Cell & Gene Therapy Quality·6 min read

Chain of Custody Gaps That Trigger 483s

Autologous CAR-T manufacturing leaves almost no room for a custody break. Here are the patterns that most often turn into an FDA observation — and what a defensible chain of custody program looks like instead.

Why chain of custody carries more weight here

In traditional biologics manufacturing, a batch is one of many, and a documentation gap is a quality event. In autologous CAR-T, the batch is a single patient's own cells, with a fixed, hours-long vein-to-vein window and no possibility of remanufacturing from another source. cGMP traceability requirements do not change based on product type — but the consequence of a custody break does. That is why inspectors scrutinize chain of identity and chain of custody controls in CGT facilities more closely than they would an equivalent gap in a small-molecule process: the same paperwork gap that is a minor deviation elsewhere is a patient-safety issue here.

Common gap patterns
01

Identity verification breaks at handoffs

Apheresis center to courier to manufacturing facility to courier to infusion site: each handoff is a point where patient identity confirmation can be assumed rather than actively verified and documented. A gap here is not just a paperwork issue — it is the exact failure mode chain of identity exists to prevent.

Why it draws scrutiny

Inspectors look for positive, two-person or system-enforced identity verification at every transfer point, not a single check at the start of the process.

02

Manual or swivel-chair tracking across disconnected systems

When custody status is tracked by moving data manually between a courier's system, a manufacturing LIMS, and a hospital's EHR, the handoffs between systems — not the systems themselves — are where records silently diverge.

Why it draws scrutiny

Data integrity expectations apply to the full record, including the transitions between systems, not just each system in isolation.

03

Vein-to-vein timeline deviations that are not systematically trended

A single late shipment or delayed infusion gets resolved case-by-case, but if timeline deviations are not logged and trended across patients, a systemic pattern — a courier route, a site's handling process — never surfaces until an inspector asks for the aggregate view.

Why it draws scrutiny

A deviation system built for occasional biologics events, rather than for a process where timing is itself a critical parameter, tends to under-capture this category.

04

CDMO and courier custody not covered by the sponsor's own oversight

A sponsor's internal procedures can be airtight while the custody handoff to a third-party courier or CDMO is governed only by a quality agreement that has never been audited against real-world practice.

Why it draws scrutiny

The sponsor remains accountable for the full chain regardless of who physically holds custody at a given point.

05

Label and identifier mismatches across systems

Different systems along the chain sometimes use different identifiers for the same patient or product — a barcode here, a name and date of birth there — creating reconciliation gaps that are hard to catch until something does not match.

Why it draws scrutiny

A single, consistent identifier carried through every system and handoff is the control that actually prevents misidentification; reconciling different schemes after the fact is not equivalent.

06

Temperature or handling excursions not linked to a disposition decision

A custody transfer that included a temperature excursion sometimes gets logged as a shipping event without a documented link to the QA disposition decision for that specific patient's product.

Why it draws scrutiny

The excursion record and the release/disposition record need to trace to each other explicitly — a logistics log and a QA decision that live in separate systems, unconnected, is a common finding.

The six patterns above are common industry-recognized risk areas in CGT chain of custody, not citations to specific published 483s on custody failures specifically — FDA does not always make that level of CGT-specific detail public. What is publicly documented is that CAR-T manufacturers do draw real 483s and warning letters for exactly the category of control failure these patterns belong to: in a 2023 inspection, FDA cited a CAR-T manufacturer for inadequate non-viable particulate monitoring during open manufacturing manipulations in its aseptic processing area — a real-time process control gap in the same family as the custody and handoff gaps above, where a documented GMP control was not being executed as required. It is a reminder that these are not theoretical risks; they are the kind of finding FDA actually issues against this product class.

The current CAR-T inspection landscape

Across tracked CAR-T CDMOs with FDA inspection records, outcomes split roughly 17 No Action Indicated, 21 Voluntary Action Indicated, and 3 Official Action Indicated — so most inspections close without major enforcement, but a meaningful share still escalate. That tail is exactly what a defensible chain of custody and aseptic process control program is built to prevent.

What a defensible program looks like instead
  • Positive identity verification, documented at every handoff — not assumed from the prior step.
  • A single system of record (or fully validated integration between systems) rather than manual re-entry across platforms.
  • Timeline and deviation data trended across patients, not just resolved case-by-case.
  • CDMO and courier custody procedures audited directly, not just referenced in a quality agreement.
  • One consistent identifier carried through the entire chain, end to end.
  • Every excursion explicitly linked to the disposition decision it affected.

Is your chain of custody program built for CAR-T's realities?

Quintessential Quality builds and runs internal audit programs and chain of custody/identity systems specifically for CAR-T and autologous cell therapy manufacturers — designed around these exact failure points, not adapted from a small-molecule template.