However, that costs time and money, and the document looks the same as the one on the FDA Web site.
Due to the lack of easily obtainable information, an engineer or regulatory affairs officer may decide to design tests that will meet FDA requirements and ISO matrices. That’s a costly and time-consuming misconception. ISO 10993 has documents beyond its’ Part One that specify how to perform required tests.
Initially, these requirements were evaluated based on testing procedures specified by the United States Pharmacopeia (USP). They continue to evolve to satisfy the testing requirements of as many countries as possible. After all, the purpose of ISO is to provide one set of standards that medical devices or materials must meet for worldwide approval. At this point, both the USP and the American Society for Testing and Materials (ASTM) have generated their equivalents of ISO 10993-1. One force driving this evolution is the Japanese Ministry of Health, Labor, and Welfare. Early versions of ISO 10993 were relatively loose and allowed a lot of room for interpretation. Previous guidance documents for the Japanese market were very stringent, with no room for interpretation.
The safety of products such as hyaluronic acid may be evaluated by injecting them in a clinically relevant tissue or site, such as a cross-section of knee tissue as shown here.
The current Japanese guidance document, titled, “Test Methods for Biological Safety Evaluation of Medical Devices, Assessment of Medical Device, Notice 36” (2003), partially embraces ISO 10993, but it has technical specifications to perform tests that differ from the procedures specified in ISO 10993. As a consequence, many Japanese regulators do not readily accept testing conducted according to ISO 10993. Such reluctance will more than likely continue until the technical sections of these two standards are harmonized. The technical conduct of many of the studies differ enough that the endpoint requirements of a Japanese test don’t fulfill the requirements of ISO and vice versa. Therefore, be prepared to perform additional testing if you intend to submit your product for Japanese approval.
It is said that a little knowledge is a dangerous thing. The best course of action is to contact a contract research organization (CRO) that specializes in biocompatibility testing and performs these tests according to good laboratory practices (GLPs). This step will be the most prudent use of your time and will give you an idea of the road ahead. You will also want to get FDA agreement on your device classification and testing scheme if you haven’t already done so.
Describe in accurate detail to the CRO what your product is and how it will be used. Allow the experts there to discuss what tests they can offer to meet the requirements. A fully equipped CRO should have a standard battery of tests to meet both ISO and Japanese requirements. Refrain from designing your own study, which might or might not fulfill the requirements of multiple testing categories. This approach will wind up costing more money and may not be accepted by regulators unless agreed to in advance with a pretest meeting.
Consider the following three important questions to ask the CRO: How much sample do I need to provide? How long will testing take? And, most importantly, how much is this going to cost? An experienced CRO may ask you similar questions, e.g., how much sample can you provide and when do you intend to file the submission?
Depending on your material or device, the sample requirements may seem excessive, but they are what the CRO needs to conform to the regulations. Don’t be discouraged by the sample quantities being requested. The CRO will advise you on the requirements of the guidelines, but ultimately will test the product in whichever way you specify. If you manipulate the sample requirements, be prepared to present justification to regulators for doing so. In a case in which altering the sample requirements is justifiable, the company applying for the approval will have to stand behind any attempt to alter requirements. Be advised that the cost of goods alone doesn’t offer much of a justification, particularly in Japan.
If the need for biocompatibility testing came as a surprise to you, there is a good chance you are finding out about it late in the process. Test systems for biocompatibility, whether in vitro or in vivo, have a shelf life. The in vivo assays use animals within very specific weight ranges, which fall in the middle of the animal’s growth curve. The in vitro assays use the cells when they are at a certain confluency or density. Therefore, the tests need to be timed to start when the animals are within the proper weight range or the cells are at the proper density. This choreography of “growing up cells,” or acquiring animals, isn’t going to start until the CRO receives the test article. The end result is that there will be a lag from the time you send the sample to the time the testing actually starts. This lag will vary from lab to lab and test to test. However, the actual test time requirements do not change. In this context it is important to reiterate that a 72-hour test takes 72 hours and a 26-week test takes 26 weeks to complete. After the actual test time, there may be additional evaluations, such as histopathology or clinical chemistry. All of the data must then be evaluated and packaged by the study director in the form of a report.
Sample Preparations and Material Requirements
For biocompatibility testing using cytotoxicity (ISO 10993-5), the test samples are either tested directly, as in an agar overlay, or they are extracted, as in the minimum essential medium (MEM) elution. The extraction is a process in which the test material is typically subdivided, placed in an extraction vessel, and covered with the extraction vehicle. Polar and nonpolar extraction vehicles are to be used separately. Examples of polar extraction vehicles would be water, cell culture media (as an MEM), and physiologic saline (0.9% NaCl). Saline is the preferred polar vehicle for biological assays. Examples of nonpolar vehicles would be cottonseed oil and sesame seed oil.
When you are required to do biocompatibility testing, the test should be performed on the finished product. The sample should go through any cleaning, processing, polishing, sterilization, etc. that the final product will go through. Occasionally, the test sample won’t lend itself to extraction, possibly because of its size or configuration. Sometimes the sponsor may prepare a scaled-down mock sample. The mock sample has to go through all the processes and have all the components in the proper proportions as the actual device. This may require retooling your whole production line, which may not be worth the potential material savings.
Careful consideration should be given to each test article to determine an appropriate approach to the testing. For example, some devices may be a thin, sealed titanium can containing electronics. Extracting this device with the electronics contained within would not be appropriate because there is little to no likelihood that a patient will be exposed to the electronics. Extracting the device with the electronics contained would add mass to the test article and, therefore, increase the amount of vehicle the can is exposed to. The result is a dilution of any minor extractants leaching from the can, and potentially masking any toxic response.
Likewise, in the case of hemodialyzers, the fibers will increase the volume of vehicle to the test article and mask the results of components that are present in lesser amounts. The fibers and housing should be tested separately, with the fiber-adhesive junction tested as part of the housing.
The material requirements vary from test to test, particularly if you are doing direct-exposure tests. If you are doing extractions, most tests can be performed with 10 ml of extract per vehicle per extraction event. Exceptions to that general rule include the rabbit pyrogen test and repeat dose studies. Other studies may require more sample depending on the protocol. The extraction ratios have gone through some alterations over the years but keep coming back to the same ratios.
Fortunately, one area in which the world of biocompatibility is in harmony is in the extraction ratios. The ratio for test articles with a thickness less than 0.5 mm is 6 cm2 per 1 ml of vehicle; test articles with a thickness of 0.5 mm or greater require 3 cm2 per 1 ml of vehicle. Therefore, depending on the thickness, a 10-ml extract would require 60 or 30 cm2 of test material. Surface area ratios should be the primary method of preparing extracts. Regulators around the world expect the submitters to provide testing that conforms to these ratios. If a fluid pathway or a clinically relevant extraction is used, you will be required to back-calculate and show that the surface area–to-volume ratios met or exceeded the specified ISO ratios. Fluid pathway may meet or exceed the ratios, whereas clinically relevant ratios will likely fall short of the specified ratios the product will need to be retested. These approaches should be avoided.
To put that concept into perspective, compare it with an empty manila folder, the cover of a paperback journal, or a staple from a standard desk stapler. These all have a thickness of about 0.5 mm. The typical small office sticky note measures 3.5 × 4.8 mm and has a single-side surface area of 16.8 cm2.
When the surface area is calculated in the laboratory, all sides are calculated; so the sticky note has a total surface area of
33.6 cm2. The thickness of a sticky note is about 0.1 mm. Therefore, two sticky notes would be used to create 11.2 ml of extract.
Surface area–to-volume ratios are the preferred method of creating the extract. But when the geometry is complex, a weight-to-volume ratio of 0.2 g per 1 ml is used. The definition of complex geometry varies according to perspective. The CRO may consider only tubings, slabs, cylinders, disks, and rings, and are unlikely, for example, to calculate the surface area of a porous device or screw. Without your intervention, the CRO will usually proceed with extracting at a weight-to-volume ratio. If you have calculated surface areas in your computer-aided design (CAD) drawings and the CRO used the weight-to-volume ratio, the submission may be rejected and get sent back to you because of incorrect or inconsistent extraction procedures. So, if you have used CAD to calculate surface areas, provide those data to the CRO.
When a device is complex and composed of various materials, the weight-to-volume ratio may be the more appropriate method to prepare the extract. This way, the volume-to-material ratio remains consistent with the finished product. For example, a simple three-material catheter may have the shaft made from one material and the hub made from another, plus the adhesive that holds the two together. All such materials are present in the finished device in differing thicknesses.
Even though the surface areas may be easily calculable, a surface area–to-volume ratio could be inappropriate for such a case. The shaft has far more surface area than the hub, but the hub is far thicker, and there is far more material in the hub than the shaft. Then there is the interface between the shaft and the hub, where the adhesive is present. To extract all the components at a consistent ratio, a weight-to-volume ratio should be performed. In this example, the whole device would be extracted. But larger devices need to be disassembled and sorted by material. Then the extract sample is prepared with each material present in the same percentage as it exists in the finished device.
The prepared samples are combined with the appropriate polar solvent in one vessel and the appropriate nonpolar solvent in another vessel. The vessels are then placed in an oven at 37°, 50°, or 70°C at 72, 72, or 24 hours, respectively, or an autoclave at 121°C for one hour. The temperature chosen should be the highest possible that doesn’t degrade or deform the test article. The temperature and time points should be your decision to direct the CRO. Unless the test article contains a biologic or pharmaceutical that is going to degrade at 50°C, stay away from choosing 37°C for 72 hours. Even though this is currently specified in the ISO documents as an acceptable choice, it’s possible that the low temperature may require a 96-hour incubation time somewhere in the future.
The first part of this primer for medical device biocompatibility testing describes a few ways to avoid potential pitfalls that could delay a product launch. It is essential to follow strict testing protocol; don’t waste valuable resources planning your own criteria. Be sure to start the process early enough to allow for thorough and complete testing. As with any business or scientific endeavor, communicate all details accurately; and, to be as up to date as possible on all requirements, consult other qualified professionals to provide expertise. The second part of this primer will deal with issues such as genotoxicity, sensitization, hemocompatibility, and implantation.
“Biological Evaluation of Medical Devices,” ISO 10993, parts 1–12 (Geneva: International Organization for Standardization, various dates).
“Use of International Standard ISO 10993, Biological Evaluation of Medical Devices—Part 1: Evaluation and Testing” G95-1 (Rockville, MD: Department of Health and Human Services, FDA, 1995).
“Testing Methods to Evaluate Biological Safety of Medical Devices, Notice from the Office Medical Devices Evaluation Number 36” (Pharmaceutical and Food Safety Bureau, Ministry of Health, Labour and Welfare, March 19, 2003).
“<87>Biological Reactivity Tests In Vivo.” United States Pharmacopoeia, most recent version.
“<88>Biological Reactivity Tests In Vitro.” United States Pharmacopoeia, most recent version.
“<1031> The Biocompatibility of Materials Used in Drug Containers, Medical Devices and Implants.” United States Pharmacopoeia, most recent
“Standard Practice for Selecting Generic Biological Test Methods for Materials and Devices.” ASTM F748-06.
Laurence Lister is director of biocompatibility services at Toxikon Corp. (Bedford, MA).