- Email: [email protected]
The Clinical Relevance of Hip Simulator Testing of High Performance Implants
The Clinical Relevance of Hip Simulator Testing of High Performance Implants Aiguo Wang, PhD, Aaron Essner, MSc, and Joe Cooper, BA Effective in vitro hip simulation testing should reproduce clinical failures, clinical wear mechanisms, and clinical successes. These validated test outcomes are the only criteria for predicting the failures and successes of new bearing surfaces and geometries. The authors have successfully reproduced the clinical failures of Charnley’s polytetrafluoroethylene (PTFE) and Hylame polyethylene, and successfully predicted clinical successes, including first generation highly crosslinked UHMWPE (Crossfire) over gamma-inert sterilized conventional UHMWPE (N2vac). When approached with new implant technologies, surgeons must know how to evaluate the validation of testing methods used by manufacturers, to ensure the test outcomes of these products predict good survivorship. Semin Arthro 17:49-55 © 2006 Elsevier Inc. All rights reserved. KEYWORDS wear, simulation testing, high performance, hip arthroplasty
T
oday’s total hip arthroplasty patients comprise a changing population that grows younger and more active every year. As patients’ expectations increase, so does the demand for new bearing technologies capable of long-term wear under harsh conditions. Testing is a crucial phase in the development of these new technologies, and there are many parameters affecting the validity and predictive potential of implant testing. Any new bearing technologies must be subjected to rigorous and exacting in vitro testing, to meet the implant demands of present and future patient populations. The clinical relevance of this testing is critical. There is a long clinical history of bearing material failure in acetabular components. A study by Min and coworkers on the Harris-Galante cup (Harris-Galante II, Zimmer, Warsaw, IN) revealed a 17.3% failure rate of the polyethelene liner after 8 years.1 Previous clinical failures include the early failure of Hylamer polyethelene (DePuy, Warsaw, IN) in total hip2 arthroplasty. Teflon and polyester are two other bearing surfaces used in total hip arthroplasty, both having poor clinical histories.3 New bearing technologies face not only the failures of past generations, but also the increasing patient demand for better implant performance. Disease also challenges implant perforBiomechanics and Tribology Laboratory, Stryker Orthopaedics, Mahwah, NJ. Address reprint requests to Aiguo Wang, PhD, Science and Clinical Studies, Stryker Orthopaedics, 325 Corporate Drive, Mahwah, NJ 07430; [email protected].
1045-4527/06/$-see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1053/j.sart.2006.06.005
mance, such as obesity, as well as diseases linked to obesity such as diabetes and hypertension.4 Other challenges include suboptimal implant placement (sometimes because of new trends or surgical technique) such as minimally invasive surgery (MIS), and establishing realistic patient expectations. Laboratory in vitro simulation testing must reproduce the mechanisms, rate, and total magnitude of wear. The ultimate goal of these reproductions is the accurate prediction of the clinical failure or success of bearing surfaces and design. As bearing surface technologies advance to meet increasing patient expectations, a validated simulation testing model with a clinical history of predictive success must be used to effectively analyze and develop these technologies. The objective of this paper is to review the criteria for effective simulation testing and provide the reader with the background necessary for the investigation of the testing methods of new implant technologies.
Wear Study Parameters When designing a wear study, several investigational parameters must be considered. Lubrication is critical to the simulation of physical joints, and its composition is a determinant of the testing result. Lubricants for simulator testing need careful consideration regarding protein content. Previous research has found that protein in water-based lubricants promotes the wear rate of ultra-high molecular weight polyethylene (UHMWPE) in artificial joint testing.5 However, soluble proteins can denature as a result of frictional heating and produce a “solid.” This solid is generally com49
50 posed of albumin and globulin, similar to cooked eggwhite in texture, and was found to be an effective solid lubricant for UHMWPE.6 In conducting simulation tests on polyethylene, the highest wear rate occurs with the lubricant containing an intermediate concentration of proteins, above and below which, wear will decrease.7 That is, with inadequate lubricant protein levels, UHMWPE wear is minimal. Conversely, with elevated protein levels, denaturing occurs, thereby artificially protecting articular surfaces again reducing UHMWPE wear. Therefore protein must be at or slightly below physiological levels to accurately simulate in vitro polyethelene wear.8 A significant consideration is the turnover of lubricants used in testing. Local frictional heating between a head and cup will denature proteins. This logically occurs in the body as well as the laboratory. The major difference is the ability of the body to remove this debris and supply fresh proteins/ lubricant. In vitro simulators are “dead men walking.” Turnover must be accounted for, either by frequent lubricant changes or a sufficient volume to minimize degradation, yet supply fresh proteins.9 The albumin:globulin ratio in lubrication is another important consideration. An increased albumin:globulin ratio has been shown to significantly reduce wear in both UHMWPE and polytetrafluoroethylene (PTFE).8 In a study by the authors, 32 mm cups and cobalt-chrome heads were tested with different albumin:globulin ratios for 1 million cycles. It was found that the wear rate for UHMWPE decreased almost linearly as the albumin:globulin ratio increased.10 This is because of the preferential denaturing of albumin. The ideal ratio should be as close to joint fluid as possible, something generally not found in the ordinary bovine serum normally used. Fetal bovine serum has a consistency much closer to joint fluid.8 It must also be noted that synovial fluid is usually the intended lubricant for simulation purposes, but post total joint replacement (TJR) joint fluid is actually a more accurate target. Clinically relevant loading must be applied in conjunction with multi-directional motion for accurate wear studies. Achieving loading that mimics a variety of joint reaction forces for in vivo activities will provide a more accurate reflection of clinical outcomes. Motion pattern is another parameter important to the accuracy of a wear study. The direction and pattern of motion will determine an in vitro simulation’s accuracy when compared with in vivo outcomes. Wear rates and rankings of wear resistance for various UHMWPE materials are very different under unidirectional and multi-axial motion conditions. UHMWPE strain-hardens under unidirectional motion but strain-softens under multi-directional motion because of molecular orientation-induced anisotropy at the wear surface. Unidirectional wear testing machines are not sufficient for the evaluation of polyethelene bearing materials because they do not simulate the wear created by real joints.6,11 Multidirectional wear testing is essential to a successful artificial joint simulation. Simply applying cross-path motion and clinically relevant forces is not enough. The motions and forces must be applied
A. Wang, A. Essner and J. Cooper in a clinically relevant manner in terms of relative point of application. Acetabular inserts should be mounted superiorly to femoral heads, as occurs anatomically. In other words, cross-shear motion and force should be applied through the femoral head as in a hip joint (anatomy must be replicated in simulation). If force and motion are not applied anatomically, clinical results will be inaccurate. This is partly because of the concentrated location of force application on a stationary insert that results when a femoral head is articulated within it. If the cup is moved with the head remaining stationary, stresses and motions are distributed around the cup, altering the anisotropic orientation mechanism mentioned earlier. Components must be anatomically placed during testing to replicate the anatomy of the hip joint. By placing the component correctly, applying superior force to the cup, and simulating anatomic motion of the femoral head, test results will more effectively predict clinical outcomes.
Standardization and Current Practices The American Society of Testing and Materials (ASTM) has attempted to establish testing standards. It is generally difficult to achieve consensus on these standards, however, due mostly to varied individual laboratory experiences and equipment. A particular set of conditions may not be possible to enact on all equipment in the research community, and even if this were possible, it may not produce the same results. For this reason, as well as political concerns regarding historical data, a unified standard has been very difficult to establish. It seems that clinically relevant hip simulation models, and therefore standards, must account for the issues as opposed to rote following of a prescribed test standard. Individual adjustment of lubricants and other test factors may be necessary to achieve clinical relevance and accuracy. A standard has been provided by the International Standards Organization (ISO) but this does not precisely account for lubricant details. Further improvement appears necessary. The authors have made significant contributions to the development of ASTM and ISO standards regarding lubrication, component positioning, and other testing issues. Wang and co-workers established an in vitro lubricant formulation to closely observe clinical wear rates.8 They have also determined requirements for in vitro hip-simulator testing and based them on clinical results: wear rates, wear rate rankings, and wear debris size and morphology must agree with reported clinical results.12 The satisfaction of these standards, and other requirements such as 510(k) clearance from the FDA, do not guarantee clinical success. Determining the safety of a medical device and predicting its long-term wear are two very different objectives. Surgeons must possess a background knowledge of wear testing and its importance to effectively investigate the validation of the testing used by the manufacturers whose products they use. Laboratory hip joint simulator testing is required to accurately predict wear rates and magnitudes of new bearing sur-
Hip simulator testing
51 140
PTFE/PE Wear Rate Ratio
120
PTFE / UHMWPE
100 80 60 40 20 0 0
10
20
30
40
50
60
70
Total Protein Concentration (g/L)
Figure 1 PTFE/UHMWPE wear rate ratio as a function of total protein content. Dashed lines represent lower and upper ranges for Charnley data.8
faces. The goal of simulation testing is to simulate actual biomechanics, so it must be clinically relevant to be predictive. The first step in validating a simulation testing process is to qualitatively and quantitatively reproduce clinical failure. Once clinical failure is effectively reproduced, the model may be used prospectively to predict clinical success. But the evaluation process must be continual, with ongoing checks against clinical outcomes. Methodology may need adjustment as new information is gained. This must remain a dynamic process. There are parameters of simulation testing that must be considered above and beyond any standards or individual laboratory history. Degree of motion, position of components in the simulator, and an appropriately formulated lubrication without excessively high or low protein content must be used. The use of a multidirectional pattern of motion is the only way to simulate physiological biomechanics in the hip joint. Applying force to a superiorly mounted stationary acetabular cup in conjunction with this multidirectional motion will create the cross-shear forces needed for accurate simulation. The in vitro hip joint simulator testing that the authors conduct represents an aggressive protocol. This simulation methodology is based on the principles of anatomy, physiological motion, and physiological loading. By producing as much anatomic wear as possible through simulation, test results will provide the most accurate prediction possible for clinical outcomes. In vitro simulation testing must reflect wear magnitude and rate in addition to wear mechanisms.12 Simulation testing that is more aggressive will produce a more accurate prediction of clinical results. Essner and coworkers reported a hip simulator wear comparison between a
2-axis and 3-axis machine using the same acetabular cup design and material. While similar wear mechanisms were found, the 2-axis machine produced wear rates 4 to 5 times greater than the 3-axis machine. The higher wear rate showed a reasonable reflection of the clinical wear rate of the materials used.13
Prediction of Clinical Outcomes The in vitro hip simulator testing that the authors conduct has been able to successfully predict the clinical outcomes of polyethylene bearing surfaces used in total hip arthroplasty (THA). The clinical failure of Charnley’s Polytetrafluoroethlene (PTFE) was predicted by comparing test results of conventional UHMWPE against PTFE using an MTS hip simulator with lubricants bearing different protein levels. The lubricants containing the lowest amounts of protein produced the Charnley ratios.8 Results are shown in Figure 1. After this testing was used to optimize lubricant, the model was applied to more recent material models. The clinical failure of Hylamer® was predicted through a validation test comparing Hylamer® to conventional UHMWPE, where gamma-in-air sterilized Hylamer® was found to have wear rates 132% higher than gamma-in-air sterilized conventional UHMWPE.14 This data were compared with published results14 on the same materials, and not only were similar rankings reproduced, but absolute magnitudes were also closely replicated as shown in Figure 2. It should be noted that this was accomplished without any additional or artificial treatment of the samples. A valid model must be able to reproduce in vivo ranking, wear rates, and magnitudes of known materials under the conditions seen in clinical use. If
52
A. Wang, A. Essner and J. Cooper
Figure 2 In vitro and clinical linear wear rates of gamma-air sterilized Hylamer® and conventional UHMWPE. Note that in vitro volumetric data were converted to linear values assuming 106 cycles approximates one year of clinical use.14,15
Figure 3 In vitro and clinical performance of EtO gas and gamma-irradiation sterilized UHMWPE. Note that wear rates were normalized to gamma-irradiation sterilized values to allow comparison of volumetric and linear data.16-18
Hip simulator testing
53
Figure 4 In vitro and clinical linear wear rates of Duration® Stabilized UHMWPE and gamma-air sterilized UHMWPE. Note that in vitro volumetric data were converted to linear values assuming 106 cycles approximates one year of clinical use.20
artificial treatments such as accelerated chemical aging must be used, considerable suspicion should be raised about a particular model. In 1996, data indicating the relative difference between gas sterilized, nonirradiated UHMWPE compared with gammasterilized UHMWPE was presented.16 The nonirradiated material was found to have more than double the wear of the irradiated UHMWPE (Fig. 3). There have now been multiple clinical publications that have reported this same observation in vivo.18 This model has also been used prospectively since the ability to predict clinical success is the goal of simulation testing. Duration® stabilized polyethylene was evaluated extensively in the laboratory before clinical introduction. This work found approximately 30% lower wear for this material compared with gamma-air irradiated UHMWPE, the control material of the time. A multi-center study in started in 1996 demonstrated that the predicted wear rate of Gamma-in-Air sterilized UHMWPE compared with Duration UHMWPE was similar to the results produced by the hip simulator before the study.19 With Duration® polyethylene now nearing a history of 10 years, there are multiple studies that have confirmed the predicted reduction in wear. Figure 4 shows our compiled laboratory data (⬎100 samples) indicating a 27% reduction in wear for Duration polyethylene over Gammain-Air material. Strikingly, 5-year clinical data shows a reduction of 28%, with the same volumetric values.20 More recently, Crossfire polyethylene was evaluated in our laboratory. We predicted an approximate 90% reduction in wear when compared with conventional polyethylene
(N2vac) as shown in Figure 5.21 One in vivo study has shown a reduction in wear rate of 72% for this material,22 while another has found approximately 94% at 54 months using gamma-air sterilized UHMWPE as a control.23 Why the difference in observed reduction? The study showing 94% utilizes a RSA measurement technique, which is generally acknowledged as being more accurate than the radiographic technique used in the other study. It should be noted that 2 year data published using a radiographic technique showed a 42% difference,24 while the 5 year study shows 72%. As time and use go on, the absolute amount of wear generally increases, minimizing error and overcoming measurement resolution limitations. Minimal effect is anticipated from slight differences in controls since all were gamma-irradiation sterilized at a similar dose. The important observation to note is the similarity between our in vitro data and the RSA study. Focusing on clinical relevance and anatomical/biological issues has led our simulation testing to its predictive ability. The testing that we conduct has enabled us to predict results that are proven correct by clinical outcomes. Predicting clinical failure and then success should be the standard by which to measure the validation of in vitro hip simulation testing.
A Look Ahead Laboratory simulation testing reproduces the mechanisms, rate, and total magnitude of wear. This is specific to testing conditions, however. Other in vivo issues exist that may not be accounted for in ideal simulation testing. Clinical realities such as rim loading or impingement and third body wear
A. Wang, A. Essner and J. Cooper
54
Figure 5 In vitro wear data showing a 91% reduction in wear for Crossfire UHMWPE compared with nitrogen-vacuum packed gamma-irradiation sterilized UHMWPE.21 A 94% in vivo reduction was found when compared with gamma-air sterilized UHMWPE.23 Note that in vitro volumetric data were converted to linear values assuming 106 cycles approximates one year of clinical use. Also note the difference in control materials.
need to be accounted for in simulation models. Any new bearing technology should be evaluated under these nonideal circumstances to fully characterize its performance. These areas are just as detailed and complex as the ideal test conditions that were described above and therefore require the same level of research and optimization. How much of what type of debris should be used for third-body abrasive testing? How much force should be applied when simulating neck to rim impingement? What separation distance should be used when modeling subluxation or so called “micro-separation” during the swing phase of normal gait? These are all areas of current pioneering development. With these newer models, the existing gaps between predicted and observed wear will continue to decrease. We will begin to see the accuracy of prediction improve, enabling our ability to create bearing surfaces that show greater and greater reductions in wear.
Conclusion In vitro hip simulation testing is capable of predicting clinical outcomes. However, not all simulator testing is equal; there are a number of factors that influence the outcome of testing, including protein quantity and type in lubrication, direction of motion and force, and appropriate implant positioning. Determine if a given simulation test is validated before accepting products or technologies developed with it for use. The model must first reproduce clinical failure, including the mechanism, rate and magnitude of wear. This must be ac-
complished without artificial treatments, focusing on the same conditions that products or technologies failed clinically. Only once a model has been able to reproduce known failures can it be applied to product or technology development. Ideally, the model should also have a demonstrated history of predictive success. These qualifications are the best criteria by which to judge the validation of simulation testing methods.
References 1. Min BW, Song KS, Kang CH, et al: Polyethylene liner failure in secondgeneration Harris-Galante acetabular components. J Arthroplasty 6:717-722, 2005 2. Chmell MJ, Poss R, Thomas WH, et al: Early failure of Hylamer acetabular inserts due to eccentric wear. J Arthroplasty 11(3):351-353, 1996 3. McKellop H, Clarke I, Markolf K, et al: Friction and wear properties of polymer, metal, and ceramic prosthetic joint materials evaluated on a multichannel screening device. Biomed Mater Res 15:619-653, 1981 4. Namba RS, Paxton L, Fithian DC, et al: Obesity and perioperative morbidity in total hip and total knee arthroplasty patients. J Arthroplasty 20(suppl 3):46-50, 2005 5. Wang A, Polineni VK, Essner A, et al: Role of Proteins and Hylauronic Acid in the Lubrication and Wear of UHMWPE Acetabular Cups. Transactions of the Society for Biomaterials 24th Annual Meeting, 1998, p 218 6. Wang A, Essner A, Polineni VK, et al: Lubrication and wear of ultrahigh molecular weight polyethelene in total joint replacements. Tribol Int 31:1-3, 1998 7. Polineni WK, Wang A, Essner A, et al: Effect of Lubricant Protein Concentration on the Wear of UHMWPE Acetabular Cups Against Cobalt-chrome and Alumina Femoral Heads. Transactions of the Society for Biomaterials 23rd Annual Meeting, 1997, p 154
Hip simulator testing 8. Wang A, Essner A, Schmidig G: The effects of lubricant composition on in vitro wear testing of polymeric acetabular components. J Biomaterial Res Part B: Appl Biomaterials 68B: 45-52, 2004 9. Wang A, Polineni VK, Essner A, et al: Quantitative analysis of serum degradation and its effect on the outcome of hip joint simulator wear testing of UHMWPE. Transactions of the 45th Annual Meeting of the Orthopaedic Research Society, 1999, p 73 10. Wang A, Schmidig G, Essner A: The effects of lubricant protein composition on wear rates of UHMWPE and PTFE acetabular cups. Proceedings of the 6th World Biomaterials Congress, 2000, p 1075 11. Wang A, Sun DC, Stark C, et al: Wear Testing Based on Unidirectional Motion: Fact or Artifact? Fifth World Biomaterials Congress, May 29 – June 2, 1996, Toronto, Canada 12. Essner A, Schmidig G, Wang A: The clinical relevance of hip joint simulator testing: In-vitro and in-vivo comparisons. Wear 259:882886, 2005 13. Essner A, Wang A: Comparison of two different hip wear simulators: Transactions of the 46th Annual Meeting of the Orthopaedic Research Society, 2000, p 217 14. Essner A, Polineni VK, Wang A, et al: Hip simulator wear of “enhanced” UHMWPE acetabular inserts. Transactions of the 44th Annual Meeting of the Orthopaedic Research Society, 1998, p 774 15. Chandler HP, Smith S: Comparison of in-vivo rates of wear of Hylamer™ and conventional polyethylene acetabular liners. The 26th Annual Harvard Hip Course, Total Hip Replacement–Polyethylene: Where are We Now? Cambridge, MA, 1996 16. Wang A, Sun DC, Stark C, et al: Effect of sterilization methods on the wear of UHMWPE acetabular cups. Proceedings of the 5th World Biomaterials Congress, 1996, p 198
55 17. Wang A, Stark C, Dumbleton JH: Mechanistic and morphological origins of ultra-high molecular weight polyethylene wear debris in total joint replacement prostheses. Proc Instn Mech Engrs: Part H: J Eng Med, 210:141-155, 1996 18. Digas G, Thanner J, Nivbrant B, et al: Increase in early polyethylene wear after sterilization with ethylene oxide. Acta Orthop Scand 74:531541, 2003 19. Essner A, Wang A, Martell J, et al: In-vitro and in-vivo acetabular cup wear corroboration. Transactions of the 51st Annual Meeting of the Orthopaedic Research Society, New Orleans, 2003, p 282 20. Grimm B, Geerdink C, Tonino AJ, et al: Wear performance of a crosslinked polyethylene for total hip arthroplasty measured in-vitro and during a prospective randomized clinical study at five year followup. 7th EFFORT Congress, Lisbon, 2005, p 3676 21. Essner A, Yau S-S, Schmidig G, et al: Reducing hip wear without compromising mechanical strength: A next generation cross-linked and annealed polyethylene. Transactions of the 5th Combined Meeting of the Orthopaedic Research Societies of Canada, USA, Japan, Europe, Banff, Canada, 2004, p 80 22. D’Antonio JA, Manley MT, Capello WN, et al: Five-year experience with Crossfire® highly cross-linked polyethylene. Clin Orthop Relat Res 441:143-150, 2005 23. Röhrl SM, Nivbrant B: Extremely low in vivo wear for non-remelted crosslinked PE after 4.5 years: A RSA study. Transactions of the 51st Annual Meeting of the Orthopaedic Research Society, New Orleans, 2003, p 282 24. Martell J, Verner J, Incavo S: Clinical performance of a highly crosslinked polyethylene at two years in total hip arthroplasty: A randomized prospective trial. J Arthroplasty 55-59, 2003
T
oday’s total hip arthroplasty patients comprise a changing population that grows younger and more active every year. As patients’ expectations increase, so does the demand for new bearing technologies capable of long-term wear under harsh conditions. Testing is a crucial phase in the development of these new technologies, and there are many parameters affecting the validity and predictive potential of implant testing. Any new bearing technologies must be subjected to rigorous and exacting in vitro testing, to meet the implant demands of present and future patient populations. The clinical relevance of this testing is critical. There is a long clinical history of bearing material failure in acetabular components. A study by Min and coworkers on the Harris-Galante cup (Harris-Galante II, Zimmer, Warsaw, IN) revealed a 17.3% failure rate of the polyethelene liner after 8 years.1 Previous clinical failures include the early failure of Hylamer polyethelene (DePuy, Warsaw, IN) in total hip2 arthroplasty. Teflon and polyester are two other bearing surfaces used in total hip arthroplasty, both having poor clinical histories.3 New bearing technologies face not only the failures of past generations, but also the increasing patient demand for better implant performance. Disease also challenges implant perforBiomechanics and Tribology Laboratory, Stryker Orthopaedics, Mahwah, NJ. Address reprint requests to Aiguo Wang, PhD, Science and Clinical Studies, Stryker Orthopaedics, 325 Corporate Drive, Mahwah, NJ 07430; [email protected].
1045-4527/06/$-see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1053/j.sart.2006.06.005
mance, such as obesity, as well as diseases linked to obesity such as diabetes and hypertension.4 Other challenges include suboptimal implant placement (sometimes because of new trends or surgical technique) such as minimally invasive surgery (MIS), and establishing realistic patient expectations. Laboratory in vitro simulation testing must reproduce the mechanisms, rate, and total magnitude of wear. The ultimate goal of these reproductions is the accurate prediction of the clinical failure or success of bearing surfaces and design. As bearing surface technologies advance to meet increasing patient expectations, a validated simulation testing model with a clinical history of predictive success must be used to effectively analyze and develop these technologies. The objective of this paper is to review the criteria for effective simulation testing and provide the reader with the background necessary for the investigation of the testing methods of new implant technologies.
Wear Study Parameters When designing a wear study, several investigational parameters must be considered. Lubrication is critical to the simulation of physical joints, and its composition is a determinant of the testing result. Lubricants for simulator testing need careful consideration regarding protein content. Previous research has found that protein in water-based lubricants promotes the wear rate of ultra-high molecular weight polyethylene (UHMWPE) in artificial joint testing.5 However, soluble proteins can denature as a result of frictional heating and produce a “solid.” This solid is generally com49
50 posed of albumin and globulin, similar to cooked eggwhite in texture, and was found to be an effective solid lubricant for UHMWPE.6 In conducting simulation tests on polyethylene, the highest wear rate occurs with the lubricant containing an intermediate concentration of proteins, above and below which, wear will decrease.7 That is, with inadequate lubricant protein levels, UHMWPE wear is minimal. Conversely, with elevated protein levels, denaturing occurs, thereby artificially protecting articular surfaces again reducing UHMWPE wear. Therefore protein must be at or slightly below physiological levels to accurately simulate in vitro polyethelene wear.8 A significant consideration is the turnover of lubricants used in testing. Local frictional heating between a head and cup will denature proteins. This logically occurs in the body as well as the laboratory. The major difference is the ability of the body to remove this debris and supply fresh proteins/ lubricant. In vitro simulators are “dead men walking.” Turnover must be accounted for, either by frequent lubricant changes or a sufficient volume to minimize degradation, yet supply fresh proteins.9 The albumin:globulin ratio in lubrication is another important consideration. An increased albumin:globulin ratio has been shown to significantly reduce wear in both UHMWPE and polytetrafluoroethylene (PTFE).8 In a study by the authors, 32 mm cups and cobalt-chrome heads were tested with different albumin:globulin ratios for 1 million cycles. It was found that the wear rate for UHMWPE decreased almost linearly as the albumin:globulin ratio increased.10 This is because of the preferential denaturing of albumin. The ideal ratio should be as close to joint fluid as possible, something generally not found in the ordinary bovine serum normally used. Fetal bovine serum has a consistency much closer to joint fluid.8 It must also be noted that synovial fluid is usually the intended lubricant for simulation purposes, but post total joint replacement (TJR) joint fluid is actually a more accurate target. Clinically relevant loading must be applied in conjunction with multi-directional motion for accurate wear studies. Achieving loading that mimics a variety of joint reaction forces for in vivo activities will provide a more accurate reflection of clinical outcomes. Motion pattern is another parameter important to the accuracy of a wear study. The direction and pattern of motion will determine an in vitro simulation’s accuracy when compared with in vivo outcomes. Wear rates and rankings of wear resistance for various UHMWPE materials are very different under unidirectional and multi-axial motion conditions. UHMWPE strain-hardens under unidirectional motion but strain-softens under multi-directional motion because of molecular orientation-induced anisotropy at the wear surface. Unidirectional wear testing machines are not sufficient for the evaluation of polyethelene bearing materials because they do not simulate the wear created by real joints.6,11 Multidirectional wear testing is essential to a successful artificial joint simulation. Simply applying cross-path motion and clinically relevant forces is not enough. The motions and forces must be applied
A. Wang, A. Essner and J. Cooper in a clinically relevant manner in terms of relative point of application. Acetabular inserts should be mounted superiorly to femoral heads, as occurs anatomically. In other words, cross-shear motion and force should be applied through the femoral head as in a hip joint (anatomy must be replicated in simulation). If force and motion are not applied anatomically, clinical results will be inaccurate. This is partly because of the concentrated location of force application on a stationary insert that results when a femoral head is articulated within it. If the cup is moved with the head remaining stationary, stresses and motions are distributed around the cup, altering the anisotropic orientation mechanism mentioned earlier. Components must be anatomically placed during testing to replicate the anatomy of the hip joint. By placing the component correctly, applying superior force to the cup, and simulating anatomic motion of the femoral head, test results will more effectively predict clinical outcomes.
Standardization and Current Practices The American Society of Testing and Materials (ASTM) has attempted to establish testing standards. It is generally difficult to achieve consensus on these standards, however, due mostly to varied individual laboratory experiences and equipment. A particular set of conditions may not be possible to enact on all equipment in the research community, and even if this were possible, it may not produce the same results. For this reason, as well as political concerns regarding historical data, a unified standard has been very difficult to establish. It seems that clinically relevant hip simulation models, and therefore standards, must account for the issues as opposed to rote following of a prescribed test standard. Individual adjustment of lubricants and other test factors may be necessary to achieve clinical relevance and accuracy. A standard has been provided by the International Standards Organization (ISO) but this does not precisely account for lubricant details. Further improvement appears necessary. The authors have made significant contributions to the development of ASTM and ISO standards regarding lubrication, component positioning, and other testing issues. Wang and co-workers established an in vitro lubricant formulation to closely observe clinical wear rates.8 They have also determined requirements for in vitro hip-simulator testing and based them on clinical results: wear rates, wear rate rankings, and wear debris size and morphology must agree with reported clinical results.12 The satisfaction of these standards, and other requirements such as 510(k) clearance from the FDA, do not guarantee clinical success. Determining the safety of a medical device and predicting its long-term wear are two very different objectives. Surgeons must possess a background knowledge of wear testing and its importance to effectively investigate the validation of the testing used by the manufacturers whose products they use. Laboratory hip joint simulator testing is required to accurately predict wear rates and magnitudes of new bearing sur-
Hip simulator testing
51 140
PTFE/PE Wear Rate Ratio
120
PTFE / UHMWPE
100 80 60 40 20 0 0
10
20
30
40
50
60
70
Total Protein Concentration (g/L)
Figure 1 PTFE/UHMWPE wear rate ratio as a function of total protein content. Dashed lines represent lower and upper ranges for Charnley data.8
faces. The goal of simulation testing is to simulate actual biomechanics, so it must be clinically relevant to be predictive. The first step in validating a simulation testing process is to qualitatively and quantitatively reproduce clinical failure. Once clinical failure is effectively reproduced, the model may be used prospectively to predict clinical success. But the evaluation process must be continual, with ongoing checks against clinical outcomes. Methodology may need adjustment as new information is gained. This must remain a dynamic process. There are parameters of simulation testing that must be considered above and beyond any standards or individual laboratory history. Degree of motion, position of components in the simulator, and an appropriately formulated lubrication without excessively high or low protein content must be used. The use of a multidirectional pattern of motion is the only way to simulate physiological biomechanics in the hip joint. Applying force to a superiorly mounted stationary acetabular cup in conjunction with this multidirectional motion will create the cross-shear forces needed for accurate simulation. The in vitro hip joint simulator testing that the authors conduct represents an aggressive protocol. This simulation methodology is based on the principles of anatomy, physiological motion, and physiological loading. By producing as much anatomic wear as possible through simulation, test results will provide the most accurate prediction possible for clinical outcomes. In vitro simulation testing must reflect wear magnitude and rate in addition to wear mechanisms.12 Simulation testing that is more aggressive will produce a more accurate prediction of clinical results. Essner and coworkers reported a hip simulator wear comparison between a
2-axis and 3-axis machine using the same acetabular cup design and material. While similar wear mechanisms were found, the 2-axis machine produced wear rates 4 to 5 times greater than the 3-axis machine. The higher wear rate showed a reasonable reflection of the clinical wear rate of the materials used.13
Prediction of Clinical Outcomes The in vitro hip simulator testing that the authors conduct has been able to successfully predict the clinical outcomes of polyethylene bearing surfaces used in total hip arthroplasty (THA). The clinical failure of Charnley’s Polytetrafluoroethlene (PTFE) was predicted by comparing test results of conventional UHMWPE against PTFE using an MTS hip simulator with lubricants bearing different protein levels. The lubricants containing the lowest amounts of protein produced the Charnley ratios.8 Results are shown in Figure 1. After this testing was used to optimize lubricant, the model was applied to more recent material models. The clinical failure of Hylamer® was predicted through a validation test comparing Hylamer® to conventional UHMWPE, where gamma-in-air sterilized Hylamer® was found to have wear rates 132% higher than gamma-in-air sterilized conventional UHMWPE.14 This data were compared with published results14 on the same materials, and not only were similar rankings reproduced, but absolute magnitudes were also closely replicated as shown in Figure 2. It should be noted that this was accomplished without any additional or artificial treatment of the samples. A valid model must be able to reproduce in vivo ranking, wear rates, and magnitudes of known materials under the conditions seen in clinical use. If
52
A. Wang, A. Essner and J. Cooper
Figure 2 In vitro and clinical linear wear rates of gamma-air sterilized Hylamer® and conventional UHMWPE. Note that in vitro volumetric data were converted to linear values assuming 106 cycles approximates one year of clinical use.14,15
Figure 3 In vitro and clinical performance of EtO gas and gamma-irradiation sterilized UHMWPE. Note that wear rates were normalized to gamma-irradiation sterilized values to allow comparison of volumetric and linear data.16-18
Hip simulator testing
53
Figure 4 In vitro and clinical linear wear rates of Duration® Stabilized UHMWPE and gamma-air sterilized UHMWPE. Note that in vitro volumetric data were converted to linear values assuming 106 cycles approximates one year of clinical use.20
artificial treatments such as accelerated chemical aging must be used, considerable suspicion should be raised about a particular model. In 1996, data indicating the relative difference between gas sterilized, nonirradiated UHMWPE compared with gammasterilized UHMWPE was presented.16 The nonirradiated material was found to have more than double the wear of the irradiated UHMWPE (Fig. 3). There have now been multiple clinical publications that have reported this same observation in vivo.18 This model has also been used prospectively since the ability to predict clinical success is the goal of simulation testing. Duration® stabilized polyethylene was evaluated extensively in the laboratory before clinical introduction. This work found approximately 30% lower wear for this material compared with gamma-air irradiated UHMWPE, the control material of the time. A multi-center study in started in 1996 demonstrated that the predicted wear rate of Gamma-in-Air sterilized UHMWPE compared with Duration UHMWPE was similar to the results produced by the hip simulator before the study.19 With Duration® polyethylene now nearing a history of 10 years, there are multiple studies that have confirmed the predicted reduction in wear. Figure 4 shows our compiled laboratory data (⬎100 samples) indicating a 27% reduction in wear for Duration polyethylene over Gammain-Air material. Strikingly, 5-year clinical data shows a reduction of 28%, with the same volumetric values.20 More recently, Crossfire polyethylene was evaluated in our laboratory. We predicted an approximate 90% reduction in wear when compared with conventional polyethylene
(N2vac) as shown in Figure 5.21 One in vivo study has shown a reduction in wear rate of 72% for this material,22 while another has found approximately 94% at 54 months using gamma-air sterilized UHMWPE as a control.23 Why the difference in observed reduction? The study showing 94% utilizes a RSA measurement technique, which is generally acknowledged as being more accurate than the radiographic technique used in the other study. It should be noted that 2 year data published using a radiographic technique showed a 42% difference,24 while the 5 year study shows 72%. As time and use go on, the absolute amount of wear generally increases, minimizing error and overcoming measurement resolution limitations. Minimal effect is anticipated from slight differences in controls since all were gamma-irradiation sterilized at a similar dose. The important observation to note is the similarity between our in vitro data and the RSA study. Focusing on clinical relevance and anatomical/biological issues has led our simulation testing to its predictive ability. The testing that we conduct has enabled us to predict results that are proven correct by clinical outcomes. Predicting clinical failure and then success should be the standard by which to measure the validation of in vitro hip simulation testing.
A Look Ahead Laboratory simulation testing reproduces the mechanisms, rate, and total magnitude of wear. This is specific to testing conditions, however. Other in vivo issues exist that may not be accounted for in ideal simulation testing. Clinical realities such as rim loading or impingement and third body wear
A. Wang, A. Essner and J. Cooper
54
Figure 5 In vitro wear data showing a 91% reduction in wear for Crossfire UHMWPE compared with nitrogen-vacuum packed gamma-irradiation sterilized UHMWPE.21 A 94% in vivo reduction was found when compared with gamma-air sterilized UHMWPE.23 Note that in vitro volumetric data were converted to linear values assuming 106 cycles approximates one year of clinical use. Also note the difference in control materials.
need to be accounted for in simulation models. Any new bearing technology should be evaluated under these nonideal circumstances to fully characterize its performance. These areas are just as detailed and complex as the ideal test conditions that were described above and therefore require the same level of research and optimization. How much of what type of debris should be used for third-body abrasive testing? How much force should be applied when simulating neck to rim impingement? What separation distance should be used when modeling subluxation or so called “micro-separation” during the swing phase of normal gait? These are all areas of current pioneering development. With these newer models, the existing gaps between predicted and observed wear will continue to decrease. We will begin to see the accuracy of prediction improve, enabling our ability to create bearing surfaces that show greater and greater reductions in wear.
Conclusion In vitro hip simulation testing is capable of predicting clinical outcomes. However, not all simulator testing is equal; there are a number of factors that influence the outcome of testing, including protein quantity and type in lubrication, direction of motion and force, and appropriate implant positioning. Determine if a given simulation test is validated before accepting products or technologies developed with it for use. The model must first reproduce clinical failure, including the mechanism, rate and magnitude of wear. This must be ac-
complished without artificial treatments, focusing on the same conditions that products or technologies failed clinically. Only once a model has been able to reproduce known failures can it be applied to product or technology development. Ideally, the model should also have a demonstrated history of predictive success. These qualifications are the best criteria by which to judge the validation of simulation testing methods.
References 1. Min BW, Song KS, Kang CH, et al: Polyethylene liner failure in secondgeneration Harris-Galante acetabular components. J Arthroplasty 6:717-722, 2005 2. Chmell MJ, Poss R, Thomas WH, et al: Early failure of Hylamer acetabular inserts due to eccentric wear. J Arthroplasty 11(3):351-353, 1996 3. McKellop H, Clarke I, Markolf K, et al: Friction and wear properties of polymer, metal, and ceramic prosthetic joint materials evaluated on a multichannel screening device. Biomed Mater Res 15:619-653, 1981 4. Namba RS, Paxton L, Fithian DC, et al: Obesity and perioperative morbidity in total hip and total knee arthroplasty patients. J Arthroplasty 20(suppl 3):46-50, 2005 5. Wang A, Polineni VK, Essner A, et al: Role of Proteins and Hylauronic Acid in the Lubrication and Wear of UHMWPE Acetabular Cups. Transactions of the Society for Biomaterials 24th Annual Meeting, 1998, p 218 6. Wang A, Essner A, Polineni VK, et al: Lubrication and wear of ultrahigh molecular weight polyethelene in total joint replacements. Tribol Int 31:1-3, 1998 7. Polineni WK, Wang A, Essner A, et al: Effect of Lubricant Protein Concentration on the Wear of UHMWPE Acetabular Cups Against Cobalt-chrome and Alumina Femoral Heads. Transactions of the Society for Biomaterials 23rd Annual Meeting, 1997, p 154
Hip simulator testing 8. Wang A, Essner A, Schmidig G: The effects of lubricant composition on in vitro wear testing of polymeric acetabular components. J Biomaterial Res Part B: Appl Biomaterials 68B: 45-52, 2004 9. Wang A, Polineni VK, Essner A, et al: Quantitative analysis of serum degradation and its effect on the outcome of hip joint simulator wear testing of UHMWPE. Transactions of the 45th Annual Meeting of the Orthopaedic Research Society, 1999, p 73 10. Wang A, Schmidig G, Essner A: The effects of lubricant protein composition on wear rates of UHMWPE and PTFE acetabular cups. Proceedings of the 6th World Biomaterials Congress, 2000, p 1075 11. Wang A, Sun DC, Stark C, et al: Wear Testing Based on Unidirectional Motion: Fact or Artifact? Fifth World Biomaterials Congress, May 29 – June 2, 1996, Toronto, Canada 12. Essner A, Schmidig G, Wang A: The clinical relevance of hip joint simulator testing: In-vitro and in-vivo comparisons. Wear 259:882886, 2005 13. Essner A, Wang A: Comparison of two different hip wear simulators: Transactions of the 46th Annual Meeting of the Orthopaedic Research Society, 2000, p 217 14. Essner A, Polineni VK, Wang A, et al: Hip simulator wear of “enhanced” UHMWPE acetabular inserts. Transactions of the 44th Annual Meeting of the Orthopaedic Research Society, 1998, p 774 15. Chandler HP, Smith S: Comparison of in-vivo rates of wear of Hylamer™ and conventional polyethylene acetabular liners. The 26th Annual Harvard Hip Course, Total Hip Replacement–Polyethylene: Where are We Now? Cambridge, MA, 1996 16. Wang A, Sun DC, Stark C, et al: Effect of sterilization methods on the wear of UHMWPE acetabular cups. Proceedings of the 5th World Biomaterials Congress, 1996, p 198
55 17. Wang A, Stark C, Dumbleton JH: Mechanistic and morphological origins of ultra-high molecular weight polyethylene wear debris in total joint replacement prostheses. Proc Instn Mech Engrs: Part H: J Eng Med, 210:141-155, 1996 18. Digas G, Thanner J, Nivbrant B, et al: Increase in early polyethylene wear after sterilization with ethylene oxide. Acta Orthop Scand 74:531541, 2003 19. Essner A, Wang A, Martell J, et al: In-vitro and in-vivo acetabular cup wear corroboration. Transactions of the 51st Annual Meeting of the Orthopaedic Research Society, New Orleans, 2003, p 282 20. Grimm B, Geerdink C, Tonino AJ, et al: Wear performance of a crosslinked polyethylene for total hip arthroplasty measured in-vitro and during a prospective randomized clinical study at five year followup. 7th EFFORT Congress, Lisbon, 2005, p 3676 21. Essner A, Yau S-S, Schmidig G, et al: Reducing hip wear without compromising mechanical strength: A next generation cross-linked and annealed polyethylene. Transactions of the 5th Combined Meeting of the Orthopaedic Research Societies of Canada, USA, Japan, Europe, Banff, Canada, 2004, p 80 22. D’Antonio JA, Manley MT, Capello WN, et al: Five-year experience with Crossfire® highly cross-linked polyethylene. Clin Orthop Relat Res 441:143-150, 2005 23. Röhrl SM, Nivbrant B: Extremely low in vivo wear for non-remelted crosslinked PE after 4.5 years: A RSA study. Transactions of the 51st Annual Meeting of the Orthopaedic Research Society, New Orleans, 2003, p 282 24. Martell J, Verner J, Incavo S: Clinical performance of a highly crosslinked polyethylene at two years in total hip arthroplasty: A randomized prospective trial. J Arthroplasty 55-59, 2003