The file the science produced
Picture a twelve-person biotech, two years and roughly twenty million dollars into its first program. The lead molecule works. The tox package is clean, the potency signal is real, and the team is booking its pre-IND meeting with the confidence of people who know their science cold. Then the regulatory consultant brought in for a pre-submission read asks one plain question: where's the tech transfer report comparing the material your CDMO makes now to the material your tox studies ran on?
The room goes quiet. The comparison did happen — in someone's head, across a few Slack threads and a shared spreadsheet, back when the batches actually matched. Nobody wrote the report FDA reads.
The science isn't the problem here. The file the science produced is. And it's the same short list of gaps almost every time. This post walks the six recurring quality-system gaps that turn a strong first IND into a clinical hold, why they keep showing up, and where each one gets closed cheaply — before the pre-IND meeting, not after a hold letter.
Why the same gaps keep showing up
Every quarter, somewhere in early-stage biotech, a team discovers that the quality file FDA wants isn't the quality file the team built. The science is sound and the data is real. The records, read against the CMC standards FDA applies to a first-in-human filing, have specific holes the team never saw — because the team was building the science, not the file.
The pattern repeats because science skills and quality-file skills are different skills. A first-IND quality program is a recordkeeping discipline running in parallel to the science, capturing evidence in the structure FDA reads. When that discipline isn't there from week one, the file at submission has gaps no last-minute consultant can backfill. The clinical hold that follows is rarely about the science. It's about the file the science produced.
The scale is well documented. Between 2020 and 2024, 74% of FDA Complete Response Letters cited CMC or quality deficiencies, and roughly 40% of IND submissions are delayed for CMC reasons. The same failure modes show up in FDA's published clinical-hold analysis on oncology INDs, in the 291 Complete Response Letters analyzed by Syner-G, and in the consulting industry's running tally of why pre-IND meetings turn into hold letters. None of it is proprietary insight. It's six well-known gaps any early-stage biotech can close if it knows where to look.
Here's the shape of all six at a glance:
| # | The gap | What FDA reads | Where it's usually missing |
|---|---|---|---|
| 1 | Tech transfer discontinuity | Comparison of parameters between sending and receiving sites, plus bridging data | A CDMO memo and a batch record, no explicit before/after comparison |
| 2 | Methods not phase-appropriate | Demonstrated specificity, linearity, accuracy, precision across the spec range | Research-bench methods used as-is for release |
| 3 | Stability thin in the wrong places | In-use and shipping stability, not just long-term storage | Long-term data only; transport and clinic conditions untested |
| 4 | Process control on paper only | Deviations, change controls, and training tied to the batch records | SOPs and batch records that exist but aren't connected |
| 5 | Preclinical material ≠ clinical material | A comparability protocol bridging tox and clinical batches | Different scale or process with no bridging argument |
| 6 | Aspirational release specs | Ranges justified by multi-batch data | Specs set too wide or too tight, unsupported by data |
Now each one in turn.
Gap 1: Tech transfer discontinuity between clinical and commercial sites
The single most common CMC clinical-hold cause is a process parameter, raw material, or piece of equipment that differs between the site that made the preclinical material and the site that will make the clinical material. Teams under cost pressure routinely move from a research-grade contract manufacturer to a clinical-GMP CDMO in the year before filing. The move is necessary. The record of what changed in the move is often missing.
What FDA looks for is a tech transfer report that names every parameter that changed, assesses the impact on the product, presents bridging data demonstrating equivalent quality, and explains why the changes don't invalidate the preclinical safety data. What sponsors routinely produce instead is a CDMO selection memo, an executed services agreement, and a clinical-grade batch record with no explicit comparison to what came before.
Closing it is straightforward, but it has to start the day the CDMO contract is signed: a controlled tech transfer protocol, a comparison matrix of parameters across sending and receiving sites, batch-by-batch quality data after transfer, and an explicit decision on whether bridging studies are needed. Those records have to live in the QMS by the pre-IND meeting, not get reconstructed for the submission.
Gap 2: Analytical methods that aren't qualified for phase-appropriate use
Early-stage methods are often inherited from the research lab and used as-is to release clinical material. FDA's Phase 1 expectation is that methods are "fit for intended use" rather than fully validated — but fit-for-purpose still means demonstrated specificity, linearity, accuracy, precision, and a quantitation range that actually covers the specification. Methods that behave perfectly on the development bench can fail any of those when challenged against the FDA standard.
The usual culprit is the potency assay. For a biologic, a potency assay that hasn't shown reproducibility across multiple analysts, multiple days, and the full specification range is close to a standalone clinical-hold trigger. FDA's guidance on potency assurance for cellular and gene therapy products exists partly because the same potency gap kept appearing in filing after filing.
The fix is a method qualification protocol per assay, executed at a sample size and design appropriate to the phase, with data captured under change-controlled procedures. The protocol-and-report pair lives in the QMS, and the reviewer comparison happens during pre-IND prep — not against the submission deadline.
Gap 3: Stability data thin in the wrong places
Stability data for a first-in-human filing has to support the manufacturing date, the labeled storage condition, in-use stability across the planned trial duration, and — the one teams miss most — stability during shipment and at the clinical sites. A filing that has long-term storage data but nothing on in-use or shipping stability is filing with a hole FDA will flag.
"Stability during transport" sits on FDA's CMC clinical-hold cause list for a reason. Drug product moving between manufacturing site, clinical site, and pharmacy under variable temperature is the most common late-stage scope expansion sponsors overlook. Three records have to exist as controlled documents, and usually don't:
- Shipper qualification — evidence the container holds the labeled condition across the real shipping lane.
- Temperature-monitoring qualification — proof the monitoring itself is trustworthy.
- Condition-to-quality data — the link between what happened in transit and product quality on arrival.
The discipline that closes this treats stability as a quality-system program from day one, not a one-time study run for the IND. Protocols, pulls, and reports live as controlled documents with timestamped data, and any departure from the stability protocol goes through the deviation workflow rather than getting quietly absorbed into the next study design.
Gap 4: Process control documented but not consistently followed
This one rarely looks like a documentation gap. The SOPs exist. The batch records exist. What's missing is the connective tissue between them: deviations recorded against the SOPs, change controls run against the process, and training records proving the operators who executed a batch were trained on the current SOP version.
A reviewer reading a first-time sponsor's submission is looking for evidence the process is actually under control, not just that documents exist. A few things read as findings fast:
- a batch record signed by an operator with no training record for the SOP version in force;
- a change to the process that never went through change control;
- a deviation that went unlogged because nobody realized it mattered — the worst of the three, because it surfaces during the inspection that follows.
Closing it takes a working deviation workflow, a working change-control workflow, and a training program the QA function genuinely runs. None of that is exotic; it's QMS fundamentals. The gap appears because in early-stage biotech the QA function is often one person sitting beside scientists who don't think in QMS terms — so the fundamentals run slow. That's the same architecture problem we unpack in why slow CAPAs are usually a system problem, not a team problem.
Gap 5: Preclinical material that doesn't represent the clinical material
The bridge between preclinical safety data and clinical-grade material is one of the most scrutinized areas in any first IND. FDA's question is blunt: is the safety data generated on the preclinical material predictive for the clinical material? Answering it requires demonstrated comparability across the critical quality attributes.
The gap shows up most when the manufacturing process changed materially between tox material and clinical material — different scale, different process, different raw materials, different facility — with no comparability protocol bridging the two. Without the bridging data, the IND is asking FDA to accept that two different products produced the same safety signal. FDA can accept that with a documented rationale, or use it as grounds to hold the IND pending more data.
The pre-IND meeting is the right place to surface the comparability strategy. A sponsor who walks in with a defined comparability protocol, the critical quality attributes named, and a bridging-data plan articulated is asking FDA the right questions on the right timeline. A sponsor who walks in without one is signaling that comparability will become a hold candidate later.
Gap 6: Release specifications that are aspirational rather than evidence-based
Release specifications are the numerical thresholds clinical-grade material has to meet before it's released for human use. For a first IND, FDA expects specs tight enough to ensure safety and consistency, supported by the data generated so far, and phase-appropriate — tightening as the program matures.
The recurring failure runs in two directions:
- Too wide — specs that would pass essentially any result the methods produce. This invites the reviewer's question of how the specification ensures product quality at all.
- Too tight — ranges the actual process can't reliably hit. This produces a deviation rate that turns every batch into an investigation.
Evidence-based specifications come from multiple representative batches with statistical analysis behind the acceptable range. Treating specs as a quality-system commitment — subject to change control, justified by data, reviewed periodically — is what FDA expects. Treating them as a project deliverable produced once and never revisited is the gap.
The pre-IND meeting as the cheapest forcing function
Pre-IND meetings are PDUFA Type B meetings with a defined agenda: review of preclinical data, review of the proposed study design, and review of the product manufacturing and quality controls needed to start human studies. That third item is where the six gaps either come up early or come up too late.
A sponsor who uses the meeting to confirm the closure status of each gap gets FDA feedback while it's still actionable. A sponsor who spends the meeting on the science and skips the CMC items gets the same questions later — after submission, under clinical-hold pressure. The two paths land in very different places.
The arithmetic is unforgiving. A clinical hold on a first IND extends first-dose timing six to eighteen months. The burn in that window — payroll, CMO retainers, lease, regulatory and consulting fees — runs roughly $1.5M to $15M depending on team size and program complexity. Dilution from a down-round triggered by the delay can dwarf the burn itself. Closing the six gaps before the pre-IND meeting costs a fraction of any of those numbers.
How Complere supports a first-IND quality program
The six gaps all come down to the same thing: the records FDA reads have to exist, stay controlled, and be retrievable together. Complere is where those records live from week one, so the file the science produces is defensible by the time you sit down at the pre-IND meeting.
- Your protocols stay current, not quietly stale. The SOPs and the tech-transfer, comparability, stability, and method protocols the gaps hinge on stay under version control with a real sign-off, and your team schedules each one for review, so the protocol that's due doesn't drift into an uncontrolled draft.
- A change carries its own paper trail. When a tech transfer or a process change lands, it routes through named approvers with a written judgement of what it affects, so the comparison and the bridging decision are captured as they happen instead of reconstructed a year later.
- A departure never goes looking for a form. When something steps outside a protocol, it's logged and routed on the spot, and where it needs a corrective action it stays tied to it through to the check that it worked.
- Training and the procedure version stay married. When an SOP changes, the training on the new version is assigned and its completion tracked, so "was this person trained on the current version?" is answered by the record, not by memory.
- The judgement calls are on the record, not in the hallway. Where comparability or a specification range turns on judgement, the reasoning is written down and scored, so the rationale is retrievable rather than remembered.
- Nothing changes without leaving a mark. Every one of those records keeps its own history: who did what, when, and what it was before, so a change becomes part of the record. Your file sits in a space of your own, apart from every other customer's.
The quality judgment stays yours: the comparability strategy, the specification rationale, the method design, the effectiveness criteria. What Complere gives it is a connected place to live, so the recordkeeping discipline is part of the workflow rather than a separate job someone is supposed to do later. See how that discipline maps onto a formal validation approach and stays audit-ready from your first pre-IND meeting through your first dose.
