MDR Annex I · ISO 10993 · biological risk
Biological evaluation of medical devices
Not a test list. A scientific case for what the patient actually encounters.
A familiar polymer can become a different biological proposition after moulding, bonding, cleaning, packaging, sterilisation and ageing. Biological evaluation follows that transformation all the way to the finished device.
Trace the evidence ↓Additives · colourants · adhesives
Processing · cleaning · residues
Method · dose · residuals · ageing
Tissue · route · duration · frequency
As supplied and used
The central idea
The patient never meets a material certificate
The patient meets a finished device, carrying the history of every material choice, process step, residue, sterilisation cycle and storage condition.
That finished device is the real subject of biological evaluation.
01
Start with the device that reaches the patient #
Biological evaluation is sometimes treated as a late-stage package of cytotoxicity, sensitisation and irritation reports. That approach starts at the wrong end of the problem.
The MDR first asks whether the device is designed and manufactured to protect patients and users from risks arising from its chemical, physical and biological properties. The ISO 10993 framework then turns that broad obligation into a structured biological risk assessment.
The assessment therefore begins with the finished device in its clinically relevant condition, not with a generic material name and not with a supplier’s statement that a resin is “medical grade”.
Base materials, additives, inks, coatings, adhesives and processing aids
Machining, moulding, curing, welding, cleaning and handling
Residues, degradation products, particles and sterilisation by-products
Contact site, route, duration, frequency, population and cumulative use
If a piece of evidence does not represent the marketed device, the relevant manufacturing state or the intended clinical exposure, its reassuring conclusion may not travel as far as expected.
02
Contact turns a device description into a biological question #
The same material can present very different risks in a skin-contacting sensor, a blood-contacting catheter and a permanent implant. Biological relevance comes from exposure.
A useful contact map captures more than a category label. It reconstructs how the device and its constituents may reach the patient throughout the intended lifetime.
Who is exposed?
Patient, user or both, including vulnerable populations where relevant
What is contacted?
Intact skin, breached surface, tissue, bone, blood path or another anatomical environment
How does exposure occur?
Direct contact, indirect fluid path, released constituent, particle or degradation product
How much and how often?
Single use, repeated procedures, cumulative duration, surface area and dose-related factors
What can change over time?
Ageing, wear, corrosion, degradation, reprocessing, transport and storage
Contact classification helps organise the evaluation, but it does not mechanically dictate a fixed list of tests. The biological endpoints must still be justified using the device’s actual hazards and available evidence.
03
The biological evaluation plan is the scientific architecture #
The biological evaluation plan should be written before new studies are commissioned. Its job is to define the questions, organise the existing knowledge and expose the genuine gaps.
Device family, variants, accessories, contact pathways and intended populations
Composition, processes, prior tests, literature, clinical history and post-market information
Potential biological effects linked to materials, constituents, residues, degradation and exposure
Questions that remain unanswered after the existing evidence is critically assessed
Analytical, toxicological or biological work needed to close each remaining gap
Acceptance logic, residual uncertainty and links to the wider risk-management file
ISO 10993-1:2025 replaced the 2018 edition at international level and was reorganised around biological safety within risk management. For an MDR submission, the manufacturer should also verify the applicable EN adoption, current Official Journal citations and transition strategy rather than assuming that the newest ISO edition automatically carries harmonised status.
04
Use knowledge before generating more data #
A strong evaluation does not reward the largest stack of laboratory reports. It rewards the clearest explanation of why the available evidence is relevant, sufficient and internally consistent.
Existing information may answer a biological question completely, partly or not at all. The evaluator’s task is to distinguish among those outcomes without overstating weak evidence.
Materials, formulations, suppliers, processing, cleaning, packaging and sterilisation
Composition, extractables, leachables, residues, particles and degradation products
Prior studies on the device, justified family members or relevant materials
Peer-reviewed literature, authoritative toxicology sources and clinical experience
Complaints, vigilance, PMS, PMCF and experience with changed or aged devices
Supplier declarations and material safety data can support the story, but they rarely close it alone. They may omit formulation detail, processing aids, trace constituents or the effects of manufacturing and sterilisation.
05
Chemical characterisation asks what can reach the patient #
Chemical characterisation can reveal the difference between what the device is supposed to contain and what may actually be released from it.
Under ISO 10993-18, the work may move from material and compositional information to analytical chemistry, depending on what is already known and what remains uncertain. Extraction design, analytical methods, reporting thresholds and identification strategy must be linked to the intended exposure and the question being answered.
Questions the chemistry strategy should answer
- Which device, variant and manufacturing condition does the sample represent?
- Are the extraction conditions clinically relevant, exaggerated or both?
- How are volatile, semi-volatile, non-volatile and inorganic constituents addressed?
- How are unknown or partially identified compounds handled?
- Does the analytical sensitivity support the toxicological assessment?
- How are uncertainty, recovery and method limitations carried forward?
Chemistry is not a ritual that ends with a compound table. Its value appears only when the results are converted into a defensible exposure assessment.
06
Toxicology turns detected chemicals into a patient question #
A detected constituent is not automatically a clinical hazard, and a small analytical peak is not automatically negligible. Risk depends on identity, toxicological properties and the exposure created by the device.
ISO 10993-17:2023 describes toxicological risk assessment for medical-device constituents and connects chemical characterisation data with exposure-based safety evaluation.
The assessment should make its assumptions visible. Patient body mass, frequency of use, duration, released amount, uncertainty factors, route of exposure and special-population considerations can all change the conclusion.
Chemical characterisation without toxicological interpretation produces a list. Toxicology without a credible exposure estimate produces a theoretical discussion. The biological evaluation needs the bridge between them.
From signal to meaning
The work is not complete when the instrument stops
An analytical result becomes useful only after it is connected to exposure, biological relevance and the risk-management decision.
What leaves the device under justified conditions
What the analytical methods can identify and quantify
What the patient may receive during actual use
What toxicology and biological evidence say about that exposure
Whether the residual biological risk is acceptable
07
Test only the biological questions that remain unresolved #
Biological testing remains important, but it should be triggered by a defined evidence gap. A test is useful when its design and endpoint can materially change the safety conclusion.
Depending on the device, relevant questions may involve cytotoxicity, sensitisation, irritation, systemic toxicity, material-mediated pyrogenicity, blood interaction, implantation effects, genotoxicity, carcinogenicity, reproductive or developmental toxicity, degradation and toxicokinetics.
This is an endpoint landscape, not a mandatory menu. Each included, excluded or deferred endpoint needs a reason grounded in exposure, hazards and available evidence.
The sequence matters as well. Existing data, chemical characterisation, toxicological assessment and scientifically valid non-animal approaches should be considered before commissioning a new animal study. ISO 10993-2:2022 expressly supports replacement, reduction and refinement in animal use.
08
The wrong test article can make a good study irrelevant #
The test sample must represent the device condition that matters. That may be the finished, sterilised and aged product, a worst-case variant, a fluid-path assembly or another justified configuration.
Worst case is not always the largest device. It may be the variant with the greatest surface area, highest additive content, most demanding processing, highest sterilisation exposure, longest contact, lowest rinse efficiency or greatest patient dose.
Surface area, mass, porosity, coating and accessible fluid path
Formulation, additives, colourants, adhesives and component ratios
Temperature, curing, cleaning, welding and manufacturing residues
Method, dose, cycle count, residuals and material transformation
Beginning and end of shelf life, packaging interaction and degradation
Contact route, use duration, repeat use and maximum patient dose
Sample preparation should follow the applicable ISO 10993 principles and the scientific purpose of the study. A convenient coupon or unprocessed raw material cannot silently replace the finished device.
09
Equivalence succeeds or fails in the details #
Read-across from another device can prevent unnecessary testing, but a shared trade name or polymer family is not enough.
The comparison must cover the characteristics that determine biological exposure. Where the comparator’s detailed composition, manufacturing information or chemical profile is inaccessible, the apparent shortcut can become the least defensible part of the file.
Formulation, additives, suppliers and proportions
Manufacturing, cleaning, sterilisation and ageing
Anatomical site, route, duration, frequency and population
Release profile, residues, degradation and uncertainty
Can the available comparison show that the evaluated device does not create a greater or different biological exposure than the reference evidence already covers?
10
A biological evaluation is a living control, not a launch document #
The original conclusion remains valid only while its assumptions remain valid. Design changes and new lifecycle data must be screened for biological significance.
Re-evaluation does not automatically mean repeating every study. It means asking which evidence, assumptions and risk conclusions are affected, then generating only the additional data needed to restore confidence.
11
Build a file that lets the reviewer follow the reasoning #
The biological evaluation report should not merely summarise studies. It should close the loop opened by the plan and show how every significant biological question was resolved.
Define the device, contact and exposure
Identify biological hazards and relevant endpoints
Critically evaluate every relevant evidence source
Explain gaps, limitations and uncertainty
Connect chemistry, toxicology and biological studies
Conclude on residual risk for the intended use
Link conclusions to risk controls, PMS and change management
Failure patterns that attract questions
The file cannot explain why each endpoint was selected or omitted.
Manufacturing, residues, sterilisation and ageing were not bridged.
Detected constituents were not connected to credible patient exposure.
The true compositional, processing or dose-related worst case was missed.
Detailed composition, processing and release characteristics were unavailable.
The evidence package became a chronology of studies rather than a risk argument.
Official references
- Regulation (EU) 2017/745, Annex I and Annex II
- ISO 10993-1:2025 on biological safety evaluation within risk management
- ISO 10993-18:2020 on chemical characterisation
- ISO 10993-17:2023 on toxicological risk assessment
- ISO 10993-23:2021 on irritation testing
- ISO 10993-2:2022 on animal welfare requirements
- European Commission overview of harmonised medical-device standards
The bottom line
Biological safety is an argument built from the whole device
The MDR does not ask for a ceremonial bundle of familiar biocompatibility reports. It asks the manufacturer to control biological risk in the device that is actually placed on the market.
That requires a disciplined chain from composition and manufacturing to contact, chemistry, exposure, toxicology, biological response and lifecycle evidence.
The most convincing evaluation is not the one with the most tests. It is the one in which every test, omission and conclusion follows visibly from the risk.
Biological evaluation strategy
Need to turn fragmented biocompatibility evidence into one defensible MDR story?
Discuss your biological evaluation ↗From the knowledge centre
