Material Matters

The standard tests for the biological evaluation of biomaterials, as promoted by the International Organisation for Standardisation, have evolved over the years to become an important part of the process for ensuring, as far as is possible, the biological safety of medical devices. There is, however, some scope for improvement, especially when we consider the extremely different requirements for new medical technology products.

By: David Williams

Regulatory testing and medical device failures

I hate to criticise and this article is not meant to be critical, but rather an attempt to give a few constructive suggestions about a thorny issue. I have just looked at the latest version of the International Organisation for Standardisation’s ISO 10993 series of documents that relate to the biological evaluation of medical devices, published in June 2009.1 As virtually all readers of Medical Device Technology will know, this is a comprehensive series of standards that has evolved over many years. It has obviously been compiled with great care; many friends of mine sit on the various committees and devote much time to their tasks. New sections continue to be added, and the tests themselves have in many cases been updated with better procedures, which often reflect the need for greater accuracy and quantification of evaluation parameters. The standards contain a number of sensible caveats and warnings. These include those that remind us that there can be significant differences between species with respect to animal testing, and that we cannot guarantee correlations between in vitro tests, in vivo tests and clinical performance.

But I often wonder if the ISO 10993 series serves the community to the best extent possible, especially if it forms the basis for the evaluation of future classes of biomaterials. Let me make three points to demonstrate my concern. The first is that although medical devices perform a massive beneficial service within healthcare, problems of a lack of biological safety still occur in many clinical situations. This is in spite of the fact that the materials will have been tested within the framework of ISO 10993 and that regulatory approval has been granted on the basis of the successful negotiation of these tests. If parts of aeroplanes failed, and people died as a result of materials deficiencies even though those materials met the specifications under the industry standards, we would see planes grounded and better specifications and standards urgently put in place. The same would happen in the nuclear industry. Wherever the safety or health of consumers or citizens in general is compromised by materials failures, appropriate steps have to be taken. In this column I have, on occasion, drawn attention to issues where less than satisfactory performances have been seen, for example, with metal-on-metal orthopaedic devices and drug eluting devices such as stents and spinal fixtures. I do not underestimate the difficulty of generating tests that are more predictive of clinical performance, but the fact that failures occur with devices (and by implication the materials of their construction) that have passed FDA scrutiny and/or obtained a CE mark is prime facie evidence that the test procedures are not optimal. To avoid any doubt, I should stress here that I am well aware of the need to balance innovation and risk. There is the possibility of doing as much harm to patients by denying regulatory approval to innovative materials through the use of over-aggressive test regimes as there is by allowing doubtful materials into the market. However, it would be irresponsible not to point out the need to avoid complacency and to develop significantly better procedures.

Assessment of risk using the standard

My second and third points are concerned with the marked differences between traditional medical devices and those that are now emerging with the developing fields of bio-
nanotechnology, imaging materials and regenerative medicine. In the first category, and especially with long term implantable devices, I have argued in this column and elsewhere2 that a half century of clinical experience and laboratory testing has led to the paradigm that the best performance of biomaterials is achieved with those that are maximally chemically and biologically inert. There are few exceptions to this rule, and almost every attempt to deliberately introduce biological activity into these types of device through modification of the materials has resulted in inferior rather than improved performance.

Now let us look at the wording of the recently published ISO 10993 Part 5, which deals with in vitro cytotoxicity. I am pleased to note that more tests are described than previously, and the person responsible for the testing has a number of options. The wording of the standard still, however, gives the option of testing extracts derived from the material in question or using direct contact techniques. Furthermore, although stating that quantitative test data are preferred, it is the responsibility of the sponsor or test facility to decide whether the test should be qualitative or quantitative. With the qualitative procedures, morphological or reactivity grading scales can be used. In either case, the wording of the standard states that the achievement of a numerical grade greater than two signals a cytotoxic effect. Thus with respect to morphology, a grade of two, which is considered to be mild, is described as a response in which no more than 50% of the cells are round and devoid of intracytoplasmic granules, and where there is no extensive cell lysis and no more than 50% growth inhibition. This response allows a material to pass the test. The standard does indicate that any cytotoxic effect can be of concern and it rightly states that the in vitro cytotoxicity data is primarily an indication of the potential for in vivo toxicity; however, I am uncomfortable with a situation in which 50% of cells that are exposed to a material or its extract are visibly affected by that exposure, yet the material is considered to be nontoxic.

Biological evaluation of new technologies

Moving to the third and most significant point, we should recognise that these tests were developed with conventional medical devices in mind. The field of medical technology is changing. Let us take tissue engineering as an example of a newly emerging sub-set of these technologies. Tissue engineering has a number of manifestations, but the most widely used format is one in which cells are seeded into a biomaterial scaffold, usually within an ex vivo bioreactor, and then stimulated to generate new tissue. These procedures have been under development for approximately 20 years, and although there are now some successes, it has proved extremely difficult to generate sufficient amounts of high quality tissue in those situations. There are many reasons for this, but one of the most significant is the failure to produce optimal scaffold materials. Scaffolds are usually intended to be degradable and replaced by the newly forming tissue. It requires a considerable effort to design a material that has to support the cells whilst they are being stimulated to produce the extracellular matrix of the new tissue, and where that support concurrently degrades without the degradation process and without the degradation products having any negative effect on those cells or the new tissue. This is not a trivial process, but the majority of tissue engineering scaffolds incorporated into devices and procedures to date have had one specification in mind: the materials have had prior regulatory approval with respect to their incorporation into classical medical devices. This means that they should have passed the ISO 10993 standard tests for biological safety.

Although understandably pragmatic, this misses the whole point of the tissue engineering process. With a long term implantable device, the test procedures should be aimed at demonstrating that the material has no biological properties and can be used in the body without biological effect. In tissue engineering, the opposite is true: the scaffold material, through bulk or surface properties should actively support the cellular activity. The surface of the scaffold material should assist in optimising cell adhesion, proliferation, differentiation and other processes, typically by having molecular structures that positively interact with receptors on cell membranes. The material may be required, through its mechanical characteristics, to facilitate mechanotransduction, that is, the mechanical signalling of cells. It may be required to deliver growth factors to the cells, or specifically and directly encourage new blood vessel formation. It is disappointing to see, therefore, that the ISO 10993 series appears to assume that these tests can equally well be used for tissue engineering products. ISO 10993 Part 6, which is concerned with tests for local effects after implantation, specifically states that it can be used to assess the local effects associated with tissue engineered medical products. I do not believe that this is the case.

If we have tests that are predicated on avoiding biological consequences of using biomaterials, they are unlikely to give the right answers when we need materials that are intended to optimally interact with their biological environment. This is likely to happen more in the future as medical technologies move in different directions and our understanding of biomaterials evolves.3 Our test procedures clearly have to evolve as well.

Professor Williams retired from the University of Liverpool, after 40 years, at the end of 2007. He retains the position of Emeritus Professor there and now has a series of professorial appointments in the USA, Australia, South Africa and China. In the USA he is Director of International Affairs for the Wake Forest Institute of Regenerative Medicine. He offers consulting services from his company Morgan & Masterson, based in Brussels, Belgium. He is Editor-in-Chief of Biomaterials, the leading journal in the biomaterials field.