Every shelf life claim printed on a medical device label must be supported by stability testing evidence inside the technical documentation. MDR Annex I Section 11.4 of Regulation (EU) 2017/745 requires devices delivered in a sterile state to be packaged so that sterility is maintained under the specified storage and transport conditions until the protective packaging is damaged or opened. That obligation only holds if the manufacturer has actually demonstrated the shelf life through a documented stability programme. The evidence lives in Annex II Section 6 and consists of two linked workstreams: accelerated ageing studies that support the initial label claim, and real-time ageing studies that confirm the accelerated results as they mature. A shelf life claim without stability data behind it is a finding a Notified Body will write on the first read.

By Tibor Zechmeister and Felix Lenhard. Last updated 10 April 2026.

TL;DR

  • MDR Annex I Section 11.4 requires the sterile pack to maintain sterility through the declared storage and transport conditions until opened, and the declared shelf life must be supported by evidence.
  • Stability testing evidence is filed in Annex II Section 6 of the technical documentation, cross-referenced from the GSPR checklist line for Annex I Section 11.
  • The evidence structure is two-track: accelerated ageing studies support the initial claim at launch, and real-time ageing studies confirm the claim as it accrues.
  • Stability testing covers the sterile barrier system, the device performance, and any material properties that could degrade over time. All three layers must be in the file.
  • Shelf life is not set by what the team hopes. It is set by what the data supports.

The conversation that gives it away

A founder sends the technical file for a first-audit review. The label says "shelf life 24 months." The Annex II Section 6 folder is well organised for sterilisation, reasonably organised for biocompatibility, and silent on ageing. The stability evidence is one accelerated ageing report on the pouch, no product functional testing after ageing, no real-time data, and no justification for the 24-month number.

The question is simple. Where does 24 months come from? The answer, almost always, is that someone on the team picked it because 24 months is what the market expects. That is not a shelf life. That is a wish. Notified Body reviewers know the difference, and the finding writes itself.

What shelf life means under MDR

Annex I of Regulation (EU) 2017/745 contains the General Safety and Performance Requirements. Section 11 covers infection and microbial contamination, and Section 11.4 is the specific obligation for devices delivered in a sterile state. The regulation requires that such devices be packaged to ensure they remain sterile under the storage and transport conditions specified by the manufacturer until the protective packaging is damaged or opened. The manufacturer specifies those conditions and that duration on the label. The duration is the shelf life.

Shelf life is the period during which the manufacturer guarantees that the device, stored and transported within the declared conditions, will still perform as intended and, for sterile devices, will still be sterile. It is a regulatory claim, not a marketing number. Once it is on the label, it must be supportable on demand.

For non-sterile devices, the shelf life concept still applies wherever the device has time-dependent properties: batteries that self-discharge, reagents that degrade, polymers that change mechanical behaviour, adhesives that lose tack, electronic components whose calibration drifts. If anything in the device gets worse with time, the label either declares a shelf life or it does not, and either way the file must justify the position.

Accelerated versus real-time stability studies

Stability testing uses two complementary approaches. Both belong in the dossier, and neither replaces the other.

Accelerated ageing exposes the device to elevated temperature — and where relevant, elevated humidity — to simulate the effect of longer storage in a shorter calendar period. The theoretical basis is Arrhenius kinetics: chemical degradation rates roughly double for every ten-degree Celsius increase in temperature, so a device held at fifty-five degrees Celsius for a defined period approximates a much longer period at room temperature. The technique is routine in the sterile barrier packaging world and is documented in widely used ageing protocols that Notified Bodies recognise.

Accelerated ageing is how manufacturers support an initial shelf life claim at launch without waiting years for real-time data. It is a model, and like any model, it can be wrong. Not every failure mode follows Arrhenius. Mechanical fatigue, UV damage, battery self-discharge, and some polymer crystallisation processes do not accelerate cleanly with temperature. The accelerated protocol must be chosen with the specific failure modes of the specific device in mind, not applied as a standard recipe.

Real-time ageing runs the actual clock. The product sits in its actual packaging at the declared storage conditions for the declared duration, and the test points at six months, twelve months, eighteen months, and so on generate data that confirms or contradicts the accelerated model. Real-time studies start at launch and accumulate during the market life of the product. A launch with twenty-four-month accelerated data and real-time data running in parallel is the normal pattern. A launch with twenty-four months of accelerated data and no real-time programme running is an incomplete position.

The dossier must contain both: the accelerated protocol and report that supports the initial claim, and the real-time protocol and schedule that will confirm the claim over time.

Packaging integrity testing

The sterile barrier system is where most stability testing lives, because a sterile device that fails its packaging is unsterile on arrival. Packaging integrity testing covers seal strength, burst strength, dye penetration or other leak testing, microbial barrier integrity, and visual inspection for damage. Each of these tests is run at the start of life, at each accelerated ageing timepoint, and at each real-time ageing timepoint.

The test methods come from recognised packaging standards that the Notified Body expects to see referenced. The specific standard used is less important than the discipline of the reference: the protocol cites the standard, the report shows the acceptance criteria and the measured results, and the results sit in Annex II Section 6 alongside the sterilisation validation.

Transport simulation belongs in this part of the file too. A sterile barrier system that passes ageing in a climate chamber but has never been tested under realistic distribution vibration, drop, and pressure conditions has not been fully qualified. The stability programme is not complete without it.

What to document in Annex II Section 6

Inside the Annex II Section 6 pre-clinical and clinical evidence section, the shelf life and stability subsection typically contains: a stability master plan describing the overall approach and the rationale for the declared shelf life, the accelerated ageing protocol and report, the real-time ageing protocol and the ongoing schedule, the packaging integrity test methods and acceptance criteria, the device functional test methods and acceptance criteria at each timepoint, the transport simulation report, and any material-level stability data relevant to the failure modes identified in the risk file.

Each document is version-controlled under EN ISO 13485:2016+A11:2021. Each document is cross-referenced from the Annex I Section 11.4 line of the GSPR checklist in Annex II Section 4. The chain from GSPR line to evidence to raw data must resolve on a single sampling pass. If a reviewer asks for the data behind the accelerated ageing report and the path takes more than one step, the structure is not audit-ready.

The stability master plan is the document that ties the workstream together. It states the declared shelf life, identifies the failure modes the programme is designed to detect, specifies the timepoints and acceptance criteria, and assigns ownership for the ongoing real-time work. Without a master plan, the individual reports read as a pile of tests. With one, the file reads as a single argument.

Stability testing and sterilisation validation are adjacent workstreams in the same part of the technical file, and they are not independent. The sterile barrier system qualification is part of the sterilisation validation dossier, and the ageing of that same sterile barrier system is part of the stability dossier. The two dossiers cross-reference each other.

A common error is to treat them as separate subjects owned by separate people with separate plans that never meet. The result is a file where the sterilisation section assumes the packaging holds over the claimed shelf life, the stability section assumes the sterilisation process does not damage the packaging in ways that only show up under ageing, and nobody has tested the combined hypothesis. The correct structure tests the sterilised package after ageing. Sterilise first, then age, then run the integrity tests. That is the condition the device will actually see on the hospital shelf, and that is the condition the evidence must demonstrate.

Common gaps in startup stability dossiers

The recurring patterns when we review stability sections before a first audit:

  • A shelf life claim on the label with no stability master plan behind it. The number was chosen commercially and the evidence was never built to match.
  • Accelerated ageing only, no real-time programme running. The launch claim is supportable in theory, but the confirmation clock was never started.
  • Accelerated ageing with no Arrhenius justification. The temperature and duration were picked from a template without checking whether the assumed kinetics apply to the actual failure modes.
  • Packaging integrity tested, device functionality not tested after ageing. The pouch is fine at twenty-four months. Whether the device inside still works is an open question.
  • Transport simulation missing or run on unaged product only. The ageing and the distribution stresses were never combined in a single test.
  • Unsterilised samples used for ageing. The packaging was aged before it went through the sterilisation cycle, so the test condition does not match the market condition.
  • Real-time data accumulating in a spreadsheet but never folded back into the file. The study is running, the report has not been updated, and the file still shows the original accelerated-only position.

Each of these is a finding. None of them is rare.

The Subtract to Ship angle

Stability testing is an area where the pressure to add tests is intense and the pressure to think first is weak. It is cheaper to run one more timepoint than to argue about whether the timepoint adds evidence. A file built by accumulation ends up with a stack of reports that collectively do not answer the question the Notified Body will ask, which is: does the declared shelf life rest on a defensible argument?

The Subtract to Ship framework for MDR applies the same test here. Every stability document must trace to a specific failure mode identified in the risk file, to a specific acceptance criterion in the design specification, and to a specific line in the GSPR checklist. Tests that cannot answer "what does this show?" come out. What remains is a stability programme a reviewer can follow in one pass and a team can keep current without drowning.

The startups that get this right write the stability master plan before the first sample goes into the oven. The ones who pay twice run samples first and try to write the rationale around the data afterwards.

Reality Check — Where do you stand?

  1. Does the shelf life number on your label appear in a written stability master plan that predates the test programme?
  2. For each failure mode in your risk file that could degrade with time, is there a specific stability test with a specific acceptance criterion?
  3. Is your accelerated ageing protocol grounded in a documented Arrhenius or equivalent kinetic argument for the specific materials and failure modes of your device?
  4. Is a real-time ageing programme actually running today, with the next timepoint on the calendar and a named owner?
  5. Does your stability evidence test the sterilised package after ageing, not the unsterilised package, or the sterilised package before ageing?
  6. Is device functional performance tested at each ageing timepoint, or only packaging integrity?
  7. Is transport simulation included in the stability programme, and was it run on aged samples?
  8. Does the Annex I Section 11.4 line of your GSPR checklist resolve to a specific stability report, and does that report trace to the raw data in one step?

Frequently Asked Questions

What MDR section covers shelf life for sterile devices? Annex I Section 11.4 of Regulation (EU) 2017/745. It requires devices delivered in a sterile state to be packaged so that sterility is maintained under the specified storage and transport conditions until the protective packaging is damaged or opened. The declared shelf life is the period for which that obligation holds.

Where does stability testing evidence go in the technical file? In Annex II Section 6, the pre-clinical and clinical evidence section, cross-referenced from the Annex I Section 11.4 line of the GSPR checklist in Annex II Section 4. Accelerated and real-time ageing reports, packaging integrity tests, transport simulation, and the stability master plan all live together in this subsection.

Can a manufacturer launch with accelerated ageing data alone? Yes, provided the accelerated protocol is scientifically justified for the specific failure modes of the device, and provided a real-time ageing programme is running in parallel to confirm the accelerated results as they accrue. Accelerated data alone, with no real-time follow-up, is an incomplete position that a Notified Body will flag.

Does a non-sterile device need a shelf life claim? Not always. If nothing in the device degrades with time under realistic storage conditions, no shelf life claim is required. If anything does — batteries, reagents, adhesives, calibration, polymer properties — the label either states a shelf life supported by stability data or justifies the absence of a claim with evidence.

How is the accelerated ageing temperature chosen? From the Arrhenius relationship applied to the specific failure modes of the specific materials. A common starting point for sterile barrier packaging uses fifty-five degrees Celsius, but the temperature and duration must be justified in the stability master plan for the actual device, not copied from a template.

Does device functionality have to be tested after ageing, or is packaging integrity enough? Both. Packaging integrity confirms sterility is maintained. Device functional testing confirms the device itself still performs as intended at the end of its declared shelf life. A Notified Body will expect both layers of evidence for any device where performance could credibly change with time.

What happens if the real-time data contradicts the accelerated prediction? The shelf life claim is revisited. If real-time performance at eighteen months is already trending toward the acceptance limit, the twenty-four-month claim is not supportable and the label must change. This is why the real-time programme is not optional theatre — it is the mechanism that protects the manufacturer from discovering the problem through a field complaint instead of through planned monitoring.

Sources

  1. Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on medical devices, Annex I Section 11.4 (sterile pack shelf life), Annex II Section 6 (pre-clinical and clinical evidence), and Article 10 (general obligations of manufacturers). Official Journal L 117, 5.5.2017.
  2. EN ISO 13485:2016 + A11:2021 — Medical devices — Quality management systems — Requirements for regulatory purposes.

This post is part of the Technical Documentation cluster in the Subtract to Ship: MDR blog. Authored by Felix Lenhard and Tibor Zechmeister. Tibor has reviewed stability dossiers from both sides of the Notified Body table. Where the shelf life claim on the label is backed by a stability master plan written before the first sample went into the oven, first audits go smoothly. Where the number was chosen by marketing and the evidence was reverse-engineered, they do not.