ATMPs: Cell & Gene Therapy Manufacturing
Advanced Therapy Medicinal Products (ATMPs) — cell therapies, gene therapies, and tissue-engineered products — break several assumptions the classical GMP framework was built on, and that is why they have their own dedicated guidance. A batch may be a single patient’s dose; the starting material may be living cells donated from a specific person; the product often cannot be terminally sterilised or fully tested before it must be given; and the manufacturing process itself so shapes the product that "the process is the product" is close to literal. This page covers what makes ATMP manufacturing distinctive and how the quality framework adapts; the sterility environment they demand is the [contamination control](/topics/contamination-control) explainer.
What an explainer is not
A topic explainer is SPEQ’s synthesis of what a practice involves, cited to the standards that govern it. It does not reproduce their text, and it does not determine which of them apply to your product or process.
[ POSITION IN THE FRAMEWORK ]
7 DIMENSIONS · 26 LINKSWhen the batch is one patient, the starting material is a living donation, and the process is the product, classical GMP bends — ATMPs earn a dedicated, risk-based rulebook where chain of identity is a life-safety control.
06 · QUALITY MATURITY — ATMPS: CELL & GENE THERAPY MANUFACTURING, REACTIVE TO ADAPTIVE
Cell/gene manufacture is forced into a classical batch model; single-patient reality is handled by exception each time.
ATMP procedures exist, but comparability and chain of identity are treated as ordinary documentation tasks.
Process control, comparability, and verified chain of identity from donation to administration are established and evidenced.
Risk-based release with rapid methods is engineered around the biology's shelf life; process signals catch drift before failure.
Process understanding drives comparability and control; identity, traceability, and release are designed for personalised manufacture.
SPEQ’s shared five-stage progression, labelled synthesis — not the FDA QMM rating scale. Where does your organization sit? Score your quality system →
07 · REGULATORY & EVIDENCE
GOVERNING STANDARDS · 3
Derived from the 3 standards SPEQ maps to this subject, across 3 regulatory bodies: EMA, ICH, FDA.
RECORDS & OBJECTIVE EVIDENCE
- A qualified master and working cell bank with viral and identity testing
- Comparability data demonstrating post-change product is still the same product
- A documented risk-based release protocol for short-shelf-life product
- Verified chain-of-identity and chain-of-custody records from donation to administration
- Aseptic process simulation and contamination-control records for the cleanroom
COMMON INSPECTION FINDINGS
- Chain of identity not verified end to end for an autologous product
- Comparability not demonstrated after a process or site change
- Release strategy not defined for product administered before sterility testing completes
- Traceability from donor to patient not retained for the required period
- Classical three-batch validation assumed where single-patient manufacture applies
Why advanced therapies needed their own GMP
Classical pharmaceutical GMP assumes, broadly, large batches of a chemically-defined product made from well-characterised materials, testable and releasable before use. ATMPs violate most of those assumptions at once. An autologous cell therapy is manufactured from **one patient’s own cells** and returned to that same patient, so the "batch" is a single dose and there is no remaking it if it fails. The **starting material is a living biological donation** — variable, limited, and irreplaceable — not a specified reagent. And because the cells are alive, the product frequently has a **very short shelf life** and cannot wait for full sterility and quality testing before it must be administered.
These realities are why the EU created a **dedicated GMP guideline for ATMPs** (the Part IV guidance that stands somewhat apart from the standard EU GMP), and why the US regulates human cells and tissue-based products under a distinct framework (**21 CFR Part 1271**) alongside the biologics rules. The point of a separate rulebook is not to relax GMP but to *adapt* it — applying a risk-based interpretation that keeps patient protection intact while accommodating single-patient batches, living starting materials, and products that must be released and given before every conventional test could ever be completed.
The process is the product
For a small-molecule drug, the product is defined by its final specification and the process is the means to it. For many ATMPs the relationship inverts: the manufacturing process — the specific cells, the exact culture and modification steps, the timings — so determines the biological nature of the product that the process effectively *is* the product, and it cannot be fully characterised by end-product testing alone. A change to the process is not a change to how you make the same product; it can be a change to *what the product is*. This is the ATMP form of the principle that quality must be built in, taken to its limit.
The consequence is that process control and consistency carry even more weight than usual, and comparability becomes a central, difficult discipline: demonstrating that product made after a process change (or at a new site, or at larger scale) is still the "same" product is genuinely hard when the product cannot be exhaustively characterised. It also reshapes validation — with single-patient batches you often cannot run the classical three-validation-batch model, so process validation leans on extensive process understanding, control of the critical steps, and approaches suited to one-off manufacture. ICH Q5A (viral safety of biotech products) and the broader biologics framework sit underneath, but the ATMP-specific guidance is what adapts the expectations to living, patient-specific manufacture.
Sterility, release, and the short-shelf-life problem
Because ATMPs are typically living cells that cannot be terminally sterilised, they are made by **aseptic processing** to the same demanding sterile-manufacturing standards as any injectable — environmental monitoring, contamination control, media fills — with the added constraint that the product is biological and fragile. But the defining operational challenge is time: a fresh autologous cell product may have a shelf life measured in hours to a few days, which collides head-on with the classical model where sterility testing alone takes about two weeks and full release testing longer still.
The framework therefore permits and requires adapted approaches: **release under a defined, justified strategy before all conventional testing is complete** — using rapid microbiological methods, in-process controls, and a documented risk-based release protocol — with the outstanding results (such as the final sterility test) reviewed as they become available and a defined action plan if a result later fails after the product has been administered. This "administer before every test is complete" reality is one of the starkest differences from conventional GMP, and it is managed not by lowering the standard but by engineering the control and the risk-based release decision to protect the patient within the time the biology allows.
ATMPs in the quality system
The adaptations do not exempt ATMPs from the quality system — they intensify parts of it. **Chain of identity and chain of custody** become safety-critical in a way they are not for mass-produced drugs: because an autologous product must return to the one patient it was made from, a mix-up is not a quality deviation but a potentially fatal event, so identity tracking from donation through administration is a controlled, verified process. **Traceability** extends from the donor to the patient and must be retained for long periods. And the whole system runs on the same GMP spine — change control, deviation management, CAPA, qualification — interpreted for single-patient, living-product manufacture.
The through-line is that ATMPs are where GMP is stretched hardest by biology and personalisation, and the response has been a dedicated, risk-based adaptation rather than a carve-out. A practitioner moving from conventional manufacturing into advanced therapies has to relearn several instincts — the batch is a patient, the process is the product, release may precede full testing, and identity is a life-safety control — while keeping the underlying quality-system disciplines intact. That combination, familiar disciplines applied under unfamiliar constraints, is what the ATMP-specific framework exists to govern.
FREQUENTLY ASKED
What are ATMPs and why do they need their own GMP framework?
Advanced Therapy Medicinal Products — cell therapies, gene therapies, and tissue-engineered products. They break classical GMP assumptions: a batch may be a single patient’s dose, the starting material is a living donation, the product often cannot be terminally sterilised or fully tested before use, and the process so shapes the product that "the process is the product." The EU created a dedicated ATMP GMP guideline and the US regulates cell/tissue products under 21 CFR Part 1271 to adapt GMP to these realities without relaxing patient protection.
What does "the process is the product" mean for ATMPs?
For many ATMPs the manufacturing process — the specific cells and the exact steps — determines the biological nature of the product so completely that it cannot be fully characterised by end-product testing alone, so the process effectively defines the product. A process change can change what the product is, which makes process control, consistency, and comparability (proving post-change product is still the "same") central and difficult, and reshapes validation away from the classical three-batch model.
How are ATMPs released if they have such short shelf lives?
A fresh autologous cell product may last only hours to a few days, colliding with the ~two weeks a classical sterility test takes. The framework permits a defined, justified release strategy before all conventional testing is complete — using rapid microbiological methods, in-process controls, and a risk-based release protocol — with outstanding results (like final sterility) reviewed as they arrive and a documented action plan if a result later fails after administration. The standard is not lowered; the control and release decision are engineered to protect the patient within the biology’s time limit.
Why is chain of identity so critical for cell therapies?
Because an autologous product must return to the exact patient it was made from, a mix-up is not merely a quality deviation but a potentially fatal event. Chain of identity and chain of custody — verified identity tracking from donation through manufacture to administration — therefore become safety-critical controls, and traceability must extend from donor to patient and be retained for long periods, on top of the usual GMP disciplines interpreted for single-patient, living-product manufacture.