What CAPAs Don't Catch: 3 Hidden ATMP Cold Chain Vulnerabilities By Design

By Rui Li, principal consultant, EastWind CryoWorks

Whether or not you work directly in clinical cold chain logistics, you’ve experienced supply chain failure from the receiving end. You’ve seen it, heard about it, or at least worried about it: a package arrives damaged; a shipment gets stranded; product quality drops somewhere between the supplier and the end user. Across industries and temperature ranges, the failure modes look familiar, and so are the consequences — disruption at the point of use and imbalance at the point of supply.

When CAPA Works — And Where CAPA Falls Short

Absorbing Failure As Operating Cost?

For high-volume biopharmaceuticals or off-the-shelf research materials, some level of cold chain failure is tolerable. When it occurs, the immediate act of correction and prevention is straightforward: issue a refund, send a replacement, add redundancy. These actions absorb the loss and keep the system moving.

For high-value low-volume advanced therapy medicinal products (ATMPs), however, the same logic breaks down. When a patient is waiting for the lifesaving therapy, a refund is irrelevant. A reshipment is not a solution but a gamble — each attempt carries added cost, added risk, and no guarantee of a different outcome. Even measures often framed as proactive, such as setting aside extra products or overfilling containers, do not resolve the issue — they simply sweep it under the rug and effectively lock in overage as a recurring cost for as long as the failure risks are unresolved.

In a context where each dose may be patient-specific, time-critical, and irreplaceable, these conventional remedies are not only economically unsustainable, they are fundamentally misaligned with the nature of the product. Cold chain failures in ATMPs cannot be absorbed; they must be addressed head-on.

Don’t Fix It Unless CAPA Tells Me To?

When CAPAs go deeper into root cause investigations, some failures trace back to isolated errors: a courier mishandling a well-labelled shipper, a raw material lot introducing particulate contaminations at the container interface, or a probe’s severed wiring disrupting the function of temperature control equipment. In response, controls are tightened, monitoring is expanded, and the cold chain system appears more robust.

Yet failure patterns persist. You find broken packages — one in every 15 shipments. You see product fail post-thaw QC unless handled by that one technician with the knack. And sooner or later, you catch yourself rationalizing, “The material is just finicky, perfect is the enemy of good, we’ve done enough, and I can’t justify bleeding more cash into fixing my cold chain.” I hear this across ATMP teams with striking consistency. Why didn’t CAPA make their cold chain problem disappear?

The answer lies in what CAPA is designed to do well — and what it is not. CAPA excels at identifying and stopping errors that are proximal to where a cold chain failure is observed and presented as isolated events, affecting some products or timepoints but not others. CAPA is far less effective at uncovering vulnerabilities embedded in the product and its cold chain processes. Unlike the discrete errors CAPA is intended to address, these designed-in vulnerabilities are much harder to see. They appear as random variation with no clear dependence on monitored data and are obscured by a false sense of confidence in upstream development decisions.

Underestimating The Delicacy Of ATMP Cold Chains

The limitation of CAPA is magnified for ATMPs. Unlike conventional therapeutics, their APIs are highly biologically active and acutely sensitive to mechanical, biophysical, biochemical, and thermodynamic stresses across the cold chain. Preserving them requires not only control of standard parameters (e.g., temperature, vibration) but precise coordination of ATMP-specific factors such as cryoprotectant composition, cooling rate, duration of ambient exposure, and dynamic changes in container closure integrity during temperature transitions.

If any parameter — or interaction between parameters — is overlooked, the product may be preserved in a subtly suboptimal state. These deficiencies often surface only at the end of the cold chain, upon thawing, as irregular or degraded outcomes that are easily misattributed to possible handling or shipping errors rather than traced back to design.

Complexity That Encourages Delay

Given the inherent complexity of ATMPs, teams often defer serious attention to cold chain risks until pivotal trials or commercial deployment — after packaging, preservation, and cold chain design are locked in — instead of addressing vulnerabilities during preclinical or pre-pivotal stages, when the end-to-end cold chain system is still being defined.

This delay may reflect conscious technical deprioritization along a complex translational pathway or a natural reaction to failure modes that only emerge at scale, as more lots are produced, more sites engaged, and more regions reached. Either way, it creates a mismatch: risks become visible when intervention becomes difficult. At that point, improving cold chain reliability often falls beyond what CAPA alone can address.

Stage-Specific Opportunities

For teams with future pipelines or those approaching first-in-human trials, there is a clear opportunity to build reliability into the cold chain system from the start. For those already deep in pivotal trial, regulatory preparation, or commercial scale-up, selective interventions at high-leverage low-disruption points across the end-to-end preservation system can still meaningfully tackle hidden vulnerabilities and alleviate the persistent headache they create.

How Vulnerabilities Got Designed Into ATMP Cold Chains — And What To Do

Materials Forced To Work In Incompatible Environments

Example 1: A glass vial is shipped in dry ice and breaks while an analyst handles it during downstream characterization. The operator feels guilty, the supervisor calls for more care, and the sender laments the loss. No one asks whether the vial material chosen and the cold chain it journeyed through were fundamentally incompatible.

Example 2: A cryogenic bag that is packaged inside an overwrap bag cracks at some point along the cold chain and leaks into the overwrap bag when it is thawed at bedside. The clinical operations team accepts the outcome as expected; the care team takes pride in prophylaxis protocols preventing infections; and the primary packaging supplier is satisfied with the overall decrease in bag breakage rate since the early 2000s. No one questions whether the bag’s materials of construction were ever evaluated against the full range of thermal and mechanical stresses encountered in the real-world deployment between fill/finish and bedside.

Issue: Treating the symptoms of primary packaging breakage often results in excessive control measures over handling procedures and excessive layering of packaging. These measures increase complexity but do not eliminate failure. Product loss and contamination risk persist.

Root cause: Materials selection and cold chain design are often siloed from each other. Packaging material choices are driven heavily by familiarity from a prior academic setting despite the translational gap, recommendations from adjacent biopharmaceutical industry domains despite the environmental difference, or precedent set by similar products. Cold chain decisions focus on chain of custody, chain of identity, and equipment function. Such disconnection between materials and environments can cause validated materials to fail in realistic cold chain conditions that exceed what was tested for and the validated cold chain to leave critical process parameters unoptimized and uncontrolled despite normal equipment function.

Solution: Start by identifying which side of the material-environment balance is fixed. Does my product require a specific type of container because of its aseptic fill/finish, route of administration, and dose considerations? Or does my product require a specific modality of cold chain because of its shelf life, distribution, and biological stability considerations? If the container is constrained, design the cold chain to avoid operating beyond its stress limits. If the cold chain is constrained, test and select container closure systems that can withstand it.

Promising Tools Prematurely Dismissed

Example 1: A crimp vial system fails initial CCI testing. The therapy developer concludes that crimp vials are unreliable and switches to heat-sealed options. The real issue — insufficient crimping force — goes unnoticed.

Example 2: A popular DMSO-based cryopreservation medium yields low recovery. The therapy developer rejects DMSO altogether, not realizing that the freezing protocol and the specific cryoprotectant concentrations were suboptimal, but a small amount of DMSO would prove essential after a long unsuccessful journey through DMSO-free possibilities.

Issue: Prematurely dismissing tools leads to strenuous design paths and often forces teams to revisit abandoned options later.

Root cause: Teams overgeneralize from limited data. They fall victim to test method artifacts, irrelevant technology classification, and misinformed down-selection. The risk of such pitfalls is reinforced by insufficient post-sale technical support from many tool suppliers and by misalignment between use case expectations and material specifications.

Solution: Resolving this compounded root cause requires coordinated action: suppliers must actively guide proper technology use; test labs and CROs must rigorously control test method parameters with deterministic impact on outcomes; and therapy developers must implement a quality by design approach to cold chain development — defining design space with product and workflow guardrails, optimizing materials and process variables, and then expanding optimal conditions into tolerable operating ranges instead of committing early to arbitrary specifications.

Connection Between Preservation Parameters Not Leveraged

Example: Point-of-care QC results fall short of production QC. Teams assume operational error: differences in equipment, lack of operator proficiency, or deviations in thawing procedures. In response, the sponsor piles on controls — site requalification, infrastructure prescription, workflow segregation, extensive staff training, and expensive automation systems — aimed at tightening process execution and protocol compliance.

Issue: A common misconception is that freezing is simple and thawing is the challenge. In reality, sensitivity during thawing often reflects upstream preservation weaknesses.

Root cause: Upstream parameters are designed in isolation but determine how the product behaves when thawed in real-world settings. Formulation choices (both excipients and workflow) are disconnected from freeze-thaw performance; freezing process parameters are not fit to protect the API from the stressful cold chain conditions; and batch variabilities within and across lots are poorly characterized and not mitigated.

Solution: Treat preservation as a system of interdependent levers, from cell culture conditioning, cryoprotectant formulation, freezing, storage, and shipping all the way to thawing. Fine-tune each parameter allows relaxation of another. For pre-IND teams, anchor efforts in product-centric risk assessment and define CMC through process guardrails — not rigid prescriptions — to preserve flexibility. For pre-pivotal trial teams, learn from real workflow bottlenecks to adjust key levers, and refine CQA–CPP links to guide meaningful design decisions. For post-pivotal and post-trial teams, pinpoint mechanisms of cryobiological damage in the current practice to target high-leverage fixes and balance cold chain improvements with regulatory and operational constrains to unlock low-disruption success.

Take Home

In ATMP cold chains, the most persistent failures are not the ones CAPA can isolate and correct but those quietly embedded in how the drug product, ancillary materials, and test and operational processes are first combined. When materials operate outside their limits, when technologies are dismissed without full understanding, and when preservation parameters are engineered in isolation, vulnerability becomes part of the design.

Addressing these challenges requires a shift from reactive correction to intentional integration:

• Map interactions across product, materials, processes, and human factors.
• Translate product expectations into clear functional criteria for each technology.
• Design fit-for-purpose test protocols that reflect real-world conditions.
• Engage expert resources to avoid missteps and rework.
• Treat cold chain development as a complex system that is engineered with rigor, not shaped by fear-driven assumptions.

Only then can teams move from firefighting failures to designing them out.

About The Author:

Rui Li, PhD, is a cryobiologist and biomedical engineer and the founder and principal consultant of EastWind CryoWorks. Partnering with ATMP developers, suppliers, and global stakeholders, she designs and integrates adaptive biopreservation approaches across temperature ranges and modalities, while bridging organizations and empowering innovators in the biopharmaceutical ecosystem. Dr. Li's known for delivering targeted technical solutions under resource, timeline, and external constraints, uncovering critical gaps, and aligning teams around executable paths forward. Her work enables startups and established organizations to tackle challenges often deemed impossible, translating complex problems into robust, end-to-end biopreservation and cold chain solutions.