FROM THE EDITOR
At the end of a previous issue, I said I would cover the unsolicited-request defense. The facts went somewhere else, and I think somewhere more useful.
On October 27, FDA posted an August 12 warning letter to Amgen about a promotional video for Vectibix, its colorectal cancer drug. The letter says the video makes convenience claims and “entirely omits risk information,” even though the drug carries a boxed warning. It is addressed to Amgen’s Chairman and CEO and arrived in a year when OPDP has been sending untitled letters at a steady clip and warning letters almost never.
Amgen’s public answer was not that FDA had read the video wrong. It was that FDA had been sent the wrong video. A spokesperson said an incorrect file “was inadvertently submitted to FDA that did not contain the required risk information.” Amgen added that the file was never publicly disseminated, the version actually used included appropriate safety information, and the correct file has since been submitted.
I have no way to test that explanation, and this issue assumes it is true. That is exactly why it matters. If Amgen is right, one of the largest biotech companies in the world received a public, CEO-addressed warning letter for misbranding, and the cause was the wrong file reaching FDA. That is a process failure in the department CLOs own, not a judgment call about promotional claims.
Sources: FDA Warning Letter to Amgen Inc., MARCS-CMS 739641 (Aug. 12, 2026; posted Oct. 27, 2026); Reuters and Becker’s Hospital Review reporting (Oct. 27–28, 2026); Covington & Burling and The FDA Law Blog on 2026 OPDP enforcement. I have not seen Amgen’s response to FDA, and nothing here is a finding about what happened inside Amgen. Nothing here is legal advice.
— John LaRocca
THE LETTER IN 60 SECONDS
What FDA reviewed. A Vectibix (panitumumab) video, material code USA-954-81023, that Amgen had submitted to FDA on Form FDA 2253.
What the video said. Under the header “For adult patients with wild-type RAS mCRC,” it showed “60 minutes or less of infusion time,” “No premedication” next to a crossed-out IV bag, and colon imagery.
What it left out. According to FDA, all of the following:
Risk information. The video had none, including nothing from the boxed warning for dermatologic toxicity or the warnings for infusion reactions and pulmonary fibrosis/interstitial lung disease.
The approved regimens. These are first-line with FOLFOX, or monotherapy after progression on fluoropyrimidine-, oxaliplatin- and irinotecan-containing chemotherapy.
The limitations of use. Vectibix is not indicated for RAS-mutant mCRC (except with sotorasib in KRAS G12C-mutated disease) or where RAS status is unknown.
The legal theory. Misbranding under FD&C Act sections 502(a) and 502(n) (21 U.S.C. 352(a), (n)) and 21 CFR 202.1(e)(3)(ii)–(iii) and (e)(5). The prohibited act is introducing a misbranded drug into interstate commerce under section 301(a).
What FDA asked for. Within 15 working days:
stop the violative communications;
list every other Vectibix promotional piece with similar representations, with its 2253 submission date;
provide a plan to discontinue them;
provide a plan for corrective communications through the same media, at the same frequency and duration, to the same audiences.
Alternatively, Amgen could explain why it believes it is not in violation.
Who signed and who received it. It was signed by the Director of OPDP’s Division of Advertising & Promotion Review 1 and addressed to Robert Bradway, Chairman and CEO.
Timing. The letter was dated August 12 and posted October 27. Amgen’s “wrong file” statement came after the posting.
WHY THIS IS A WARNING LETTER AND NOT AN UNTITLED LETTER
OPDP’s 2026 record makes this letter stand out. Covington’s quarterly tracker counted 10 OPDP untitled letters and no warning letters in the second quarter. By late July, The FDA Law Blog had counted 21 untitled letters for the year. Those letters cluster around a familiar set of themes:
the full indication buried in fine print;
minimized risk;
convenience claims (a July letter objected to a TOBI Podhaler video suggesting use “in the car”);
video visuals that pull attention from the major statement.
Vectibix fits the convenience theme exactly. Infusion time and the absence of premedication are classic convenience claims. What moved this matter up a level is the claim FDA could make about the file it had: the video contained no risk information at all for a drug with a boxed warning, and FDA called those risks potentially fatal. An untitled letter says “this is misleading.” A warning letter says “this is serious enough that we may go to court,” and this one says so: failure to correct “may result in legal action without further notice,” including seizure and injunction.
On the face of the file FDA reviewed, that escalation is defensible. A promotional video with zero risk information for a boxed-warning oncology product is about as clean a violation as OPDP sees. The problem, if Amgen is right, is that the file was not the promotion.
One point to put in perspective: OPDP warning letters are routinely addressed to the company’s CEO. The addressee here is not, by itself, a signal of personal-liability theory the way the CEO-addressed Empower letter was in Issue #13. What it does mean is that the CEO’s name sits on a public FDA document alleging the company misled physicians about a cancer drug’s safety. The CEO will ask the CLO how that happened.
WHAT FORM 2253 ACTUALLY IS
Most CLOs know the 2253 as a regulatory-operations task: the form that goes with promotional materials to FDA. It is more than that, and this letter shows why.
For a biologic, 21 CFR 601.12(f)(4) requires the applicant to submit specimens of advertising and promotional labeling at the time of initial dissemination, accompanied by Form FDA 2253 (the drug analogue is 21 CFR 314.81(b)(3)(i)). The submission is not a request for review and is not pre-clearance. It is the company’s own signed record to FDA of what it put into the market.
That has three consequences.
FDA reviews the submission, not the market. OPDP does not usually go looking for your video on a conference booth monitor. It reviews what you filed, along with complaints and surveillance. When the file and the field version differ, FDA will act on the file, because the file is the company’s representation of what it disseminated.
The submission is the evidence. The letter does not say FDA saw the video in use. It cites the 2253-submitted material by its code. Under the Act, the violation is introducing a misbranded drug into commerce. When FDA’s only evidence of what was introduced is what the company told FDA it introduced, the company has handed FDA its case.
The company’s own certification is the obstacle to the defense. “The file we submitted is not the file we used” is a factual defense. But it begins with an admission that the company’s signed regulatory submission was inaccurate. That is not a comfortable posture with the agency that regulates every product you sell.
THE LIMITS OF THE WRONG-FILE DEFENSE
Assume Amgen’s explanation is entirely correct. It still answers only part of the letter.
It answers the risk-omission charge. If the disseminated video carried important safety information, FDA’s central concern was not present in the market.
It may not answer the indication charge. FDA also faulted the video for omitting the approved regimens and the limitations of use around RAS status. Amgen’s public statement addressed risk information. It did not say whether the disseminated version presented the full indication and the RAS limitations. The headline “For adult patients with wild-type RAS mCRC” may well appear in both versions. If the field version shared the same benefit presentation and differed only by adding safety information, part of FDA’s objection may survive. The CLO’s first question in any wrong-file situation is: what exactly is the difference between the two files, and which of FDA’s charges does that difference actually answer?
It does not stop the posting. FDA dated the letter August 12 and posted it October 27. The public record does not show what Amgen told FDA in its 15-working-day response, or whether FDA has evaluated it. What the record does show is that the public learned of the “wrong file” explanation from a spokesperson after FDA’s version was already online. The company’s account and the regulator’s account reached investors in the order the regulator chose. That is the dynamic I previously described under the FDA–SEC information-sharing MOU, playing out in a promotional case.
It complicates the corrective-communication request. FDA asked for corrective messages through the same media, to the same audiences, for the same duration as the original. If the violative file was never disseminated, there is no audience to correct. That is a sound argument, but it has to be made to FDA with evidence. That evidence is the dissemination record, the approved-final asset, the distribution logs, and the submission audit trail. If the company cannot produce that package quickly, the argument is only an assertion.
HOW A WRONG FILE GETS FILED
I don’t know what happened at Amgen. In my experience, the ways a wrong file reaches FDA are few and predictable, and every one of them is a process gap rather than a judgment call:
One material code on several versions. A job code like USA-954-81023 is assigned when the piece enters review and stays on the draft, the post-comment revision and the approved final. Whoever pulls the file for submission pulls by code and gets the wrong version.
Agency handoff. The creative agency renders the final cut with the ISI end-card or super, and also delivers a “clean” cut without it for other edits. Both carry the same file name.
Video-specific ISI. For video, important safety information often lives in a separate track, end slate or scrolling super added late. The approved version has it; the earlier render used for timing or review does not.
Submission compiled by someone outside review. Regulatory operations or a submissions vendor assembles the eCTD package from a shared drive. Nobody who sat in the MLR (medical-legal-regulatory) review checks the file that actually goes out.
No binding between “approved” and “submitted.” The review system shows the piece as approved. The submission system shows a file was sent. Nothing confirms they are the same bytes.
Each of these has a simple fix. Most companies have not made it, because until a letter like this one, the 2253 felt like paperwork.
BOX: BEFORE THE 2253 GOES OUT
Six controls for the promotional review and submission process
One version, one identifier. The approved final gets its own identifier or version suffix, distinct from every draft. Submission pulls by that identifier.
Lock the approved file. Capture the approved asset’s file hash (a digital fingerprint, such as SHA-256) at MLR sign-off. The submission compiler checks that hash before filing. If it does not match, the submission stops.
Two-person check on video and audio. For any non-print piece, someone who participated in MLR review views the exact submission file end to end, including the ISI track or super. That person attests to it in writing.
Agency deliverable rules. Agency contracts require a single labeled “APPROVED FINAL” rendition per material code, with no unlabeled alternates. Delivery logs identify which file is the dissemination master.
Dissemination log. The commercial team records which file was used where (booth, field force, website, congress, paid media) and when. If FDA ever cites a file you did not use, this log is your defense.
Periodic reconciliation. Each quarter, sample 2253 submissions and compare the filed asset against the disseminated asset. Treat any mismatch as a deviation with a root-cause investigation.
THE 15-WORKING-DAY RESPONSE
When an OPDP warning letter arrives, the CLO has three weeks to make decisions that will shape the public record for years. Five questions to settle in the first 48 hours:
Is FDA’s description of the material accurate? Pull the 2253 package FDA cites and the disseminated version. Compare them frame by frame before anyone drafts a word.
Which charges does each fact answer? Map every FDA objection (risk omission, indication, limitations of use, claims) against each version. Do not let a strong answer to one charge carry the others.
Do we comply, contest or both? The letter offers a route for a company that believes it is not in violation. Using that route requires evidence: a dissemination log, the approved asset, distribution data and the submission audit trail. Decide early whether you can build that package in 15 working days.
What is the “similar materials” universe? FDA asks for every promotional piece with similar representations and its 2253 submission date. That inventory is an admission if it is over-inclusive and a credibility problem if it is under-inclusive. Assign one owner.
Who needs to know before posting? Warning letters are posted, and you will not control when. Investor relations, corporate communications, the disclosure committee and the board’s compliance or audit chair should know the facts and the company’s position before the letter appears on FDA’s site, not after.
THREE WORKSTREAMS
IMMEDIATE — Reconcile Your 2253 Record
Pull a sample of 2253 submissions from the last 12 to 24 months, weighted toward video, audio and digital assets for boxed-warning products. Compare each submitted file to the disseminated version. Document the result, whether matches or mismatches, under privilege where appropriate.
BUSINESS IMPACT: If a mismatch exists, you want to find it before OPDP does. A self-identified and corrected submission error is a quality event. An FDA-identified one is a public warning letter with your CEO’s name on it.
NEAR-TERM — Close the Gap Between “Approved” and “Submitted”
Implement the six controls in the box. The two that matter most are the file-hash lock at MLR sign-off and the two-person view-through for non-print assets. Amend agency statements of work so that delivery of an unlabeled alternate render is a breach. Assign one owner for the submission step, not one owner per handoff.
BUSINESS IMPACT: The fix costs a few hours per material. The failure costs a warning letter, a corrective-messaging campaign, a response to FDA prepared on a 15-working-day clock, and a news cycle in which the company explains that its own submission was wrong.
STRATEGIC — Write the OPDP Response Playbook Before You Need It
Prepare a standing protocol for any OPDP letter. It should cover:
the 48-hour fact pull;
the charge-by-charge mapping;
the decision on whether to comply, contest or both;
the similar-materials inventory method;
the corrective-communication argument when material was never disseminated;
the pre-posting notification list (IR, communications, disclosure committee, board).
Rehearse it once on a past untitled letter.
BUSINESS IMPACT: OPDP is sending letters at least twice a month in 2026. The question is no longer whether your company receives one. It is whether your first 48 hours are spent finding files or making decisions.
“The video creates a misleading impression about the drug’s FDA-approved indication and use.”
— FDA Office of Prescription Drug Promotion, Warning Letter to Amgen Inc. (Aug. 12, 2026)
I’ve compiled a checklist to manage FDA submissions and their aftermath. https://laroccaadvisory.gumroad.com/l/ouimkw
COMING NEXT
On November 10, Weil, Gotshal & Manges partners voted 97% to dissolve the firm. For in-house legal departments with live matters there, the question is no longer which firm is winning the lateral market. It is where their files are going, who controls privilege during the transition, and what happens to a protective order that names lawyers who now work somewhere else. Issue #19 is the CLO’s playbook for when your law firm shuts down.
ABOUT THE AUTHOR
John LaRocca is a New York and New Jersey-barred attorney with 37 years of legal experience, including approximately 26 years as GC or CLO at specialty pharma and biotech companies including Actavis, Eagle Pharmaceuticals, Aerie Pharmaceuticals, and Y-mAbs Therapeutics. His transactions include the Aerie/Alcon acquisition (~$930M) and Y-mAbs/SERB acquisition ($412M, 105% premium). He runs LaRocca Advisory Services, a fractional CLO practice for PE-backed specialty pharma and commercial-stage biotech. Subs



John, you raise an important distinction between a failure in regulatory judgment and a failure in the integrity of the submission process. It highlights how an administrative error can create significant regulatory and reputational consequences.
One question: If a company can conclusively demonstrate that the promotional material actually disseminated was fully compliant, should the FDA distinguish between a documentation or submission-control failure and an actual patient-safety or misbranding violation when determining the appropriate enforcement action?