Welsh Dragon Silhouette - Wikimedia Commons

Morgan & Masterson Consulting

Biological Safety and Risk Analyses

David F Williams, PhD, DSc, FREng, FLSW
Author, Scientist & Consultant
David F Williams
Author, Scientist & Consultant

Biological Safety and Risk Analyses

Most regulatory authorities require evidence of biological safety of medical devices, such evidence being provided by successful completion of standard test procedures, which are directed towards satisfactory performance related to a series of biological end points. These end points may include cytotoxicity, sensitization, acute and chronic toxicity, thrombogenicity, mutagenicity, carcinogenicity and reproductive toxicity. The precise tests that are used for such determinations are not usually mandated by regulators, but it is usual, and indeed sensible, to follow the procedures recommended by the International Standards Organization (ISO) in their ISO 10993 series. The selection of tests to be used for any new submission is not normally problematic and there are many commercial test houses that can advise on this, and carry out such tests under GLP conditions. However, there are some contentious issues, where a much deeper knowledge of biocompatibility and toxicology is required, some of which are covered here.

Interpretation of ISO 10993

The fact that the standard tests for biological safety are not mandatory, although strongly recommended, indicates that there is some latitude about their interpretation. Even with the simplest of the tests described in ISO 10993, that for cytotoxicity, there are several possible procedures that can be used. In the most popular, which determines the effects of biomaterials extracts on cells in culture, a material is considered to pass the test (i.e., is non-toxic) if it is graded 2 or less, but fails the test (i.e., is toxic) if it has a grade more than 2. A grade of 2 is defined as having not more than 50% of the cells being rounded, devoid of intracytoplasmatic granules, no extensive cell lysis and not more than 50% growth inhibition. Grade 3 has not more than 70% of cell layers containing rounded cells or are lysed, the cell layers not being completely destroyed but not more than 50% growth inhibition. It is easy to see how subjective this process can be. Similar situations arise with other end points, and borderline cases need expert judgement as to whether there is an unacceptable biological risk.

Questions also arise when there is no adequate test to determine performance with respect to certain endpoints. Two important biological responses may be mentioned, those of carcinogenicity and thrombogenicity. The former refers to the possibility that biomaterials could be the cause of cancer. Most regulators require complete investigations of mutagenicity testing as a type of surrogate for carcinogenicity but, from a biological perspective it is possible for non- mutagenic substances to give rise to cancer, so there are residual concerns. The problem is that there are no reliable animal models for implant-induced tumor formation – when I carried out animal experiments in my Liverpool laboratories. I could produce tumors with all currently used biomaterials in rats. The justification of claims that biomaterials are not carcinogenic requires careful crafting. With respect to thrombogenicity, there are standard in vitro tests for effects on the blood clotting cascades and on platelet function, but their relevance to intravascular devices is far from clear, and all test data has to be assessed by those experienced in the field.

Chemical characterization and toxicology

It has been a long-held view that if you know the exact composition of a material, you should be able to understand the toxicological profile of that material by close examination of the individual constituents. This is the basis for the recent surge in interest in a full chemical characterization of biomaterials, now that analytical techniques, such as ICP-MS, HPLC, FTIR and GC-MS, are in common use and have remarkable sophistication in detection and quantification of such constituents. The ISO 10993 series now has a section dealing with chemical characterization of biomaterials, or to put it more correctly, the characterization of extracts from biomaterials. Sample of the material may be placed in pure water and / or a variety of solvents, and a large array of analytical techniques used to define any molecules or ions released into the medium. So far, so good; the question then arises, however, as to what you do with that data. I have worked with several medical device companies on this difficult issue, with materials such as silicone elastomers and polyurethanes. Sometimes, more than 100 species are ‘detected’, usually at miniscule levels, and often without precise molecular information. It is then necessary to consider each constituent in turn and to evaluate, on the basis of published literature and databases, the potential risk related to all possible forms of toxicity. A major difficulty arises when general toxicological data has been obtained by routes such as inhalation or ingestion, which bear no relationship to implantation or extra-corporeal circulation. These situations are very challenging.

Gap analysis

A term that is being used more and more frequently in the determination of biological safety is ‘gap analysis’; in the present context this means the identification of gaps in knowledge about the biological performance of a device, where filling of those gaps should facilitate regulatory approval. It is difficult to make generic comments about what might seem a nebulous area, but one thing that a consultant has remember here is that, rather like carrying out an FMEA exercise, all possible biological endpoints have to be considered, along with a subsequent consideration of whether it is in the client’s interest to fill those gaps experimentally, which could take time and be very costly, or to justify to the regulator why there are such gaps, or whether to advise the client that such gaps render the project economically non-viable 

Failure analysis

In spite of the best endeavors of both the company and the regulator, devices may ultimately ‘fail’ in some patients. If these events are rare, or minor, or just become detectable on the examinations of performance through post-market surveillance, the situation may be readily addressed through routine company protocols, perhaps necessitating a CAPA process, which in a Corrective And Preventive Action system. Once again these should be simple, not requiring external advice. Occasionally, however, the failure may be of such a magnitude, either in the total number of cases or the serious injuries or deaths that result from this, that an extensive failure analysis process has to be initiated, often mandated by, and controlled through the home regulatory body. In the USA, most of these cases will inevitably end with product liability litigation, which I will cover in the section on Expert Witness.

On rare occasions, the investigation has to be more profound than determining fault and settling claims for damages. This is usually because there may be many patients (perhaps world-wide) who have been implanted with a device that is now being shown to be associated with a ‘defect’, which could cause death or injury on a significant scale. I have been involved with several such scenarios. In one case, which I can mention since this was long ago and all claims settled, a mechanical heart valve was shown to fail in a few patients after several years function. By the time this was realized, over 80,000 patients had received the valve. The critical question concerned the decisions as to whether such patients should have their valves surgically removed and replaced with another prosthetic valve. A group was established, approved by the Board of the company and the FDA, with responsibility of determining the characteristics of those patients who should be considered for re-operation. Of significance was the fact that statistically there was an overall risk of death at reoperation of 6%; the risk of valve fracture overall could not be assessed since this was a time-dependent fatigue process. As a metallurgist, my role was to determine the precise mechanism of failure and identify those valves (size, date and place of manufacturer, the name of the welders and other manufacturing personnel etc.) most at risk, while other team members, who were biostatisticians, epidemiologists, cardiac surgeons and so on, determined the patients most at risk. My report was very detailed, and has never been made public, but I was very satisfied with the outcome of that consulting experience.

Morgan & Masterson Consulting

We offer global, high-level consulting for scientific and infrastructure aspects of cutting-edge medical technologies.