (2%) | 2 (6%) |
| Not submitted on Form 2253 | 5 (4%) | 3 (9%) | 0 | 2 (6%) |
The batch leaned toward how benefit is shown. Efficacy was the lead category in 75% of its letters, and omitted risk fell from 56% of earlier letters to 27%. The pre-2025 group is 32 letters, so treat small gaps with caution. After the batch, the mix moved back toward earlier years: omitted risk at 48%, broader use at 27%.
The format explains much of this. Among the 58 untitled letters about a TV ad, 67% cite risk presentation, against 27% of the 62 that are not about a TV ad. The TV letters read like a checklist from 21 CFR 202.1(e)(1)(ii). The major statement has to appear in audio and on-screen text together, in plain language, for long enough to read, with no visuals likely to interfere. Letters describe scene changes, music and characters during the risk statement. For example, FDA told Vanda in October 2026 that eleven scene changes in a car chase competed with the major statement for the viewer's attention.
Efficacy letters follow a pattern too. FDA quotes the claim, finds the label or trial result it rests on, and says the claim goes further. It then notes that a disclaimer such as "individual results may vary" does not correct the impression. The Tremfya letter of February 2026 is typical. It said that "many people" or "~1 out of 2 patients" reaching clinical remission at one year overstated results of 38% and about 39% in the two studies.
Thirty-one untitled letters are aimed at healthcare professionals. Their profile differs: 71% cite efficacy and 32% cite superiority, but only 13% cite risk presentation.
Letters since the batch
Thirty-three untitled letters are dated after 9 September 2025. The pace slowed through 2026: 19 untitled letters from January to June, then 6 from July to October.
Several follow earlier letters. FDA wrote ARS Pharmaceuticals about a Neffy TV ad in September 2025 and about another Neffy TV ad in January 2026. The second letter says the first letter addressed similar claims. ImmunityBio received untitled letters in September 2025 and January 2026 for Anktiva webpages. FDA closed both out on 27 January 2026. On 13 March 2026 it sent a warning letter about an Anktiva TV ad and a podcast, and said ImmunityBio "continues to promote Anktiva in a similarly misleading manner" despite the untitled letters.
Close-outs
FDA's table lists no close-out for any untitled letter before March 2021, so the analysis counts only the 104 letters since then. Eighty-six of them (83%) have a close-out letter. The median wait from letter to close-out is 76 days. The fastest was 13 days (Althera's Roszet) and the slowest 280 days (ARS's Neffy).
| Group | Letters | With a close-out | Median days |
|---|---|---|---|
| 2021 to 2024 | 16 | 16 (100%) | 63 |
| 9 September 2025 batch | 52 | 45 (87%) | 85 |
| 2025 before the batch | 3 | 3 (100%) | 46 |
| After the batch | 33 | 22 (67%) | 58.5 |
No letter in the batch was closed out in under 60 days. Seven of the 52 have no close-out after 14 months: UCB (Rystiggo), Supernus (two Qelbree letters), Novartis (Kisqali), Nalpropion (Contrave), Amgen (Imdelltra) and Alexion (Ultomiris). Three letters since the batch that are more than six months old are also open: two to Novo Nordisk and one to Bristol Myers Squibb for Cobenfy.
A close-out letter says FDA evaluated the response and "it appears that you have addressed the violations". The same letter says it does not stop later action. ImmunityBio shows what that means.
Who received them
By untitled letters: AstraZeneca 6, Bristol Myers Squibb 5 (counting Karuna and Mirati, which FDA's table names as Bristol Myers Squibb companies), Novartis 5, UCB 4, AbbVie 4, and Pinnacle Biologics, BeOne Medicines and Pfizer 3 each. Because FDA writes per piece, these counts measure how many pieces FDA cited, not how many problems a company has.
Six products appear in three letters each, counting warning letters: Photofrin, Anktiva, Bimzelx, Voquezna, Jardiance and Fasenra.
Warning letters
The 25 warning letters about drug promotion look different. Twenty-two cite omitted or minimized risk (88%). Ten cite a failure to submit the piece to FDA at first use, against 5 of 120 untitled letters. Eight concern a spokesperson or influencer, and nine are online video. Seventeen of the 25 warn that failure to correct may bring seizure or injunction. None of the 120 untitled letters says that.
Ten of the 25 are dated 9 September 2025 or later. Two of those, to Eli Lilly and Novo Nordisk, concern the same video, "An Oprah Special: Shame, Blame, and the Weight Loss Revolution".
What a reviewer can use
- Check the major statement as a viewer. For a TV ad, watch the risk statement with the sound off and then with the picture off. FDA's test is whether text and audio match and whether the visuals compete.
- Read each superlative against the label. "Only" draws superiority letters. A December 2025 Nexlizet ad called bempedoic acid the only nonstatin approved to reduce the risk of heart attack in statin-intolerant patients, and FDA pointed out that certain PCSK9 inhibitors are approved for that too.
- Trace every number to a table. The Tremfya letter is about a subgroup result stretched into "many people". The Qelbree letter is about children shown with their symptoms gone entirely.
- Do not rely on the disclaimer. FDA says in letter after letter that a disclaimer does not cure a misleading impression.
- Submit at first use. Not submitting on Form FDA 2253 is rare in untitled letters and common in warning letters. The rule is 21 CFR 314.81(b)(3)(i).
- Keep a list of similar claims. Every letter asks the company to list all promotional pieces with similar representations, with their 2253 dates.
The MLR review guide covers the process around these checks.
Caveats
The categories are labels. Claude Opus 5.5 read each letter against the eight categories and quoted a sentence for each. A script confirmed every quote appears in the letter. We checked 22 letters by hand: the model gave 48 labels, 46 were right, and 3 labels it should have added were missing. In 17 of 22 letters the set was exactly right. A second model, Claude Sonnet 5.5, agreed with the first on 88% of labels across all 145 letters and chose an identical set for 111 of them. The largest gap was the "missing required disclosure" label, which the second model added to 12 letters.
These are the published letters. FDA said it sent thousands of letters in September 2025. Only a small fraction is public, and the data cannot say how many more FDA sent.
The warning letters are selected by rule. FDA files warning letters under different offices from year to year. We included letters issued by OPDP, and letters that open by saying FDA reviewed a promotional communication and cite the advertising provisions. Letters to online pharmacies and compounders are in the warning letter search, not here. An unusual letter could be missed.
Dates follow FDA's table. Some close-out letters cite a different date for the letter they close than the table shows. Two close-out letters and one untitled letter have no text layer and were read by optical character recognition.
Method. scripts/opdp/ in this site's repository fetches the letters,
labels them, and builds the dataset. scripts/analyze-opdp-letters.py reproduces
every figure above.
Frequently asked questions
How many OPDP untitled letters are there?
FDA's page lists 120, dated 18 May 2017 to 26 October 2026. FDA archives earlier letters from 1997 to 2016.
Why were there so many untitled letters in 2025?
Fifty-two of the 63 are dated 9 September 2025, when FDA announced a crackdown on direct-to-consumer drug advertising. Most were about TV ads.
What do OPDP untitled letters cite most?
Overstated or unsupported efficacy, in 80 of 120 letters. Risk information not presented clearly and conspicuously is second, in 56.
How long does FDA take to close out an untitled letter?
The median is 76 days for letters since March 2021, from 13 days to 280 days. For the September 2025 batch the median is 85 days.