Regulatory glossary

ISO 10993: Biological Evaluation of Medical Devices, Explained

ISO 10993 is the multi-part standard for the biological evaluation of medical devices, usually shortened to biocompatibility. Part 1 is the framework and the part you start from; the numbered parts under it cover individual endpoints and methods.

The most important thing about it is what it is not. ISO 10993-1 describes a risk-based evaluation, not a test checklist. Testing is what you do when the evaluation cannot answer a question from information you already have.

Contact type and duration set the endpoints

The framework categorises a device on two axes, and the pair determines which biological endpoints need addressing.

Nature of body contact:

• Surface devices: intact skin, mucosal membranes, or breached surfaces
• External communicating devices: blood path indirect, tissue/bone/dentin, or circulating blood
• Implant devices: tissue/bone, or blood

Duration of contact:

• Limited: up to 24 hours
• Prolonged: 24 hours to 30 days
• Long-term: over 30 days

Duration is cumulative, which is the part people get wrong. A device used for an hour a day over a month is prolonged or long-term contact, not limited.

The endpoints

Depending on category, the evaluation addresses some subset of: cytotoxicity, sensitisation, irritation or intracutaneous reactivity, acute systemic toxicity, material-mediated pyrogenicity, subacute and subchronic toxicity, genotoxicity, implantation effects, haemocompatibility, chronic toxicity, carcinogenicity, and reproductive or developmental toxicity.

Cytotoxicity, sensitisation and irritation apply to essentially everything. The rest scale with contact and duration.

Addressing an endpoint does not mean testing it. You address it with whatever evidence answers the question: chemical characterisation, published literature, data from an equivalent material with the same processing history, or a test.

Chemical characterisation comes first

This is the shift in the current thinking, and where older programmes still default to the wrong order.

ISO 10993-18 chemical characterisation, feeding a toxicological risk assessment under 10993-17, is the front of the process rather than an alternative to it. You characterise what can come out of the device, assess whether those constituents present a risk at the exposure your device produces, and test only where that assessment leaves a question open.

Running the full battery first is slower, more expensive, uses more animals, and frequently answers questions the chemistry had already closed.

What FDA expects

FDA recognises ISO 10993-1 with its own guidance, and the guidance matters as much as the standard. Two points recur:

Your evaluation is of the final, sterilised device, in its final configuration, with its actual manufacturing and processing history. Data on the raw material is not data on your device. Sterilisation, cleaning agents, mould release and adhesives all change the answer.

A change in material, supplier or process reopens the evaluation. Not automatically the whole battery, but a documented assessment of whether the change affects the biological risk.

Where biocompatibility programmes go wrong

• Testing before characterising. Expensive, slow, and often unnecessary.
• Evaluating the material rather than the device. The device is what contacts the patient, after everything you did to the material.
• Contact duration understated. Cumulative, not per use.
• Supplier change treated as a purchasing decision. A different resin from a different plant is a biocompatibility question.
• Equivalence claimed too loosely. Relying on data from a similar device requires demonstrating the same material, the same processing and the same contact, not just the same product category.
• No link to the risk file. Biological risk belongs in the ISO 14971 risk management file like any other, and an evaluation that lives only in a test report is disconnected from the device's overall residual risk.