Originally Published MDDI January 2006
It is often difficult to identify the reasons implanted devices fail. But certain testing techniques can provide insights into failure causation.
By Dan Mazzucco, Anton Bowden, and Kevin Ong
The implantable medical device industry has grown exponentially over the last 30 years. Since 1976, when FDA began regulating the sale of implantable medical devices, the number of approved medical devices has dramatically increased.
Analyzing medical device design to improve implant performance presents a unique challenge to engineers and clinicians. Unlike products created for use outside the body, implants are tough to test in a realistic environment. Products need to be validated in ever more lifelike scenarios, from computer simulation, to benchtop testing, to an animal model, to, finally, clinical tests. For implants, the jump from animal to human can be particularly challenging. It is often hard to simulate human geometry and loading conditions in an animal model. In addition, testing becomes particularly challenging when bringing a new product into a market filled with relatively successful competitors. Doctors and patients are not likely to sign up for an unproven procedure or product if proven alternatives exist. And, if the consequences of product failure can be fatal, adoption is even less likely.
Once a new product's safety has been validated in an animal model, the device is implanted in humans during clinical testing. Usually, these tests quickly weed out products that do not measure up to existing standards of care. But for products that are expected to interact with the body for long periods of time, it may take 10 years or more to reveal subtle improvements in product performance. Nonetheless, the potential benefits of improved quality of life and reduced healthcare costs motivate FDA and industry to expedite market introduction of new implants. And in a competitive market, delivering a reliable and innovative device to market first can give a manufacturer an edge in market share. For these reasons, FDA and industry may bring implants to market before proof testing is completed.
FDA and industry also monitor devices that are implanted and respond when negative trends emerge. FDA has developed the Manufacturer and User Facility Device Experience (MAUDE) database to monitor the performance of medical devices. This database is often the best source available to use to evaluate the market performance of a medical device. However, there is a potential for typographical errors and underreporting, and there is often a time lag between an event and its publication in the database.
Even implants with very high success rates can be costly to both the individual and society when they do not succeed. For example, total knee replacement (TKR) is widely considered as “very successful” by the orthopedic community.1 Ninety percent of knee implants last for 10 years without needing surgical intervention. It is difficult to make a toaster that works for 10 years, yet we expect such longevity of TKRs. But if revision surgery is needed, the societal cost is huge—the annual Medicare reimbursements amount to $204 million.2 This cost to society is compounded by our litigious culture, in which undesirable device outcomes often lead to expensive legal disputes. And these societal costs pale in comparison with the human cost of a failed medical device. One in 200 TKR patients dies in the days following surgery.1 Those odds are fairly intimidating for an elective surgery.
Further complicating matters is that the device alone does not decide clinical outcome. Surgical techniques play a critical role in a product's success, as do patient factors, including biological response, comorbidities, fear, and pain. This means that a thorough review of a patient's physical history alone may not be able to pinpoint a device's role in a poor clinical outcome. More specifically, revision surgery, explantation, or other poor outcomes are possible even for a well-designed device built to specification.
Many failure modes are possible, and the events leading to failure may not be obvious upon examination. Identifying causes of failure is challenging because the incidence of removal is relatively rare. As a result, those interested in improving clinical outcomes (e.g., manufacturers, clinicians, and researchers) use many techniques to extract as much data as possible from those retrieved implants. The following two case studies showcase several useful tools for postretrieval analysis.
Retrieved Total Disk Replacement
Artificial-disk technology has been used in Europe for about 20 years, but these devices have only recently been cleared for use in the United States. In October 2004, the Charité total disk replacement (TDR) from DePuy Spine Inc. (Raynham, MA) became the first FDA-approved artificial disk. Several other TDR makers are at various stages of the clinical trial or FDA review process. Despite the attention given to these disks in the popular press, there have not been many detailed studies of retrieved devices. One area of particular concern is the long-term wear characteristics in the components of these devices. The response of spinal tissue to particulate matter is an area of ongoing investigation. There has been little quantitative information presented in the scientific literature regarding the quantity of wear debris. Additionally, stress levels in total disk arthroplasty have not yet been studied extensively. Stress levels are closely associated with fatigue wear and surface damage of hip and knee replacement components.
A 49-year-old woman was implanted with a Charité TDR. About three years later, the implant was removed because of intractable lower-back pain and intermittent left leg and buttock pain. Radiographic evidence showed gaps in the bone-metal interface between the TDR end plates and the vertebral end plates. The firm that analyzed the TDR tried to investigate the failure mechanisms of the device and to retrospectively examine its in vivo performance. To accomplish this task, the firm used several analytical tools.
The first tool was microcomputed tomography (µCT). CT has extensive clinical use; on a smaller scale, it can be used to characterize retrieved implants. It generates 3-D maps with 10- to 75-µm resolution. This technology has been used in the past to characterize bone in an in vitro setting. But recently, it has also been used to characterize materials as varied as polyethylene acetabular cups and metallic stent struts. This method is nondestructive, and the only permanent change to the material comes from a small dose of radiation (about as much radiation exposure as one would get from a transatlantic flight). Using this imaging technique, the testing firm's analysts could visualize cracks and plastic creep both along the surface and through the volume of the core. They also observed pit formation around the edge of the component, where the core and rim interacted with the metal end plates of the implant.
Other microscopic evaluation tools included interferometry and Fourier transform infrared (FTIR) spectroscopy. The average roughness and waviness of the plastic component was evaluated using white-light interferometry. By measuring and examining alterations in machining marks, analysts can determine the presence or absence of adhesive or abrasive wear. The analysis showed only minimal wear at the central apex of the device. FTIR spectroscopy determined the chemical signature of the implant, which showed that degradation by oxidation had occurred near the surface of the implant, but not in the interior. Light microscopy performed on 200-µm-thick slices confirmed the crack geometry and directionality observed by µCT.
Finally, the analysis firm developed a detailed finite-element model of the device to study the stress magnitudes and distributions in the TDR. It obtained the geometry of the implant core from the undamaged portion of the device, as represented in the µCT data. The geometry of the metal end plates was obtained using 3-D point digitization. Using mechanical testing, the firm found material property data for the implant core. The data were then modeled numerically using a validated nonlinear plasticity model. The firm exercised the model under direct compression loading and compression and flexion loading. Results from the analysis showed that the maximum effective stress in the core coincided with observed plastic deformation and fracture patterns in the retrieved device (see Figures 1 and 2). Also, the stress magnitudes were similar to those typically observed in TKRs.
The data gained from these analyses helped determine the failure modes for this particular implant. It also helped the firm examine other likely failure modes for similar implants. OEMs can now apply that knowledge toward the design of next-generation TDR implants.
Retrieved Total Hip Replacement
Unlike TDR, total hip replacement (THR) has been in clinical use for almost half a century. The clinical performance of THRs is affected by a combination of factors. Design factors (e.g., cup shape and size), patient factors (e.g., activity level, weight, and bone quality), and surgical factors (e.g., extent of reaming and implant orientation) all affect performance. Furthermore, the longevity of THRs is limited by implant wear. Ultimately, wear induces complications, such as implant loosening and osteolysis, and results in revision surgery. In the United States, the annual number of THR revisions increased almost twofold between 1990 and 2002, from 24,000 to 43,000.3
Consequently, retrieved THR implants are much more common than retrieved TDR implants. The example that follows is typical of a retrieved THR implant. A 5-ft, 7-in.-tall, 205-lb man with degenerative joint disease received a left hip replacement at age 65. Five years postoperation, the patient had left hip pain after a fall. One and a half years later, the polyethylene component was removed and replaced because of wear. The removed implant, which was retained and sterilized after revision surgery, was then analyzed.
One of the most common methods of evaluating a retrieved polyethylene component is mechanical characterization using the small-punch test. The small-punch test evaluates a small (6.4-mm-diam, 0.5-mm-thick), disk-shaped section of material by pushing a hemispherical indenter through its center. Typically, the small-punch test is performed on a section taken from the surface of a specimen and also one from the subsurface. The test shows whether the implant has retained its mechanical properties through its time in storage and through service time in vivo. In this example, small-punch testing of specimens taken from the worn and unworn regions of the THR showed no differences in ultimate and peak loads. FTIR spectroscopy showed that oxidative degradation occurred near the surface of the implant, but not in the interior.
Assessing wear depth, wear volume, and morphology of retrieved devices can also provide useful insight into the design and performance of the implants, as well as guide future design development. Advances in volumetric wear evaluation in retrieved acetabular components have been developed and validated using high-resolution µCT in concert with image registration techniques.4 A Scanco µCT 80 high-resolution scanner, made by Scanco USA Inc. (Southeastern, PA), was used to image the retrieved hip implant at 74-µm resolution. After image registration of the implant, wear volume was measured by image data thresholding and subtraction. The measured wear volume was 526 mm3, corresponding to a volumetric wear rate of 81 mm3 per year. That rate is consistent with reported wear rates for similar cup designs. The wear pattern on the cup indicated a small amount of rim damage, probably caused by impingement with the neck of the femoral stem. Previous retrieval studies of acetabular components also have reported a high incidence of impingement.5
Preclinical evaluation of THR implant designs can also be performed using computational models of the pelvis. For example, realistic geometry and material properties for the finite-element models were obtained from CT scans of a cadaver pelvis. After simulating various joint loads, bone quality, cup implant positions, and reamed acetabular geometry and surface conditions in the model, the extents of cup fixation and of migration were computed as functions of cup size and shape. Based on these analyses, cup stability in terms of cup migration was reduced for implants with 2 mm of press-fit at the rim. That particular design is consistent with the recommended 1- to 2-mm press-fit for uncemented components. A press-fit of that size had been employed during surgical implantation of the retrieved hip replacement.
Using the tools and techniques described in this article, the analysis firm retrospectively quantified the extent of implant damage and compared it with values reported in literature for well-functioning and retrieved devices. The small amount of rim damage to this particular component may have been initiated by the fall that the patient suffered prior to the revision surgery.
These two case studies demonstrate the breadth of data that various testing procedures elicit. Depending on the information required, a combination of destructive and nondestructive techniques can answer relevant questions about mechanical and chemical performance, failure modes, and the circumstances leading to implant failure. The information these analyses generate can guide clinical decision making and product development, ideally leading to improved product performance.
1. “NIH Consensus Statement on Total Knee Replacement,” NIH Consensus and State Science Statements 20, no. 1 (2003): 5.
2. F Mowat et al., “Economic Burden of Hip and Knee Arthroplasty Procedures in the Medicare Population,” in Transactions of the 51st Annual Meeting of the Orthopaedic Research Society (Washington, DC: Orthopaedic Research Society, 2005), 1148.
3. S Kurtz et al., “Prevalence of Primary and Revision Total Hip and Knee Arthroplasty in the United States from 1990 through 2002,” Journal of Bone and Joint Surgery: American 87, no. 7 (2005): 1487–1497.
4. A Bowden et al., “Validation of a Micro-CT Technique for Measuring Volumetric Wear in Retrieved Acetabular Liners,” Journal of Biomedical Materials Research, Part B: Applied Biomaterials 75, no. 1 (2005): 205–209.
5. W Shon et al., “Impingement in Total Hip Arthroplasty: A Study of Retrieved Acetabular Components,” The Journal of Arthroplasty 20, no. 4 (2005): 427–435.
Dan Mazzucco is a biomedical engineering consultant at Exponent Inc. (Philadelphia). Contact him at [email protected]. Anton Bowden is an applied biometrics consultant at Exponent and can be contacted at [email protected]. Kevin Ong is a biomedical engineering consultant at the firm and can be contacted at [email protected]
Copyright ©2006 Medical Device & Diagnostic Industry