Conducting Health-Based Risk Assessments of Medical Materials

Medical Plastics and Biomaterials Magazine
MPB Article Index

Originally published September 1998

TESTING

Although every medical manufacturer desires to construct its devices from entirely safe materials, the reality is that not all materials are entirely safe. Generally, if one looks long enough at small enough quantities, some type of risk can be associated with every material.

A risk assessment of nitinol devices supported their use in vascular implants. Photo: David Fukumoto; Nitinol Devices & Components, Inc.

Risk can be defined as the possibility of harm or loss. Health risk, of course, is the possibility of an adverse effect on one's health. Risk is sometimes quantified by multiplying the severity of an event times the probability the event will occur, so that

While this equation appears useful in theory, in practice it is difficult to apply to the biological safety of medical devices. The process known as health-based risk assessment attempts to provide an alternative strategy for placing health risks in perspective.

Standards and Guidances

A paradigm for the risk assessment process has been detailed in a publication prepared by the U.S. National Academy of Sciences.^1,2 Although devised primarily for cancer risk assessment, many of the provisions also apply to the assessment of other health effects. The major components of the paradigm are (1) hazard identification, (2) dosage-response assessment, (3) exposure assessment, and (4) risk characterization.³

This general approach to risk assessment was adapted to medical devices via the draft CEN standard Risk Analysis, published in 1993,⁴ and more recently via the ISO standard, ISO 14538—Method for the Establishment of Allowable Limits for Residues in Medical Devices Using Health-Based Risk Assessment, published in 1996.⁵ At the present time, FDA is also working to develop a health-based risk assessment protocol adapted to medical devices. Informally called the Medical Device Paradigm, the document is not yet generally available.^6,7

Some manufacturers may object that regulators are once again attempting to impose a "drug model" on medical devices. However, we shall see in the following pages that judicious application of these risk assessment principles can provide a justification for using materials that carry with them some element of risk, and that may, under traditional biocompatibility testing regimes, be difficult to evaluate or be deemed unsuitable for medical device applications.

Method

Hazard Identification. The first step in the risk assessment process is to identify the possible hazards that may be presented by a material. This is accomplished by determining whether a compound, an extract of the material, or the material itself produces adverse effects, and by identifying the nature of those effects. Adverse effects are identified either through a review of the literature or through actual biological safety testing.

Dose-Response Assessment. The second step is to determine the dose response of the material—that is, what is the highest weight or concentration of the material that will not cause an effect? This upper limit is called the allowable limit. There are numerous sources in the literature of data from which to determine allowable limits; some will be more applicable than others, and some may require correction factors.

Exposure Assessment. The third step is to determine the exposure assessment by quantifying the available dose of the chemical residues that will be received by the patient. This is readily done by estimating the number of devices to which a patient is likely to be exposed in a sequential period of use (for instance, during a hospital stay) or over a lifetime. For example, a patient might be exposed to 100 skin staples following a surgical procedure, or to two heart valves in a lifetime; thus, the amount of residue available on 100 skin staples or two heart valves would be determined.

Risk Characterization. Characterizing the risk constitutes the final step of the process. The allowable limit is compared with the estimated exposure: if the allowable limit is greater than the estimated exposure by a comfortable safety margin, the likelihood of an adverse event occurring in an exposed population is small, and the material may be used.

Case Studies

We can best get a sense of how these standards work by looking at some actual medical case studies that illustrate the risk assessment process.⁸

Nitinol Implant. Nitinol is an unusual alloy of nickel and titanium that features the useful property of "shape memory." A nitinol part can be given a particular shape at a high temperature, then cooled to a low temperature and compressed into some other shape; the compressed part will subsequently deploy to its original shape at a predetermined transition temperature. This feature is particularly beneficial for vascular implant applications in which the shape of the device in its compressed state eases the insertion process. The nitinol deploys as it is warmed by the surrounding tissue, expanding to take on the desired shape of a stent, filter, or other device. The transition temperature depends on the alloy's relative concentrations of nickel and titanium: a typical nickel concentration of 55–60% is used in medical devices, since this gives a transition temperature at approximately the temperature of the body (37°C).

Hazard Identification. One concern with using nitinol in implant applications is the potential release of nickel into the body. Although nickel is a dietary requirement, it is also highly toxic—known to cause dermatitis, cancer subsequent to inhalation, and acute pneumonitis from inhalation of nickel carbonyl, and to exert a toxic effect on cellular reproduction. It is a known sensitizer, with approximately 5% of the domestic population allergic to this common metal, probably through exposure from costume jewelry and clothing snaps. The biocompatibility question at hand is whether or not in vivo corrosion of nitinol releases unsafe levels of nickel.

Dose-Response Assessment. A search of the world medical literature revealed that the recommended safe level of exposure to nickel in intravenous fluids is a maximum of 35 µg/day.⁹ This value can be taken as an allowable limit of nickel exposure for a 70-kg (154-lb) adult.

The intravenous fluid data are based on subjects that are comparable to the patients who will be receiving nitinol implants. The data are for humans (not animals), for ill patients (not healthy workers or volunteers), and for similar routes of exposure (intravenous fluid and tissue contact). For these reasons, no safety correction factor need be applied to the allowable limit of exposure.

Exposure Assessment. The available dose of nickel from nitinol implants can be estimated from data found in the literature. In one study, dental arch wires of nitinol were extracted in artificial saliva, and the concentration of nickel measured in the supernatant. Corrosion reached a peak at day 7, then declined steadily thereafter. The average rate of corrosion under these conditions was 12.8 µg/day/cm² over the first 28 days.

Risk Characterization. A comparison of the available dose with the allowable limit for intravenous fluid levels shows that there is approximately a threefold safety margin, assuming that the implanted device is a full 1 cm² in surface area. (Devices with less surface area will contribute even less to the nickel concentration and have an even larger safety margin.) Considering the high quality of the data, a threefold safety margin is sufficient to justify using nitinol in vascular implants.

Wound-Dressing Formulation. Today's wound dressings are highly engineered products, designed to maintain the moisture content and osmotic balance of the wound bed so as to promote optimum conditions for wound healing. Complex constructions of hydrocolloids and superabsorbers, these dressings are sometimes used in direct tissue contact over full-thickness wounds that penetrate the skin layers.

Hazard Identification. There have been reports in the literature of patients succumbing to cardiac arrest from potassium overload, with the wound dressing as one of the important contributors of excess potassium in the bloodstream. The effects of potassium on cardiac function are well characterized. Normal serum levels for potassium are 3.8 to 4 milliequivalents per liter. As the potassium concentration rises to 5—7 mEq/L, a patient can undergo cardiac arrest and die. The biocompatibility issue to be explored is whether or not a wound-dressing formulation might release dangerous levels of potassium if used on full-thickness wounds.

Dose-Response Assessment. An increase of approximately 1 mEq/L of potassium is likely to provoke mild adverse events in most patients. Assuming that the average person's blood volume is 5 L, a one-time dose of 5 mEq of potassium may begin to cause adverse reactions. This value can be considered to be the allowable limit of potassium for most patients.

Exposure Assessment. Let us suppose that each dressing contains 2.5 g of potassium bicarbonate. Since the molecular weight of potassium bicarbonate is 100 g/mole, each dressing contains 0.025 mole of sodium bicarbonate, or 0.025 mEq of potassium ion. If a patient were to use four dressings in a day, the available dose of potassium would be 0.1 mEq/day.

Risk Characterization. Comparing the available dose of potassium (0.1 mEq) to the allowable limit (5 mEq) shows that there is a 50-fold safety margin. Considering that patients may be small in size, may have kidney impairment, or may receive potassium from additional sources such as intravenous fluids, this safety margin is too small, and so the dressing should be reformulated.

Perchloroethylene Solvent. A manufacturer of metal fabricated parts uses perchloroethylene to clean the finished pieces. Perchloroethylene has many advantages as a cleaner and degreaser: it is highly volatile, does not damage the ozone layer, and is very effective as a precision cleaning solvent (see Figure 1). The most common use of perchloroethylene is in the dry cleaning industry, but it is also commonly used in the electronics industry to clean circuit boards.

Figure 1. The chemical structure of perchloroethylene.

Hazard Identification. The downside of perchloroethylene is that it is highly toxic, with a material safety data sheet several pages in length listing adverse effects ranging from dizziness to death. Biocompatibility testing on solvent-cleaned parts would be meaningless; the solvent concentration on the part is so small that any effects of the solvent would be masked by the natural biological process of the test animals. The biocompatibility question that must be answered is whether or not sufficient residual perchloroethylene remains on the cleaned metal parts to pose a health hazard.

Dose-Response Assessment. Threshold limit values (TLVs) are values that indicate the maximum level of a chemical that a healthy worker could take in on a daily basis over the course of his or her work life without experiencing any adverse effects. The TLV for perchloroethylene is 50 ppm/day (50 ml of perchloroethylene per 10³ liter of air) by inhalation. The average person inhales 12,960 L of air per day, making this equivalent to 650 ml of perchloroethylene per day. Since the vapor density of perchloroethylene is 5.76 g/L, the TLV is equal to 3.7 g of perchloroethylene per day by inhalation.

Because TLVs for inhalation—as opposed to direct tissue exposure—are determined based on healthy individuals (not ill patients), we will divide the TLV by an uncertainty factor of 100, i.e., 10 to account for a different route of exposure and 10 to account for healthy-to-ill persons. By this method, we obtain an allowable perchloroethylene limit of 37 mg/day.

Exposure Assessment. To calculate an available dose of perchloroethylene, we need some additional information. In this case, the manufacturer brought a number of cleaned metal pieces into equilibrium within a closed jar, then analyzed the headspace above the pieces by using high-pressure liquid chromatography to determine the concentration of perchloroethylene released. The concentration of perchloroethylene was undetectable by high-performance liquid chromatography. Since the limits of this analytical method are 2 ppb, this value was taken as the concentration of perchloroethylene in the headspace. Taking the weight of the metal pieces, the number of pieces tested, and the volume of the headspace, it was calculated that the amount of perchloroethylene per single piece was a maximum of 1.0 ng/piece. If we suppose that a patient might be exposed to a maximum of 50 pieces over a lifetime, then the maximum available dose of perchloroethylene from the pieces would be 50 ng.

Risk Characterization. A comparison of the available dose (50 ng) to the allowable limit (37 mg/day) indicates an ample safety margin.

Ligature Material. A manufacturer purchases commercial, black fishing line to use as a ligature in a circumcision kit. Because the ligature is not "medical grade," a cytotoxicity test is routinely conducted as an incoming inspection test. It was assumed that a negative cytotoxicity test would be associated with an acceptable incidence of skin irritation.

Hazard Identification. A newly received lot of the fishing line failed the cytotoxicity test. The extraction ratio of this material—of indeterminate surface area—was 0.2 g/ml, with a 0.1-ml aliquot of sample extract being applied to a culture dish. Thus, 0.2 g/ml x 0.1 ml = 0.02 g represents a toxic dose of fishing line.

Dose-Response Assessment. A titration curve was obtained on the sample extract. If the sample was diluted 1:2, the test was still positive; however, if the sample was diluted 1:4, the test was negative. Thus, 0.02 g/4 = 0.005 g of fishing line, the maximum dose that is not cytotoxic. This value was called the allowable limit of fishing line.

Exposure Assessment. Each circumcision kit contained about 12 in. of line, but only about 4 in. of the material was ever in contact with the patient. Since an 8-yd line was determined to weigh 5 g, the available dose of fishing line was calculated to be 5 g/288 in. x 4 in. = 0.07 g.

Risk Characterization. A comparison of the available dose (0.07 g) with the allowable limit (0.005 g) convinced the manufacturer to reject the lot of fishing line.

Sources of Data

Data for calculating the allowable limit of exposure to a material can come from many sources, most of them promulgated by industrial and environmental hygienists and related agencies.¹⁰

Threshold Limit Values (TLVs) are time-weighted average concentrations of airborne substances. They are designed as guides to protect the health and well-being of workers repeatedly exposed to a substance during their entire working lifetime (7–8 hr/day, 40 hr/wk). TLVs are published annually by the American Conference of Governmental Industrial Hygienists (ACGIH).¹¹ Biological Exposure Indices (BEIs) are also published annually by ACGIH. These are the maximum acceptable concentrations of a substance at which a worker's health and well-being will not be compromised.

Other published guides include Workplace Environmental Exposure Levels (WEELs), from the American Industrial Hygiene Association; Recommended Exposure Limits (RELs), from the U.S. National Institute for Occupational Safety and Health; and Permissible Exposure Limits (PELs), from the U.S. Occupational Safety and Health Administration.^12–14 In the United States, PELs have the force of law.

Another important limit measurement, Short-Term Exposure Limits (STELs), are defined as the maximum concentration of a substance to which workers can be exposed for a period of up to 15 minutes continuously, provided that no more than four excursions per day are permitted, and with at least 60 minutes between exposure periods. The STEL allows for short-term exposures during which workers will not suffer from irritation, chronic or irreversible tissue damage, or narcosis of sufficient degree to increase the likelihood of injury, impair self-rescue, or materially reduce work efficiency. Some substances are given a "ceiling"—an airborne concentration that should not be exceeded even momentarily. Examples of substances having ceilings are certain irritants whose short-term effects are so undesirable that they override consideration of long-term hazards.

Uncertainty Factors

An uncertainty factor is a correction that is made to the value used to calculate an allowable limit. It is based on the uncertainty that exists in the applicability of the data to actual exposure conditions. Typically, uncertainty factors range in value from 1 to 10. For example, a correction factor of 10 might be applied for data obtained in animals rather than humans, or to allow for a different route of exposure. In other words, for every property of available data that is different from the actual application, a correction factor of between 1 and 10 is applied. If our first example had been of a small amount of data obtained in animals by a different route of exposure, an uncertainty factor of 1000 might be applied.

Safety Margins

A safety margin is the difference or ratio between the allowable limit (after correction by the uncertainty factor) and the available dose. How large does a safety margin need to be? Generally, a safety margin of 100x or more is desirable, but this can depend on the severity of the risk under consideration, the type of product, the business risk to the company, and the potential benefits of product use.

Conclusion

Medical device manufacturers have two predominant questions when it comes to material biocompatibility. The first is: "We have a material that we absolutely must use in our device, but it fails a biocompatibility test. Can we justify using the material anyway?" The second is: "We have a material that we absolutely must use in our device, but carcinogenicity and/or chronic toxicity testing are required. Can we justify omitting these tests?" Judicious application of health-based risk assessments can help with both of these issues, often providing a fast, cost-effective answer to both questions.

References

1. Risk Assessment in the Federal Government: Managing the Process, Washington, DC, National Research Council, 1983.

2. Hays AW, Principles and Methods of Toxicology (3rd ed), New York, Raven Press, pp 26–58, 1994.

3. Ecobichon DJ, The Basis of Toxicology Testing, Boca Raton, FL, CRC Press, 1992.

4. CEN BTS 3/WG 1—Risk Analysis is available through the British Standards Institute.

5. Available from the Association for the Advancement of Medical Instrumentation, 3330 Washington Blvd., Ste. 400, Arlington, VA 22201.

6. Draft copies of the Medical Device Paradigm may be obtained by contacting Dr. Melvin Stratmeyer, FDA Center for Devices and Radiological Health, HFZ-112, Division of Life Sciences, Office of Science and Technology, FDA, Rockville, MD 20857.

7. Brown RP, and Stratmeyer M, "Proposed Approach for the Biological Evaluation of Medical Device Materials," in Proceedings of the Medical Design and Manufacturing East 97 Conference and Exposition, Santa Monica, CA, Canon Communications, pp 205-9–205-18, 1997.

8. Stark NJ, "Case Studies: Using the World Literature to Reduce Biocompatibility Testing," in Proceedings of the Medical Design and Manufacturing East 97 Conference and Exposition, Santa Monica, CA, Canon Communications, pp 205-1–205-7, 1997.

9. Stark NJ, "Literature Review: Biological Safety of Parylene C," Med Plas Biomat, 3(2): 30–35, 1996.

10. Hayes AW, Principles and Methods of Toxicology (3rd ed), New York, Raven Press, pp 366–367, 1994.

11. American Conference of Governmental Industrial Hygienists, 1331 Kemper Meadow Dr., Cincinnati, OH 45240.

12. American Industrial Hygiene Association, 2700 Prosperity Ave., Ste. 250, Fairfax, VA 22031.

13. National Institute for Occupational Safety and Health, Hubert H. Humphrey Bldg., 200 Independence Ave. SW, Rm. 715H, Washington, DC 20201.

14. Occupational Safety and Health Administration, U.S. Department of Labor, Washington, DC 20210.

Nancy Stark, PhD, is president of Clinical Design Group, Inc. (Chicago), a consulting and contracting firm for medical device safety, efficacy, and performance. The company frequently performs health-based risk assessments for device manufacturers.

Copyright ©1998 Medical Plastics and Biomaterials