Batch Release Decision for a Seeded Cartilage Construct
A university-hospital collaboration produces a porous collagen-PLGA scaffold seeded with autologous mesenchymal stem cells for a phase I/IIa study in focal cartilage defects, with 12 patients enrolled. Production batch B-2412 yields eight constructs. Release testing shows: endotoxin 0.35 EU/mL against a limit of 0.5 EU/mL; sterility with no growth at 14 days; residual glutaraldehyde 8 ppm against an internal limit of 5 ppm; ISO 10993-5 cytotoxicity showing 68% cell viability at 24 h against an acceptance criterion of greater than 70%; and seeded cell viability of 88%. The unseeded scaffold is CE-marked as a device, but the seeded construct has no marketing authorisation; the hospital ethics committee approved a hospital-exemption style pathway. The first patient is scheduled in 10 days. The laboratory director argues the cytotoxicity result is borderline and that because the cells are autologous, the risk is low.
1. Should batch B-2412 be released? Work through each test result and the reasoning.
Read the worked answer
No. The batch must not be released, and the reasoning is about pre-specified criteria, not about how convincing the director's argument sounds.
- Cytotoxicity, 68% versus greater than 70%. The criterion is a defined pass/fail threshold, generally applied as a viability cut-off in ISO 10993-5. A result below it is a failing result. Borderline is not a category in a specification; either the specification already allows for assay variability (it does not appear to) or the result is Out of Specification.
- Residual glutaraldehyde, 8 ppm versus 5 ppm. This is a hard specification failure, and it is plausibly the cause of the cytotoxicity result. Two failures that could share a root cause should not be assessed in isolation.
- Endotoxin, 0.35 versus 0.5 EU/mL. Passes, but with modest margin; for an intra-articular implant the local inflammatory response matters, and a borderline endotoxin result alongside a failed cytotoxicity test reduces confidence in the washing and purification steps.
- Sterility, no growth at 14 days. A negative sterility test is not a sterility assurance level. For a product implanted within days, release should rely on a validated aseptic process and, where applicable, parametric release. Ask whether the 14-day incubation has even completed before the scheduled implant date - if not, the result is not yet a result.
- Cell viability 88%. Good, but cell viability does not offset a chemical residue failure.
- The autologous argument. Autologous cells remove the risk of allogeneic immune rejection, not the risk from residual crosslinker, particulate, endotoxin or a failing process. An autologous product that injures the patient is just as reportable.
- Decision. Reject the batch, open a formal deviation, and do not release 'with a note'. Release criteria written in advance exist precisely to remove this negotiation from the pressure of a scheduled patient.
2. What regulatory and professional pathways apply, and what must the ethics committee and the patient be told?
Read the worked answer
The product sits between two regimes, and the gap must be managed openly.
- Classification. A scaffold plus autologous cells is a combined product. The scaffold alone is a CE-marked device; the seeded construct is not. The hospital must state, in writing, the basis on which it is being supplied - here, an ethics-committee-approved hospital-exemption style pathway - and what would trigger a change in that basis (for example, a different cell type, a commercial supply route, or wider use).
- Quality system. ISO 13485 alone is not sufficient for a cell-based product; good tissue practice and GMP expectations for aseptic cell handling, environmental monitoring, donor and batch traceability and cleanroom classification apply. The QMS must cover the whole chain, not only the scaffold manufacturing.
- MDACS. The device component should be listed, and the hospital should clarify with the Medical Device Division how a combined product under an exemption route should be recorded. Do not assume the device listing covers the seeded construct.
- Ethics committee. Notify as a protocol deviation and a batch failure before the scheduled date. The committee approved a trial with defined release criteria; silently substituting a patient because the schedule is tight is a serious breach. Provide the failed results, the root cause investigation status and the plan.
- The patient. Disclose that the batch did not meet release criteria, that the implant has been postponed, and what the alternative timelines are. Do not present it as a minor quality issue. Consent for an investigational product must be informed about manufacturing holds.
- Documentation. Keep the release decision, the deviation, the ethics correspondence and the patient communication in the trial master file.
3. If the batch is rejected, how do you manage the supply gap and prevent recurrence?
Read the worked answer
Treat it as a process capability problem, not a single bad batch.
- Supply gap. Immediately assess whether any previously released batch is available and whether the patient can safely wait. Re-sequence the surgical list, tell the patient and the referring surgeon early, and record the clinical impact. Do not bring forward a less mature batch to fill the slot.
- Root cause. Investigate the washing steps after crosslinking, the glutaraldehyde concentration and contact time, the number and volume of washes, the collagen lot and its variability, the seeding density and the assay itself. Check whether the cytotoxicity result is genuine by reviewing the positive and negative controls and the cell line passage; a failed assay run would change the investigation but not permit release without a valid retest protocol.
- Corrective action. Validate an improved washing regime with residue testing at each step, and demonstrate removal of glutaraldehyde to below the internal limit across three consecutive batches before resuming release. Consider replacing glutaraldehyde with a less cytotoxic crosslinker if the process cannot be controlled.
- Strengthen release. Build process capability data so that limits are based on what the process reliably achieves; set an OOS procedure that prohibits repeat-until-pass; require cytotoxicity and residue results to be available and reviewed before scheduling the patient.
- Supplier and change control. Qualify collagen lots, and require change notification and re-validation for any raw material or process change. An unannounced lot change is a common root cause in scaffold failures.
- Learning loop. Feed the deviation into the risk management file, the trial documentation and staff training, and report to the ethics committee at the next scheduled review with evidence that the corrective actions were effective.