The question I get from OEM packaging engineers isn't usually how to run accelerated aging. It's whether accelerated data alone will get them through the reviewer, or whether they need real-time data sitting in the file before they hit submit.
The honest answer: accelerated aging can carry your submission, but only if you understand what it is actually buying you and where it stops being enough. That line is where most of the confusion lives, and it is exactly the judgment call the standards decline to make for you.
What the Standards Actually Require
ISO 11607-1:2019 treats stability, meaning real-time aging evidence, as a requirement for the sterile barrier system, with Annex B cataloging the test methods that back a conformity claim (ISO 11607-1:2019). Read that plainly: the standard wants real-time data eventually. It does not say you must have five years of it before you can market a five-year device.
That gap is where ASTM F1980-21 does the heavy lifting. The standard states that accelerated aging data may be regarded as sufficient evidence to support expiration-date claims until real-time aging data becomes available, and that real-time studies must ultimately confirm the accelerated results using the same evaluation methods (ASTM F1980-21). So the framework is intentional. Accelerated aging supports the initial claim. Real-time aging confirms it.
Here is what that means for a 510(k) or PMA package. You can submit and clear on accelerated data. Reviewers accept it routinely. But the acceptance is conditional on a real-time program that is running in parallel, with a documented pull schedule and acceptance criteria. Submitting accelerated data with no real-time arm in motion is the version that draws questions. The reviewer isn't asking for five years of data at submission. They're asking for evidence that you understand the accelerated result is a prediction, not a final answer.
Where Accelerated-Only Actually Holds Up
Accelerated-only data is on its strongest footing when the shelf-life claim is modest, the materials are well-characterized polyolefin films and Tyvek with no moisture-sensitive components, and the aging parameters are conservative and justified. A two- or three-year claim on a conventional pouch or tray system, with a real-time arm started at the same time as the accelerated study, is a package reviewers see constantly.
The expectation tightens as the risk climbs. Longer shelf-life claims, novel materials, moisture-sensitive contents, implantables, and PMA-class devices all shift the reviewer toward wanting to see real-time confirmation initiated, and sometimes early real-time pull points already reported. An implantable neurostimulation program (NCT03029624) documented exactly this structure: accelerated aging per ASTM F1980 run in parallel with a real-time arm, followed by ASTM F88 seal-strength and ASTM F1886 seal-integrity testing on the aged samples. The parallel design is the tell. Higher-risk programs don't choose between the two methods. They run both and they start early.
The Q10 Model Is an Assumption You Have to Defend
Accelerated aging works by using temperature to compress time, and the math rests on the Arrhenius relationship expressed through a Q10 temperature coefficient. For every 10 degrees C above your assumed ambient storage temperature, the chosen Q10 governs how much the degradation rate accelerates, and that factor is what converts your real-time shelf life into an accelerated aging duration (ASTM F1980-21).
A Q10 of 2.0 is the common default, and for most medical packaging polymers it is defensible as a conservative choice. But the standard requires you to justify the selected Q10, and an unsupported value undermines the entire predicted shelf life (ASTM F1980-21). This is the piece I see engineers skip. They plug in 2.0, run the clock, and never write down why 2.0 is appropriate for their material system.
Think about what the number actually claims. A Q10 above 2.0 means your accelerated study is overpredicting shelf life, which is the direction that gets you in trouble. A Q10 below 2.0 means the study is more conservative than reality. Unless you have material-specific kinetic data, 2.0 is the value you can stand behind, but you have to state that it is an assumption and not a measured property of your package.
Temperature and Humidity: The Two Decisions That Sink Protocols
Temperature selection is a balancing act. Go too low and the study takes forever. Go too high and you induce failure modes that would never happen on a hospital shelf. If your adhesive-coated Tyvek header or a multilayer film starts deforming at 58 degrees C, running at 55 degrees C leaves you no margin, and a warped seal at the aging chamber tells the reviewer nothing about real storage. That is not aging. That is destruction. Pick a temperature high enough to be useful and low enough to stay representative, and document how you knew the difference.
Humidity is the decision that quietly failed a lot of older protocols. If your package or device contains moisture-sensitive materials, controlled humidity during aging is not optional, and you need a rationale for the level you chose. A moisture-sensitive hydrated tissue graft program documented in US Patent 10,828,141 ran its ASTM F1980 study at a controlled 20 to 25 percent RH precisely because the contents demanded it. If your system is polyolefin film and Tyvek that doesn't meaningfully interact with moisture, ambient humidity is fine, but you still owe the protocol a sentence explaining why. The failure mode isn't running at the wrong humidity. It's saying nothing about humidity at all.
Don't Extrapolate Past What Your Data Supports
The temptation with a clean accelerated result is to stretch it. You aged to support three years, the seals looked great, so surely you can claim five. Don't. The aging factor you calculated supports the duration you ran, at the parameters you justified, on the materials you tested. Extrapolating a shelf-life claim beyond the ratio your study actually covered is a claim without data underneath it, and a reviewer who does the arithmetic will catch the gap. If you want five years, run the study that supports five years, and start the real-time arm out to five years, too.
When Real-Time Data Contradicts the Accelerated Study
This is the scenario nobody plans for and everybody should. You cleared on accelerated data predicting a five-year shelf life. At the year-three real-time pull, seal strength drops below your specification. Now what?
The accelerated prediction was wrong, and the label claim is no longer supported. That is a genuine problem, and it can mean a labeling change, a shelf-life reduction, revalidation, or a field action depending on severity. The reason the standards insist real-time must confirm accelerated is precisely to surface this before it becomes a patient-safety issue.
The way you protect yourself is upstream. Start real-time aging at the same time as the accelerated study, not two years later. Build in an early real-time pull, at six months or a year, so a diverging trend shows up while you still have room to react. And treat a trending-but-still-passing result, seal strength declining but inside spec, as a signal that deserves a documented risk assessment, not a shrug. The engineer who has an early real-time pull already reported has a story to tell the reviewer. The one who deferred real-time until an auditor asked has a gap.
The Underlying Obligation
All of this rolls up to design validation. Packaging shelf-life stability evidence supporting your labeled expiration date is part of design and development validation under ISO 13485:2016 Clause 7.3, which now flows through the FDA framework as the QMSR. The QMSR retains the obligation that packaging protect the device and maintain its condition through processing, storage, handling, and distribution. The standards tell you the method. The QMS tells you it has to be traceable, justified, and defensible.
Accelerated aging gets you to market. Real-time aging keeps you there. The submission that clears cleanly is the one where both are running, the Q10 and temperature and humidity choices are each written down with a reason, and the shelf-life claim never reaches past the data that supports it.