Watchdog Wednesday: Is FDA’s Manufacturing Guidance A Trojan Horse?

Last week in this space, we looked at what a Form 483 can do to a pending application, and how a single set of inspectional observations can reset a review clock. This week, we turn to the sequel: what happens after the observations land and firms want to respond as swiftly and successfully as possible.

The FDA’s draft guidance on responding to Form 483 observations can be interpreted simply as a tidy how-to guide for firms that want to answer well. But judging by the comments it received, the industry read it rather differently. The recurring worry, voiced most sharply by the biggest names in the docket, is that a document ostensibly about formatting has smuggled in something heavier.

The metaphor practically writes itself, and several commenters reached for it in all but name. A guidance that presents itself as a gift – a template to speed the resolution of a problem – may also be carrying a garrison of new substantive obligations inside its wooden flanks.

The Pharmaceutical Research and Manufacturers of America (PhRMA) warned that the draft “could be interpreted to go beyond the topic of responding to FDA 483 observations and to suggest new substantive current good manufacturing practice (‘CGMP’) expectations.”

For example, “the Draft Guidance contains extensive discussion on the process for investigating manufacturing deviations, some of which is consistent with CGMP, but some of which appears to go beyond current regulatory requirements,” PhRMA says.

If that sounds a little like Inception (an idea planted so deep inside a formatting document that a compliance officer eventually mistakes it for settled law), that is the fear. The Swiss company Lonza located the trap in a single phrase, observing that the instruction to include “at least” a list of elements “may imply that all referenced elements are mandatory components of the FDA Form 483 response,” when in practice the applicability of any given element varies by observation.

The stakes are not academic. A recommendation buried in nonbinding guidance can harden, over a few inspection cycles, into the thing every investigator expects to see.

A Form 483 response is, by regulation, voluntary, and the draft says as much. But the Biotechnology Innovation Organization (BIO) argued that the surrounding language undercuts the promise, warning that a framing in which a weak or absent response raises enforcement concerns “risks creating the impression that responses are effectively mandatory, notwithstanding their characterization as voluntary.”

BIO’s fix is to have the FDA state clearly that “enforcement decisions are based on the totality of inspectional findings and compliance history, and not solely on whether a response is submitted or the extent to which a response conforms to the recommendations in this guidance.” Several trade groups want the same reassurance in different words: reaffirm that a Form 483 response is an opportunity, not a new obligation with a due date.

PhRMA pressed the point into the mechanics. Where the draft recommends assembling a cross-functional “investigation team,” the association noted that a mature quality management system already has one. “A new cross-functional team is not needed if the QMS already has sufficient procedures and processes to investigate deviations,” the association wrote, adding that spinning up a parallel team could itself count as a departure from the firm’s own system. The message to the FDA was to credit what firms already do rather than mandate a second way of doing it.

Troy supplied this column’s opening metaphor; Procrustes supplies the middle one. The draft’s model response format is, in the words of Hyman, Phelps & McNamara, effectively “one size fits all,” and the law firm warned the FDA’s own 483s are anything but uniform. Some observations are simple, others sprawl across dozens of products and multiple sites. Stretching every response onto the same bed (or lopping it to fit) ignores that “some investigations simply cannot be completed within a 15-day timeframe,” particularly where samples must be cultured or additional laboratory analyses run.

The 15-business-day window drew the most consistent fire. BioPhorum, speaking for fourteen member companies, asked FDA to “clearly distinguish expectations for initiation, planning, risk assessment and communication from expectations for completion of investigations” and corrective actions within that window.

Basically, the industry wants the FDA to acknowledge that no one finishes a root-cause investigation in three weeks.

Several firms flagged a Tenet-style asymmetry: the clock runs in only one direction. Firms face a hard 15-day expectation, while the draft says nothing about when the FDA will answer. Becton, Dickinson and Company proposed a symmetrical fix, suggesting the agency commit to “FDA expects to deliver a response to establishments within 30 days after receiving the 483 response,” a timeline it called proportionate to the one imposed on industry.

Two smaller irritations drew outsized attention because they touch cost and confidentiality. The first is the instruction that “all attachments should be signed.” Hyman, Phelps & McNamara called the purpose unclear and the burden real, noting that a response can carry “dozens or even hundreds of attachments,” and that knowingly submitting false statements is already a felony under federal law, which makes a signature on every exhibit “an ineffective use of the limited time” a firm has.

The second is the requirement to identify the response preparer, including any consultant or outside counsel. The Association for Accessible Medicines (AAM) argued that identifying the establishment and its signatory should suffice, and that references to other guidance documents “should remain supportive” and “not an expansion” of what counts as CGMP compliance.

Hyman, Phelps & McNamara went further, warning that forcing consultants to be named could “have a chilling effect on the quality and accessibility of qualified” expert support, since many such engagements run through counsel and are meant to stay privileged.

Writing on behalf of what we can only assume is a client with US manufacturing facilities, Hyman, Phelps & McNamara’s Karla Palmer also raised a question of geographic fairness.

Because the draft says FDA “may evaluate these activities during the next inspection,” and because foreign plants are inspected far less often than domestic ones, the firm warned that relying on the next visit to verify corrective actions “would effectively put foreign establishments at an advantage over domestic manufacturers and could unintentionally incentivize companies to move manufacturing operations overseas.”

That disparity runs counter to the current administration’s onshoring push, Palmer said, suggesting the FDA could remedy the problem by using remote regulatory assessments to verify foreign corrective actions before reinspection rather than waiting years for the next one.

BIO circled the same ground from a different angle, asking the FDA to clarify how the guidance applies to foreign inspections and to those conducted under mutual recognition agreements.

If you will indulge one final metaphor, fixing manufacturing problems can feel like a real Odyssey.

The FDA now has a docket full of stakeholders telling it, in effect, to be clearer about what it wants and humbler about what it can require. For a smaller sponsor, the practical read is to treat the draft’s format as a strong checklist rather than a mandate, protect your consultant and counsel relationships, and do not promise in 15 days what a real investigation needs 15 weeks to prove. May your voyage conclude successfully.

See you tomorrow.