FDA published the final version of its guidance on formal meetings this month, finalising a draft that had been open since September 2023. The guidance is procedural and nonbinding, and most of the coverage of it concerns scheduling. The part that matters to whoever is drafting is section VII, which describes the meeting package as a document with a defined shape, an order of contents, a question limit and a navigation expectation. Read alongside the recommendation that the requester will generally be asked to summarise the discussion at the end of the meeting, it amounts to a specification for a piece of writing that binds a programme for years.
The package is a document, not a bundle
The guidance asks that the package be "organized according to the proposed agenda" and describes it as "a sequentially paginated document with a table of contents with appropriate electronic linkage, appropriate indices, appendices, and cross references." The stated purpose of all of that structure is to "enhance reviewers' navigation across different sections within the package, both in preparation for and during the meeting."
That is a drafting instruction rather than a publishing one. A package assembled as a set of files and stitched together at the end will not have working internal linkage, which is why cross-references belong in the draft rather than the publishing phase. The reviewer is expected to move around this document while the meeting is happening, and that expectation only sharpens as the reviewer picks up AI tooling of their own, since structure is what makes a document tractable to read quickly.
Ten questions, including subquestions
The content list runs to six items, "preferably in the order listed." The fifth is the one that shapes everything else: "a list of the final questions (up to 10 total questions including subquestions) for discussion grouped by FDA discipline and with a brief summary for each question to explain the need or context for the question."
The parenthesis in that sentence is the constraint, because ten total including subquestions limits the usual workaround of one headline question with multiple subparts beneath it, and grouping by discipline means the clinical, CMC and statistical questions each have to stand on their own. Every question also carries a short piece of prose explaining why it is being asked, so the package has to argue its way to each question rather than simply list them, which is the same move as going from a one-line indication to a cited synopsis. Grouping by discipline also means writing to the lens each reviewer actually applies, in the way CDER set out for safety.
The cap sits against a scarcity that is worth stating alongside it, since outside breakthrough therapy and RMAT designations FDA "generally will not grant more than one of each of the Type B meetings for each potential application," which for many programmes means one pre-NDA or pre-BLA meeting and ten questions to spend at it.
You write the history of what the agency already told you
The second content item asks for a background section containing "a brief history of the development program and relevant communications with the FDA before the meeting," along with substantive changes to the development plan and the current status of development.
The sponsor writes that history, and it is the sponsor's own account of what the agency has previously said, submitted to the agency that said it, and it becomes part of the record the review team reads next time. The TUDRIQEV approval earlier this month is the long version of why that matters, since the complete response letter quoted a March 2021 meeting minute back to the sponsor and observed that the agency's advice had not changed in the intervening five years. A background section that summarises prior advice selectively is legible as such to anyone holding the original minutes.
What the data section may not be
The sixth item is data organised by discipline and by question, and the guidance is direct about what does not belong: "protocols, full study reports, or detailed data generally are not appropriate for meeting packages." What is wanted instead is summarised material that describes results "with some degree of quantification."
Two sentences in that passage carry more weight than their placement suggests. The first is that "the trial endpoints should be stated, as should whether endpoints were altered, or analyses were changed during the course of the trial." That is a disclosure expectation sitting inside a summary document, and it is the same territory as an analysis population that moves late and has to be reconciled everywhere downstream. The second is that "merely describing a result as significant does not provide the review division with enough information to give the most constructive advice or identify important problems the requester may have missed." Both are instructions about how to write a results paragraph.
Length is a failure mode
The guidance sets out when a meeting can be rescheduled or cancelled, and the package appears in three of those provisions. A meeting can be rescheduled where "the review team determines that the meeting package is inadequate," and also where "there is insufficient time to review the material because the meeting package is voluminous," which the guidance is careful to note can happen "despite submission within the specified time frames and the appropriateness of the content." A meeting can be cancelled where the package is "grossly inadequate," and a request can be denied where the package "does not provide an adequate basis for the meeting discussion."
Volume is therefore its own risk, independent of whether the content is right and whether it arrived on time. Against that, the guidance is explicit that denials "will be based on a substantive reason, not merely on the absence of a minor element." The thing to avoid is not an imperfect package but an unreadable one. A meeting can also be denied for being "premature for the stage of drug or biological product development," and the guidance gives the example of an end-of-phase 2 request that concedes the phase 2 efficacy and safety data will not be in the briefing document, which is phase-appropriate expectations applied to the meeting rather than the dossier.
Three days to rewrite your own agenda
FDA sends preliminary responses no later than five calendar days before Type B (end-of-phase), Type C, Type D and INTERACT meetings, and no later than two calendar days before Type A and other Type B meetings. For the first group the requester then has three calendar days to say whether the meeting is still needed and, if it is, to send "a revised meeting agenda indicating which questions the requestor considers as resolved and which questions the requestor will want to further discuss."
That revised agenda is a drafting task on a three-day clock, performed against fresh written positions from the agency, and the guidance notes it "could alter the core attendees from the FDA." Deciding which questions are now closed is a judgement about the programme, recorded in a document, and it belongs to the same category as review windows that leave no slack for a slow internal turnaround. Adding new material at this point does not work, since preliminary responses "are not intended to generate the submission of new information or new questions," and new data may simply go without comment.
The summary you give in the room seeds the minutes
The recommendation in section X on summing up is easy to read past. "Generally, the requester will be asked to present the summary of the discussion, advice, and/or agreements to ensure that there is mutual understanding of meeting outcomes and action items." FDA staff can then add or clarify points, "and these items can be added to the meeting minutes."
Recording is prohibited, so this spoken summary, composed live, is the sponsor’s first chance to shape the official record, which FDA then writes and issues. At pre-NDA and pre-BLA meetings run under the Program, the requester and FDA should also summarise "agreements regarding the content of a complete application" and any agreement on delayed submission of minor components, which means the contents of the eventual dossier can begin to harden in this same summary.
Thirty days, then twenty
FDA issues finalised minutes within 30 calendar days. They are written in bulleted form from established templates, they cover agreements, disagreements, issues for further discussion and action items, and they are "not intended to represent a transcript." The guidance also notes that FDA may put things in the minutes that were never said in the meeting, giving pediatric requirements, data standards and abuse liability potential as examples, and that the minutes will distinguish this material from the discussion itself.
Two written routes follow from that, both of them narrow. On a significant difference in understanding, the requester raises it with the project manager and, if it persists, submits "a description of the specific disagreements" to the application, or by letter to the division director where there is no application. The minutes then either stand as the official documentation or receive an addendum, and that addendum "will also document any remaining requester objections." Separately, clarifying questions may be sent "as a Request for Clarification to the FDA within 20 calendar days following receipt of the meeting minutes," confined to confirming advice already given rather than raising new issues, with a written response due within 20 calendar days that references the original minutes. Both routes run in writing and both expire, which puts the accuracy of a document the sponsor did not draft on a clock the sponsor has to watch.
What this means for the writing
The meeting package is usually treated as correspondence, drafted quickly by whoever has the clinical detail, and filed once the meeting is over. This guidance describes something functionally closer to the first section of the submission. It has a content order, a navigation requirement, a question budget, a disclosure expectation about changed endpoints and analyses, and a length beyond which it stops being read properly. The prose in it draws the advice that the programme is then measured against, which is why a pre-IND package written before the modality is settled is a harder writing problem than it appears, and why manufacturing questions can warrant their own meeting rather than a shared one.
What follows the meeting is also writing, on tight clocks: a revised agenda in three days, a disagreement filed against minutes that otherwise stand, a Request for Clarification inside twenty. None of that is background material sitting behind the programme, and the window in which any of it can still be corrected is measured in days rather than review cycles. The TUDRIQEV file shows an objection recorded in 2021 still doing work in 2026, and five years and two rejections is the interval over which these documents are read back. Post-action meetings sit in this same guidance, as Type A within three months of a rejection and Type B after that, which is the point at which programmes like Outlook’s fourth cycle, Atara and Pierre Fabre’s third and Regenxbio’s resubmission without a new study have to decide what to ask next.
One caution on reading this as news
This is a final guidance that closes out a September 2023 draft, and, as with CBER's revision of SOPP 8401, there is no published redline. Much of what is described here, including the thirty-day minutes commitment and the twenty-day clarification window, has roots in the 2023 draft and in the 2017 guidance it replaced. The useful question is not what changed this month but what the current document asks of the writer, which it sets out at more length than most procedural guidances bother with. The meeting types and goal dates in it were negotiated under user fee commitments, so the next PDUFA cycle is where any further change to them will surface first. It sits in the same category as the other procedural documents a submission has to survive, where the source of truth is the record rather than the document and provenance is defined by whoever wrote it down.
The guidance is at fda.gov/media/172311/download, under docket FDA-2017-D-6530.