For the past 30 years, the search for highly wear-resistant materials for total hip arthroplasty has been innovative. To improve the life expectancy of tribologically stressed load bearing couples, an alumina-on-alumina combination was introduced based on this ceramic excellent wear resistance 10 and biocompatibility. 19 The first use of alumina ceramic for the articulating components of total hip prostheses was by Boutin in France. 1 Biomechanical research using simulator tests have proved low friction, 21 and low wear (only approximately 1/20 of metal-on-polyethylene combination). Initial clinical results were less satisfactory than expected. 11 Alumina quality and component designs were far from perfect, but wear was not thought to be attributable to the ceramic-on-ceramic total hip prosthesis. The occurrence of wear was regarded as a consequence of loosening or was attributed mainly to “clinically exceptional” situations and/or to the poor quality of the material. 8 Since 1977, considerable changes were made in the material and in the geometry of the ceramic components of the Ceraver-Osteal® (Ceraver, Roissy, France) total hip prosthesis. Clinical results have reached values comparable with results of other designs and even better results have been reported for young and active patients. 30 In addition, most of the revisions were considered relatively simple procedures because loss of bone stock loss was not observed. 25 From these results it seems that aseptic loosening of the socket remained the main cause of long-term failure. The mechanisms leading to loosening still have not been explained satisfactorily. In addition to in vitro studies, comprehensive retrieval studies have to be performed to better understand failures and to avoid them in the future.
The aim of the current study was to investigate the wear patterns of a series of explanted alumina components retrieved for aseptic loosening of the socket. Wear quantification on macroscopic and microscopic scales in combination with microstructure analysis and clinical and radiographic data were used to identify the main risk factors involved in the in vivo alumina wear process.
MATERIALS AND METHODS
Eleven explanted Ceraver-Osteal® bearings were included in the current study. They were implanted between 1977 and 1988 in patients who were treated at the authors’ institution. Ten matched pairs of retrieved alumina femoral heads and sockets and one additional socket were available for analysis. All implants had been manufactured by Ceraver-Osteal® (Roissy, France) after 1977. A smooth stem made of Ti alloy (TiAL6V4) always was cemented. The alumina femoral head was secured to the stem through conical “sleeving”. The sockets and femoral heads were made of pure dense surgical grade aluminum oxide (Al2O3). The nominal femoral head diameter was 32 mm, whereas the outside diameters of the socket ranged from 50 to 58 mm.
There were six females and five males. The mean age of the patients at the time of hip replacement was 55 ± 17 years (range, 17–74 years). The average weight of the patients was 69 kg (range, 52–86 kg). The initial diagnoses were idiopathic femoral head osteonecrosis in one patient, primary osteoarthritis in five patients, posttraumatic femoral head osteonecrosis in two patients, posttraumatic osteoarthritis in two patients, and congenital dislocation of the hip in one patient. The mean implantation time was 11.3 ± 5.2 years (range, 1.7–18.4 years). All prostheses were retrieved at the time of revision for aseptic loosening of the socket.
Clinical and Radiologic Parameters
Clinical and radiologic evaluations were completed by an independent observer. Evaluation was performed in four stages by reviewing the charts and radiographs obtained: (1) before implantation; (2) immediately after implantation, at the latest followup; (3) before revision; and (4) at an intermediate followup before revision. Anteroposterior (AP) radiographs of the pelvis were used to locate the positions of the components relative to the teardrop line. Socket component migration of more than 3 mm either vertically or horizontally as seen on serial radiographs was considered definite migration. Any change of socket inclination by more than 5° was considered definite tilting. 5
Small samples (10 mm × 5 mm) of both components were cut using a diamond saw. After polishing and thermal etching (1500° C, 10 minutes), scanning electron microscopy pictures were taken and analyzed with an image analyzer (QWin, Leica Imaging System Ltd, Cambridge, UK). A minimum of 500 grains was digitized and grain distribution descriptive parameters were computed to characterize alumina quality in terms of average grain size, grain size distribution, and porosity percentage.
After reception, and before examination, the femoral heads were removed from stems and all specimens were cleaned in a 10% formol bath for 48 hours. Macrographs of the cleaned implants were taken. They were examined visually using a magnifier for evidence of gross wear pattern. Explants then were classified according to wear features.
Macroscopic wear was assessed by measuring the change in dimensions of both components, using a coordinate measuring machine. 7 Measurements were done using a Kemco 400 CMM (Keeley Measurement Company, Derby, UK) equipped with a Renishaw PH9 MkII probe head with a TP25W touch trigger probe (Renishaw Metrology Limited, Gloucestershire, UK). Once a datum plane (XY) and the origin of the coordinate system (0.0.0) were set, three-dimensional coordinates of multiple points describing, two-dimensional profiles were computed by measuring the position of 30 points at 6° intervals in the vertical (YZ) plane across the worn region of each component. For each component, two measurements were obtained. Differences between measured and nominal radii were computed to evaluate penetration.
Microscopic wear features were assessed by Talysurf measurements. 23,24,32 Surface roughness measurements (average arithmetic roughness, Ra and total roughness, Rt) were taken of worn and unworn areas using a Talysurf 6 contacting profilometer (Rank Taylor Hobson, Leicester, UK). Roughness measurements were taken along 6 mm lines using five, 0.8-mm cutoff lengths or 4 mm lines using 12, 0.25-mm cutoff lengths. One measurement was taken per area of interest.
All results were analyzed by one way analysis of variance (ANOVA) using a multiple group ANOVA test. In case of significant differences, post hoc Student’s t test was applied in selected cases. When looking for correlations, Spearman’s rank correlation tests were computed. The limit of significance was fixed at 0.05.
Retrieved Prostheses Analysis
After gross examination, components were classified into three groups according to their wear patterns. Group 1 was defined as severe wear (n = 2): femoral heads and cups in this group showed a visible loss of material and important sphericity deviations. Group 2 was defined as stripe wear (n = 5): the upper part of the femoral heads showed a localized oblong worn area. The worn area of the counterface generally was not obvious. Group 3 was defined as low wear (n = 4): this category was established when none of the features above were visible, except occasional spots or scratches.
In addition, some components had obvious signs of rim wear. A particular category was not established for rim wear because this phenomenon was likely to appear once the cup had loosened and/or tilted but did not trigger loosening. Metallic scratches and brownish deposits also were observed occasionally.
Alumina Microstructure Characterization
Scanning electron microscopy observation of alumina microstructure allowed the grain size distributions of grain areas and grain diameters to be characterized. A constant improvement in the quality microstructure of the cups was observed. The mean grain areas for cups was 3.34 ± 1.69 μm2 versus 4.69 ± 2.41 μm2 for femoral heads (p = 0.036). The mean grain diameter for cups was 1.79 ± 0.45 μm versus 2.02 ± 0.53 μm for femoral heads.
Dramatic changes were observed in ceramics manufactured before 1980 and ceramics manufactured after 1980. According to the manufacturer, before 1977, the average grain size was 4.3 ± 2.5 μm. After 1979, the average grain size decreased to 3.4 ± 0.8 μm, the range of grain size distribution was reduced (1.5–5 μm). However alumina purity increased from 99.5% to 99.8% and density increased from 3.93 g/cm3 to 3.94 g/cm3. However, improvements were much more progressive from 1977 to 1988. When mean grain areas of the components are plotted versus their implantation date, a linear decrease (r = −0.65; p = 0.002) with time was observed (Fig 1). Furthermore, very few components had a microstructure with a maximum grain size remaining below the 5 μm limit.
According to the classification into the groups, explants were analyzed using coordinate measuring machine and profilometry techniques to quantify wear.
Coordinate Measuring Machine Analysis
Using the coordinate measuring machine coordinates of multiple points describing twodimensional profiles of femoral heads and cups, the maximum penetrations and linear penetration rates were computed.
There was no clear relationship between macroscopic wear features and the time of implantation. For the entire series the maximum penetration values ranged from 13 to 2600 μm for femoral heads and from 8 to 670 μm for cups. The corresponding linear wear rates ranged from 2 to 165 μm/year and from 1 to 48 μm/year for femoral heads and cups, respectively. The maximum frequency would be most appropriately described with the median values of 5 μm/year for femoral heads and cups. 29
When all these data are gathered according to the three categories established (Table 1), components with stripe and low wear had similar values (maximum penetration < 130 μm, and penetration rates < 10 μm/year), both significantly less important than those observed with severely worn components (p < 10−6).
Mean Talysurf values computed for each group are shown in Table 2. No significant differences were observed between data obtained respectively from worn femoral heads and cups even if the heads showed a tendency for having rougher surfaces than the cups. When all individual data were gathered according to the three categories, no differences were observed on unworn regions. However, when considering worn regions, low wear components still had very smooth surfaces because Ra values remained below 0.05 μm whereas Rt values remained below 0.55 μm. On the contrary, components with stripe and severe wear had significantly rougher surfaces with Rt values increasing to 4 μm (p < 0.05). The main difference between the surfaces of components with stripe and severe wear did not seem to be the magnitude of roughness but rather the extent of the worn region over the unworn region.
A significant linear correlation (r = 0.78, p < 10−4) was observed when maximum grain diameters were plotted versus Rt values assessed on worn regions (Fig 2). However, no clear relationship was obtained between macroscopic wear features (penetration) and any grain size distribution parameter.
Patients’ Clinical Data
The relationship between wear features and clinical data revealed that patients could be divided into two groups: one group consisting of young, heavy, and tall males and a second group consisting of older, lighter, and smaller women.
Duration of implantation
No simple relationship could be observed between any wear parameter and the duration of implantation. In addition, increased wear was not associated with a reduced service life but was seen in components having been in use more than 15 years.
Initial Cup Inclination
The initial cup inclination was plotted versus roughness values measured on worn regions. A second order relationship was obtained, suggesting an optimum value around 45°. On a macroscopic scale (penetration rate), only an initial cup inclination greater than 50° seemed to be associated with an increased wear magnitude.
A linear correlation was observed between Rt values measured on worn areas and patients’ weights (r = 0.78, p < 10−4). The corresponding plot for macroscopic wear features revealed an exponential relationship (Fig 3). High stress seemed to be able to enhance wear magnitude.
Combination of Unfavorable Factors
Among available data, risk factors were considered to be all parameters leading to an increase in the load the joint was submitted to. These risk factors were patients young age (< 50 years), male gender, and high weight (> 80 kg). In addition, maximum grain diameter larger than 6 μm on either component was considered unfavorable. Furthermore, parameters that might have impaired the load distribution over the femoral head surface such as an initial cup inclination either too steep (≥ 50°) or not enough (≤ 40°), also were considered. Finally, the occurrence of either cup migration or tilting was observed. Ascribing the same power to each individual risk factor, the sum was computed, which defined a risk factor index for each case. Microscopic (Rt on worn regions) and macroscopic (penetration rates) wear features seemed to be significantly correlated with this risk factor index (r = 0.63, p = 0.005; r = 0.67, p = 0.002, respectively).
Furthermore, severely worn components were associated with a significantly increased risk factor index than that in the components with stripe wear (p = 0.003) and components with low wear (p = 0.0001) (Fig 4). Between these two last groups, the difference was not significant (p = 0.06).
In the current study, various and complementary techniques were used to accurately quantify the wear patterns of a series of retrieved alumina-on-alumina hip prostheses. Usually, in clinical wear studies, wear rates of hip components may be assessed by radiographic analysis or by direct measurement of explanted components. 4,8,12,14,17,20,28,31 However, the use of an accurate method of measurement is essential, especially when dimension changes lower than 100 μm have to be recorded. The use of a coordinate measuring machine with a resolution less than 1 μm fulfilled these requirements. 9 However, the precision of the results might be impaired by the difficulty in precisely defining the center of the components. However, this source of error only was observed in the case of severe wear spreading over the entire surface. In the most frequent cases of stripe and low wear patterns, large unworn surfaces were identified easily allowing a precise definition of component’s center.
Another limitation, always associated with this type of retrieval study, also should be mentioned. This series of implants, obtained at the time of revision for aseptic loosening of the socket, is not representative of the entire group but rather represents a worst case selection. Moreover, the number of available components was limited and did not correspond to a homogeneous population. Patients were either young, active, tall, and heavy males or older, less active, smaller, and lighter females. Furthermore, these components were two different designs: one was a cemented massive alumina cup; the other was noncemented held via a Ti threaded shell. However, knowing all these limitations, a comprehensive retrieval study was performed and reliable data were obtained on the in vivo alumina wear behavior.
Macroscopically, a classification into three groups was established according to the wear patterns: severe, stripe, and low wear. Quantitative analysis revealed that on a macroscopic scale, components with low and stripe wear had similar patterns whereas on a microscopic scale, components with stripe wear had roughness values similar to those obtained with severely worn components. As a result, these three groups may be characterized as follows: (1) components belonging to the low wear group (Group 3) had a limited wear on microscopic and macroscopic scales (Rt < 1 μm and penetration < 50 μm); (2) components with stripe wear (Group 2) had rougher surfaces than components in Group 1; the Rt value remained below 2 μm and contained localized worn regions in which depths from 10 to 150 μm were present; and (3) the severely worn components (Group 1) were characterized by the extent of the worn region visible on both components and corresponding to a penetration larger than 150 μm.
Among the 11 available components, two had evidence of massive wear. The remaining nine pairs had linear wear rates below 15 μm/year. This work confirmed that the wear of alumina ceramic components may take two different forms: it may be very limited with a negligible incidence on the long-term behavior of the system or it may be catastrophic, leading to a rapid destruction of the bearing surfaces.
Published alumina wear rates measured in vivo seem to be highly variable. Boutin and Blanquaert 2 reported wear rates ranging from 5 to 9 μm/year, whereas Mittelmeier and Heise1 22 reported a wear rate of approximately 10 μm/year. Plitz and Griss 26 gave an example of massive wear: approximately 0.9 mm after 3 years of implantation. These discrepancies may be related to material and design considerations. In fact, these abovementioned data were published more than 15 years ago and since that time numerous improvements have been achieved in terms of alumina quality and prosthesis designs. In the current series, the two severely worn components were implanted before 1980. After that time, wear rates always remained below 15 μm/year with a median value of 5 μm/year of use. To explain the occurrence of massive wear, various statements also have been published. Plitz and Griss, 26 Plitz and Hoss, 27 and Walter and Plitz 35,36 reported gross wear on the retrieved prostheses they examined and stated that such a wear is unavoidable and is related to the intrinsic nature of alumina. This was in contrast to Heimke and Griss, 15 who stated that severe wear was associated with exceptional situations (cup inclination > 50°). Among prostheses reviewed in the current study, none broke, but two had severe wear features. They were not removed early as in the series of Boutin et al, 3 but were in use for more than 15 years. The occurrence of this exaggerated wear seemed to be the result of a combination of unfavorable factors and with a delayed time to revision. The loosed prostheses still were functioning, leading to poor biomechanical conditions. Alumina quality improvements have led to the disappearance of rapid catastrophic wear. The sole limitation of the use of alumina-on-alumina hip prostheses seems to be the occurrence of socket loosening, and no longer is there a risk of massive wear, mechanical failure, or both.
In an attempt to identify the main risk factors, wear of alumina components can be related to material and biomechanical and clinical aspects. 3,11,15 The primary risk factor was the alumina quality and more specifically the grain sizes that seemed to be directly related to microscopic wear patterns. Under wear, alumina grains are pulled out leading to increasing defects as the grain sizes enlarge resulting in rougher surfaces. 18 However, no clear relationship was observed between alumina quality and macroscopic wear features. Thus, if grain size is implicated in the wear process, it is not the sole parameter.
On the clinical side, an important factor was the weight of the patients. The biomechanical aspects encompassed the positioning of the prosthesis, and the loosening (migration, tilting or both of the socket, by modifying the contact geometry, also can promote wear. From a laboratory study, 33 it is known that the volume of material removed by wear is proportional to the load applied and the sliding distance of the articulating surfaces. In the specific case of alumina wearing against alumina, wear volumes are related to the roughness of the surfaces. 13 Translated into clinical terms, the volume of material removed each year is proportional to the amount of penetration. 16 The load was variable but would be proportional to the mass of the patient, and the sliding distance would be determined by the time the joint had been in service and by the level of the patient’s activity. 6 The activity generally was inversely related to age. 34 Thus, penetration rate, which is a measurement of wear rate, should be proportional to patient mass, femoral head roughness, and, to a lesser extent, inversely proportional to age.
The lack of correlation between roughness parameters and penetration rates may be explained by the fact that studies dealing with the effect of surface roughness on wear rates have been performed with uniform rough surfaces and not with surface damages such as patches of relatively large, discrete scratches on an otherwise undamaged surface, as encountered with retrieved components. 9 Dowson et al 9 investigated this situation in the case of polyethylene wear and observed that single scratches could be responsible for dramatic wear.
In addition, any departure from an optimal cup inclination of 45° was associated with increased microscopic wear magnitude. On a macroscopic scale, a socket that was too vertical seemed to increase wear rates. As reported previously by several authors, 8,23 in those exceptional situations the contact area between the femoral head and the socket is decreased and the maximum load is transferred to the head by the edge of the socket. The higher than normal stress then exerted on the ceramic surfaces is responsible for the grain excavation process to occur. 3,23
Furthermore, alumina wear on microscopic and macroscopic scales seemed to be time independent. This result, which already has been reported in multiple publications, 3,8,14,16 is in favor with the hypothesis that wear magnitude was the result of a combination of unfavorable factors rather than the unavoidable fate of alumina bearings.
Moreover, the lack of relationship between macroscopic and microscopic wear features, and between macroscopic wear features and clinical data, suggest that no single mechanism is predominant in causing loosening. The occurrence of low wear even on microscopic scale, indicates that numerous mechanisms are contributing in varying degrees to failure in a one prosthesis. Aseptic loosening may not be attributable to alumina wear as the dominant factor in contrast to the wear occurring in a metal-on-polyethylene sliding couple.
1. Boutin P: [Total arthroplasty of the hip by fritted aluminum prosthesis. Experimental study and 1st clinical applications]. Rev Chir Orthop Reparatrice Appar Mot 58:229–246, 1972.
2. Boutin P, Blanquaert D: [A study of the mechanical properties of alumina-on-alumina total hip prosthesis]. Rev Chir Orthop Reparatrice Appar Mot 67:279–287, 1981.
3. Boutin P, Christel P, Dorlot JM, et al: The use of dense alumina-alumina ceramic combination in total hip replacement. J Biomed Mater Res 22:1203–1232, 1988.
4. Bragdon CR, Jasty M, Kawate K, et al: Wear of retrieved cemented polyethylene acetabula with alumina femoral heads. J Arthroplasty 12:119–25, 1997.
5. Brand RA, Pedersen DR, Yoder SA: How definition of “loosening” affects the incidence of loose total hip reconstructions. Clin Orthop 210:185–191, 1986.
6. Callaghan JJ, Salvati EA, Pellicci PM, et al: Results of revision for mechanical failure after cemented total hip replacement, 1979 to 1982. J Bone and Joint Surg 67A:1074–1085, 1985.
7. Derbyshire B, Hardaker CS, Fisher J, et al: Assessment of the change in volume of acetabular cups using a coordinate measuring machine. Proc Inst Mech Eng [H] 208:151–158, 1994.
8. Dorlot JM, Christel P, Meunier A: Wear analysis of retrieved alumina heads and sockets hip prostheses. J Biomed Mater Res Appl Biomater 23:299–310, 1989.
9. Dowson D, Taheri S, Wallbridge NC: The role of counterface imperfetions in the wear of polyethylene. Wear 119:227–293, 1987.
10. Fruh HJ, Willmann G, Pfaff HG: Wear characteristics of ceramic-on-ceramic for hip endoprostheses. Biomaterials 18:873–876, 1997.
11. Griss P, Heimke G: Five years experience with ceramic-metal-composite hip endoprostheses. Part I. Clinical evaluation. Arch Orthop Trauma Surg 98: 157–164, 1981.
12. Gualtieri G, Nigrisoli M, Dallari D, et al: A study of the mechanical and biological behavior of bioceramic hip arthroplasty followed-up after more than 10 years. Chir Organi Mov 78:213–222, 1993.
13. Hailey JL, Ingham E, Stone M, et al: Ultra-high molecular weight polyethylene wear debris generated in vivo and in laboratory tests; The influence of counterface roughness. Proc Inst Mech Eng [H] 210: 3–10, 1996.
14. Hall RM, Unsworth A, Siney P, et al: The surface topography of retrieved femoral heads. J Mater Sci 7:739–744, 1996.
15. Heimke G, Griss P: Five years experience with ceramic-metal-composite hip endoprostheses. Part II. Mechanical evaluation and improvements. Arch Orthop Trauma 98:165–171, 1981.
16. Isaac GH, Wroblewski BM, Atkinson JR, et al: A tribological study of retrieved hip prostheses. Clin Orthop 276:115–125, 1992.
17. Jazrawi LM, Bogner E, Della Valle CJ, et al: Wear rates of ceramic-on-ceramic bearing surfaces in total hip implants: a 12-year follow-up study. J Arthroplasty 14:781–787, 1999.
18. Lancaster JK, Mashal YA, Atkins AG: Particle Detachment Processes in the Dry and Lubricated Wear of Ceramics. In Dowson D (ed). Wear Particles. New York, Elsevier Science Publishers BV 237–246, 1992.
19. Leray JL, Cristel P: Tissue Tolerance of Wear Debris from Arthroplasties. In Winter GD, Gibbons DF, Plenk H (eds). Biomaterials 1980. Advances in Biomaterials. Vol 3. New York, John Wiley & Sons 197–206, 1982.
20. McGovern TE, Black J, Jacobs JJ, et al: In vivo wear of Ti6A14V femoral heads: A retrieval study. J Biomed Mater Res 32:447–457, 1996.
21. McKellop H, Clarke IC, Markolf KL, et al: Friction and wear properties of polymer, metal and ceramic prosthetic joint materials evaluated on a multichannel screening device. J Biomed Mater Res 5:619–653, 1981.
22. Mittelmeier H, Heisel J: Sixteen-years’ experience with ceramic hip prostheses. Clin Orthop 282:64–72, 1992.
23. Nevelos AB, Evans PA, Harrison P, et al: Examination of alumina ceramic components from total hip arthroplasties. Proc Inst Mech Eng [H] 207:155–162, 1993.
24. Nevelos JE, Ingham E, Doyle C, et al: Analysis of retrieved alumina ceramic components from Mittelmeier total hip prostheses. Biomaterials 20:1833–1840, 1999.
25. Nizard RS, Sedel L, Christel P, et al: Ten-year survivorship of cemented ceramic-ceramic total hip prosthesis. Clin Orthop 282:53–63, 1992.
26. Plitz W, Griss P: Clinical, Histo-Morphological and Material Related Observations on Removed Alumina-Ceramic Hip Joint Components. In Weinstein A, Gibbons D, Brown S, Ruff W (eds). Implant Retrieval: Material and Histological Analysis. Washington DC, National Bureau of Standards Specific Publication 601, United States Department of Commerce 131–156, 1981.
27. Plitz W, Hoss HU: Wear of Alumina-ceramic Hip-joints: Some Clinical and Tribological Aspects. In Winter GD, Gibbons D, Plenk H (eds). Biomaterials 1980. Advances in Biomaterials. Vol 3. New York, John Wiley & Sons 187–196, 1982.
28. Plitz W, Walter A, Jager M: [Material-specific wear of ceramic/ceramic sliding surfaces in revised hip endoprostheses—clinical and technological considerations]. Z Orthop Ihre Grenzgeb 12:299–303, 1984.
29. Refior HJ, Plitz W, Walter A: Ex Vivo and In Vitro Analysis of the Alumina Alumina Bearing System for Hip Joint Prostheses. In Sedel L, Rey C (eds). Bioceramics Vol 10. Oxford, UK, Elsevier Science Ltd 127–130, 1997.
30. Sedel L, Nizard RS, Kerboull L, et al: Alumina-alumina hip replacement in patients younger than 50 years old. Clin Orthop 298:175–183, 1994.
31. Sychterz CJ, Engh Jr CA, Swope SW, et al: Analysis of prosthetic femoral heads retrieved at autopsy. Clin Orthop 358:223–234, 1999.
32. Ungersböck A, Rahn B: Methods to characterize the surface roughness of metallic implants. J Mater Sci Mater Med 5:434–440, 1994.
33. Walker PS, Bullough PG: The effects of friction and wear in artificial joints. Orthop Clin North Am 4:275–293, 1973.
34. Wallbridge N, Dowson D: The walking activity of patients with artificial hip joints. Eng Med 11:95–96, 1982.
35. Walter A, Plitz W: Wear of Retrieved Alumina-Ceramic Hip-Joints Prostheses. In Vincenzini P (ed). Ceramics in Surgery. Amsterdam, Elsevier Science Ltd 253–259, 1983.
36. Walter A, Plitz W: Wear Mechanisms of Alumina-Ceramic Bearing Surfaces of Hip-Joints Prostheses. In Ducheyne P, Van de Perre G, Aubert AE (eds). Biomaterials and Biomechanics. Amsterdam, Elsevier Science Ltd 55–60, 1984.
Jonathan P. Garino, MD; and Laurent Sedel, MD Guest Editors