- Email: [email protected]
Finite element analysis of the cervico-trochanteric stemless femoral prosthesis
Clinical Biomechanics 18 (2003) S53–S58 www.elsevier.com/locate/clinbiomech
Finite element analysis of the cervico-trochanteric stemless femoral prosthesis Ching-Lung Tai a
a,b
, Chun-Hsiung Shih b,c, Weng-Pin Chen a,*, Shiuann-Sheng Lee b, Yu-Liang Liu a, Pang-Hsin Hsieh b, Wen-Jer Chen b
Department of Biomedical Engineering, Chung Yuan Christian University, Chungli 320, Taiwan, ROC Department of Orthopedic Surgery, Chang Gung Memorial Hospital, Kweishan 333, Taiwan, ROC c Department of Orthopedic Surgery, Chung Shan Hospital, Taipei 106, Taiwan, ROC
b
Abstract Objective. To investigate the biomechanical performance of a newly designed cervico-trochanteric stemless prosthesis by comparing the stress distribution with that of the traditional stem-type porous-coated anatomic prosthesis. Design. Three-dimensional finite element models were created for the intact femur, cervico-trochanteric implanted femur and porous-coated anatomic implanted femur. The stress distributions on the femur and the implant were compared. The effects of using two or three screws fixation for the cervico-trochanteric implanted femur were also investigated. Background. Local bone loss after implantation of traditional stem-type prostheses remains an unsolved problem during the long-term application of total hip replacement. The stress shielding effect and osteolysis were thought to be the two main factors that result in local bone loss after prosthesis implantation. In order to eliminate the mechanical and the biological causes of bone loss after total hip arthroplasty, a newly designed stemless femoral prosthesis was investigated. Methods. Three-dimensional finite element models were created for the intact, cervico-trochanteric (with two or three fixation screws), and porous-coated anatomic implanted femora with the geometry of a standardized composite femur. Analysis was performed for a loading condition simulating the single-legged stance. The von Mises stress distributions of each model were analyzed and compared. Results. The results can be summarized as follows: (1) Von Mises stress in the proximal, medial femur for the cervico-trochanteric implanted model was higher than that of the intact model and the porous-coated anatomic implanted model; (2) stress-shielding effect of the cervico-trochanteric models (with two or three fixation screws) were eliminated as compared with the porous-coated anatomic model; (3) no obvious difference in von Mises stress distribution for the cervico-trochanteric implanted model with two or three fixation screws. Conclusions. The cervico-trochanteric femoral prosthesis may reduce the stress-shielding effect of the proximal femur and achieve a more physiological stress distribution on the proximal femur than that of the porous-coated anatomic prosthesis. Relevance The new concept of cervico-trochanteric stemless prosthesis has proven to possess several advantages based on the current results, and may be an alternative for traditional stem-type prostheses in future clinical applications. Ó 2003 Elsevier Science Ltd. All rights reserved. Keywords: Cervico-trochanteric stemless femoral prosthesis; Finite element analysis; Stress-shielding effect; Composite femur; Osteolysis
1. Introduction Since the introduction of Charnley hip prosthesis in the early 1960s, total hip arthroplasty (THA) has proven
*
Corresponding author. E-mail address: [email protected] (W.-P. Chen).
to be a successful surgical procedure due to the improvements of prosthetic design, biomaterials and surgical technique (Harris, 1984; Callaghan, 1992). However, there is still a great concern of bone loss associated with stress shielding and osteolysis (Kroger et al., 1997; McCarthy et al., 1991; Chao and Coventry, 1981). Stress shielding is a mechanical cause of bone loss and is characterized by adaptive remodeling changes in the proximal femoral cortical bone following stem
0268-0033/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(03)00085-8
S54
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
implantation (Marchetti et al., 1996; Lewis et al., 1984; Engh and Bobyn, 1988). The proposed mechanism of stress shielding is based on WolffÕs law of remodeling. The redistribution of stress results in a decrease in the bone mineral density around the proximal femur, which may influence the longevity of the prosthesis. Osteolysis is another cause of bone loss following THA. This biological process of periprosthetic bone resorption is associated with the generation of particulate debris, especially the polyethylene wear particles (Shih et al., 1994; Murray and Rushton, 1992). To eliminate these two well-known complications of THA, the surface replacement arthroplasty was once widely used in the 1980s (Trentani and Vaccarino, 1981; Head, 1981; Bassett et al., 1982; Buechel et al., 1994; Herbeterts et al., 1983). However, it was proved a failure in design because of significant interface stress concentration (Huiskes et al., 1985). New types of stemless prostheses were designed in the 1990s to solve the clinical problems of stress shielding and osteolysis with reduced interface stress that ensures secure fixation (Munting and Verhelpen, 1995; Shih et al., 1997). In the current study, we used the finite element method to investigate the biomechanical performance of a newly designed cervico-trochanteric (C-T) stemless prosthesis. Three-dimensional finite element models (FEMs) were created for the intact femur, C-T implanted femur, and the traditional stem-type porouscoated anatomic (PCA) implanted femur. The stress distributions on the femur were analyzed and compared among each models. The effects of using two or three fixation screws on the C-T prosthesis were also investigated.
2. Methods The newly designed C-T femoral prosthesis and the fixation screws were made of Titanium and are shown in Fig. 1. The three holes are made with inner threads for initial locking after fastening of the screws. The traditional PCA femoral prosthesis (Howmedica, Rutherford, NJ, USA) used in this investigation is a stem-type design made of cobalt–chrome (Co–Cr) alloy. The femora used for the current study were commercially available synthetic products (Pacific Research Laboratory Inc., Vashon Island, WA, USA). They were made from a composite glass fiber/epoxy resin material to form the cortices with the internal cavity being filled with polyurethane foam. Standard surgical procedures for the implantation of the PCA and the C-T (with three screws) prostheses into the synthetic femora were performed respectively by an orthopedic surgeon (one of the authors). Since the scattering effect of the Co–Cr alloy may cause an undistinguishable contour in the computed tomography (CT) image, a methacrylate-made prosthesis with identical geometry with the Co–Cr PCA prosthesis was made by molding procedure for the purpose of CT scanning. Due to the sharp difference in density between the composite cortex, polyurethane medullary canal, titanium alloy (C-T prosthesis and fixation screws), and methacryate (PCA prosthesis), accurate definition of the boundaries was facilitated. CT scan images of the intact femur, C-T implanted femur, and PCA implanted femur were obtained at 1.25 mm intervals in the transverse planes from the femoral head down to the mid femoral shaft using a GE Hispeed
Fig. 1. The newly designed C-T prosthesis: (a) inner side, (b) outer side, and (c) configuration with three screws.
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
scanner (General Electric, Milwaukee, WI, USA). The resolution for each of the CT scans was 512 by 512 pixels, the field of view was 330 mm, and the pixel size was 0.625 mm/pixel. The cross-sectional image files of the femur and the implant were transferred to a customwritten automatic contouring program for the detection of the contours for the cortical bone, cancellous bone, and prosthesis. The three sets of parallel contours were then input into the Solidworks 99 CAD program (Solidworks Corp., Boston, MA, USA) for the reconstruction of 3-D solid models. In the Solidworks program, the C-T (with three screws) implanted femur was modified by removing the third screw (distal one) to create the C-T (with two screws) implanted femur. The four solid models were then transferred to a finite element pre- and post-processing program––Mentat 2000 (MSC Software Corp., Los Angeles, CA, USA) for the generation of 3-D meshes. The 10-node second-order tetrahedral elements were automatically generated in the Mentat 2000 program. The threads and the tips of the fixation screws were not modeled in the FEM in order to simplify the model setup. The geometry of the screw was assumed cylindrical with 6 mm in diameter and 36 mm in length as obtained from the measurement of the actual screw. Four different FEMs were created, namely, the intact femur, C-T (with two screws) implanted femur, C-T (with three screws) implanted femur, and PCA implanted femur. All materials were assumed to be linear, elastic, homogeneous and isotropic (Mann and Bartel, 1995; Mann et al., 1997; McNamara et al., 1997). According to the specifications provided by the manufacturer, the PoissonÕs ratios for all materials were set to be 0.3 and the modulus of elasticity used for the cortical bone (composite glass fiber/epoxy resin), cancellous bone (polyurethane foam), C-T (titanium) and PCA (Co–Cr alloy) prosthesis were set to be 14.2 GPa, 50 MPa, 110 GPa and 220 GPa, respectively. The numbers
S55
Fig. 2. The loading configuration of generated 3-D finite meshes for the (a) intact, (b) C-T, and (c) PCA implanted femora.
of elements used for each material in the FEMs are listed in Table 1. Analysis was performed for the loading condition simulating the single-legged stance as adopted from the literature (Mann and Bartel, 1995). Fig. 2 illustrates the three finite element meshes of the intact femur, the C-T (with three screws) implanted femur, and the PCA implanted femur with the loading condition indicated. The distal ends of the femoral models were constrained as boundary condition. The detailed meshes for the C-T prosthesis, two fixation screws, three fixation screws, and the PCA prosthesis are shown in Fig. 3. The bone/implant and screw/bone/implant interfaces in the implanted models were assumed to be fully bonded without taking into account the micromotions. The finite element analysis for each of the four models were performed using the M A R C 2000 finite element software (MSC Software Corp., Los Angeles, CA, USA) on an HP SPP/2000 supercomputer at the National Center for High-Performance Computing (NCHC, Hsinchu, Taiwan). The analysis results were retrieved back to a local workstation for post-processing.
Table 1 The number of elements for each material in the FEMs Model
Material
No. of elements
Total no. of elements
Intact
Cortical bone Cancellous bone
10,400 15,104
25,504
PCA
Cortical bone Cancellous bone PCA prosthesis (Co–Cr)
5325 5131 1198
11,654
C-T (two screws)
Cortical bone Cancellous bone C-T prosthesis (Ti) Screws (2)
8322 10,484 2831 1030
22,667
C-T (three screws)
Cortical bone Cancellous bone C-T prosthesis (Ti) Screws (3)
8378 11,746 3029 1676
24,829
S56
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58 Table 2 Maximum and minimum principal stresses of the fixation screws in the C-T implanted femur (MPa)
Maximum Minimum
Fig. 3. The 3-D FEMs of (a) C-T prosthesis, (b) two fixation screws, (c) three fixation screws, and (d) PCA prosthesis.
3. Results The von Mises equivalent stress distributions on the medial side for each of the femoral models: intact femur, C-T (with two screws) implanted femur, C-T (with three screws) implanted femur, and the PCA implanted femur were obtained from the analyses and shown in Fig. 4 with gray-scale fringes. As shown in Fig. 4, there was a local high stress on the region of the proximal femur adjacent to the distal edge of the C-T prosthesis. This was caused by the stress-concentration effect near the edge of the C-T prosthesis. The maximum and minimum principal stresses of the fixation screws of the C-T implanted femur is shown in Table 2. For the three fixation screws in the C-T implanted femur, the proximal-posterior screw was subjected to the highest and also the lowest principal stresses simultaneously under loading, and the absolute value of the maximum and minimum principal stresses were almost equal.
Fig. 4. The overall von Mises stress distribution on the medial side of the femur for (a) intact, (b) C-T (two screws), (c) C-T (three screws), and (d) PCA implanted models. (unit: MPa).
Distal
Proximalposterior
Proximalanterior
5.70 )18.28
24.57 )24.48
13.25 )9.37
In summary, the stress results can be summarized as follows: (1) von Mises stress in the proximal, medial femur for the C-T model was higher than that of the intact model and the PCA model; (2) stress-shielding effect of the C-T models (with two or three fixation screws) were obviously eliminated as compared to that of the PCA model; (3) no obvious difference in von Mises stress distribution for the C-T model with two or three fixation screws.
4. Discussion The finite element method has become a useful tool in analyzing the stresses in structures of complex shapes, loading and material behavior. Numerous applications in orthopedics have been presented and proven to be a successful tool in predicting the mechanical characteristics of skeletal parts in interesting circumstances (Cody et al., 1999; Testi et al., 1999). The convergence of FEM model plays an important role on the reliability of the final results, whereas the method used in this study to prove the convergence of model is to calculate the total strain energy of the structure. Three models of different numbers of element and node were created to perform the convergence test, and the results of the total strain energy for the three models were all within a 5% difference. The convergence test demonstrated the validity of the models established. In the C-T implanted model, the C-T prosthesis tended to move anteriorly and distally due to the moment caused by the applied forces. As shown in Table 2, it was easily understood that the distal screw was subjected to a major compressive mode while the proximal screws were subjected to a major tensile mode. In addition, both the principal tensile and compressive stresses on the proximal-posterior screw were larger than those on the proximal-anterior screw due to the anteversion of the femoral neck. Therefore, the proximal-posterior screw was subjected to the largest bending stress as compared to the other two screws under loading. Based on the result, it is reasonable to believe that the proximal-posterior screw might be a dominant component that contributed to the resistance against the major pullout force in the C-T implanted model with three screws fixation. Besides this, since there is no obvious difference in von Mises stress distribution between the two and three screws fixation conditions for the C-T
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
implanted femur (Fig. 4), the addition of the distal screw might not contribute to the elimination of stress shielding. However, it deserves to be further investigated by experimental testing to demonstrate if the addition of the distal screw improves the maximal bearing load. As shown in Fig. 4, a local high stress was found at the region of the medial proximal femur adjacent to the distal edge of the C-T prosthesis, this phenomenon was induced from the oppression of the femoral cortex by the distal edge of the C-T prosthesis under inclined resultant force acting on the prosthesis, which leads to the local high stress concentration. On the viewpoint of long-term application, the risk of bone fracture due to the existence of the local high stress concentration is of concern. A further experimental testing to measure the maximum bearing load and to observe the fracture site of the C-T implanted femur deserves to be performed and is now being carried out. Besides, the stress values at the medial proximal femur are apparently lower in the PCA model as compared with those in the intact and the C-T models (Fig. 4), which revealed significant stress-shielding phenomenon occurring after implantation of the PCA prosthesis. On the other hand, the stress distributions in the C-T model were higher than those in the intact model. The radiograph showing the offset of the C-T (53.9 mm), intact (47.7 mm) and the PCA (47.8 mm) implanted femora is shown in Fig. 5. By using the known value of the metal ball diameter (32 mm), the offset could be calculated for these models. Since the difference of the offset between the PCA (47.8 mm) and intact (47.7 mm) models was not obvious, the phenomenon of stressshielding effect induced by the implantation of traditional stem-type prosthesis was thus observed from the low von Mises stress distributions on the proximal femur of the PCA model in the current study (Fig. 4). On the other hand, the offset of the C-T model (53.9 mm) was obviously larger than that of the intact model (47.7
Fig. 5. The radiograph showing the offset of the (a) C-T (53.9 mm), (b) intact (47.7 mm), and (c) PCA (47.8 mm) implanted femora.
S57
mm), from a mechanical viewpoint, the larger offset of the C-T implanted femur might cause higher moment as compared with that of the intact femur. This might be the reason why the proximal stress of the C-T model was larger than that of the intact model (Fig. 4). Theoretically, since the geometry and constitution of the femur distal to the edge of the C-T prosthesis are identical between the C-T and intact models, both the moment of inertia (I) and the modulus of elasticity (E) should be identical, which leads to an identical flexural rigidity (EI) for these two models. Under circumstances of the same bending moment (joint force times offset), the stress or strain distribution of the femur distal to the edge of the C-T prosthesis will be identical for the C-T and intact models. In the current study, although the offset of the C-T model was larger than that of the intact model, the authors postulate that if in the future when the C-T prosthesis is fabricated with strict dimension control, under circumstances of identical offset between the intact and the C-T implanted femur, the stress or strain distribution of the C-T model will be more consistent with that of the intact model. Based on the results, we believe this stemless-type prosthesis would reduce the incidence of the stressshielding effect in the clinical application. Furthermore, the new C-T prosthesis covers all of the femoral neck and the greater trochanter area, allowing no space for wear debris to access the host bone, thus preventing the periprosthetic osteolysis. Besides this, it is noted that poor distal fixation of the traditional stem-type prosthesis may produce excessive micromotion between stem/bone interface. This leads to the failed bone-ingrowth scenario with the side effects of thigh pain and the development of a fibrous interface between the stem and bone (Whiteside, 1989). Therefore, the introduction of the C-T stemless prosthesis was considered to be able to prevent the complications of thigh pain and growth of fibrous tissue. Although the geometry of the C-T prosthesis was an enormous change as compared with that of traditional stem-type prostheses, it made no change on other components such as the metal ball, the insert, and the cup. The C-T prosthesis thus complies with the traditional prosthesis, and may alternatively replace the stem component of other modular type prostheses. In our future work, the porous coating on the inner surface of the C-T prosthesis is proposed to promote bony ingrowth for long-term consideration. In the present study, the interface between the C-T prosthesis and the underlying bone was assumed fully bonded in the FEM. This may be appropriate when the bone has fully ingrown into the porous coated surface of the C-T prosthesis. Referring to our previous study (Shih et al., 1997), the C-T stemless prosthesis was implanted with three penetrated screws and fixed with metal plate and nuts on the lateral side. The results also demonstrated that the stress-shielding effect was significantly eliminated after
S58
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
implantation of the C-T prosthesis. However, the potential drawbacks caused by the penetrated screw tips and high stress concentration caused by the fixation nuts were of concerns. In the current study, the C-T prosthesis was fixed with two or three screws without penetration of the opposite cortex. The potential drawbacks caused by the screws tips and nuts, thus, no longer existed. However, since single-cortex purchase of the screws may lead to inadequate fixation of the prosthesis on the cancellous bones, the capacity of load bearing for the C-T implanted femur deserves further investigation. By using cadaveric femora, an experimental testing is now being carried out to measure the surface strains and the maximal bearing load of the C-T implanted femur.
5. Conclusions The present study demonstrates that the stress shielding of the C-T implanted femur was significantly eliminated as compared to that of the traditional PCA implanted femur. Moreover, the C-T prosthesis possesses the following features, and might be an alternative for traditional stem-type prostheses for clinical application after future clinical trials: (1) prevention of the stress-shielding effect (decrease bone atrophy); (2) prevention of endosteal osteolysis; (3) preservation of bone stock for easy revision; (4) complied with conventional THR system.
Acknowledgements This study was supported by the grant from the National Science Council of the Republic of China (grant no. NSC-89-2320-B-033-001-M08). The computing facilities provided by the National Center for HighPerformance Computing are greatly appreciated.
References Bassett, L.W., Gold, R.H., Hedley, A.K., 1982. Radiology of failed surface replacement total hip arthroplasty. Am. J. Roentgenol. 139, 1083–1088. Buechel, F.F., Drucker, D., Jasty, M., Jiranek, W., Harris, W.H., 1994. Osteolysis around uncemented acetabular components of cobalt–chrome surface replacement hip arthroplasty. Clin. Orthop. Relat. R 298, 202–211. Callaghan, J.J., 1992. Total hip arthroplasty: clinical perspective. Clin. Orthop. Relat. R 276, 33–40. Chao, E.Y., Coventry, M.B., 1981. Fracture of femoral component after total hip replacement. An analysis of 58 cases. J. Bone Joint Surg. Am. 63A, 1078–1094.
Cody, D.D., Gross, G.J., Hou, J.H., Spencer, H.J., Glodstein, S., Fyhrie, D.P., 1999. Femoral strength is better predicted by finite element models than QCT and DXA. J. Biomech. 32, 1013–1020. Engh, C.A., Bobyn, J.D., 1988. The influence of stem size and extent of porous coating on femoral bone resorption after primary cementless hip arthroplasty. Clin. Orthop. Relat. R 231, 7–28. Harris, W.H., 1984. Will stress shielding limit the longevity of cemented femoral components of total hip replacement. Clin. Orthop. Relat. R 274, 120–123. Head, W.C., 1981. Wagner surface replacement arthoplasty of the hip. J. Bone Joint Surg. Am. 63A, 420–427. Herbeterts, P., Lansinger, O., Romanus, B., 1983. Surface replacement arthroplasty of the hip. Acta. Orthop. Scand. 54, 884–890. Huiskes, R., Strens, P., Heck, J.V., Sloof, T., 1985. Interface stress in the resurfaced hip. Acta Orthop. Scand. 56, 474–478. Kroger, H., Vanninen, E., Overmyer, M., Miettinen, H., Rushton, N., Suomalainen, O., 1997. Periprosthetic bone loss and regional bone turnover in uncemented total hip arthroplasty: a prospective study using high resolution single photon emission tomography and dualenergy X-ray absorptiometry. J. Bone Miner. Res. 12 (3), 487–492. Lewis, J.L., Askew, M.J., Wixson, R.L., Kramer, G.M., Tarr, R.R., 1984. The influence of prosthetic stem stiffness and of a calcar collar on stresses in the proximal end of the femur with a cemented femoral components. J. Bone Joint Surg. Am. 66A, 280–286. Mann, K.A., Bartel, D.L., 1995. Coulomb frictional interfaces in modeling cemented total hip replacement: a more realistic model. J. Biomech. 9, 1067–1078. Mann, K.A., Bartel, D.L., Ayers, D.C., 1997. Influence of stem geometry on mechanics of cemented femoral hip components with a proximal bond. J. Orthop. Res. 15, 700–706. Marchetti, M.E., Steinberg, G.G., Greene, J.M., Jenis, L.G., Baran, D.T., 1996. A prospective study of proximal femur bone mass following cemented and uncemented hip arthroplasty. J. Bone Miner. Res. 11 (7), 1033–1039. McCarthy, C.K., Steinberg, G.G., Agren, M., Leahey, D., Wyman, E., Baran, D.T., 1991. Quantifying bone loss from the proximal femur after total hip arthroplasty. J. Bone Joint Surg. [Br] 73B, 774–778. McNamara, B.P., Cristofolini, L., Toni, A., Taylor, D., 1997. Relationship between bone-prosthesis bonding and load transfer in total hip reconstruction. J. Biomech. 30, 621–630. Munting, E., Verhelpen, M., 1995. Fixation and effect on bone strain pattern of a stemless hip prosthesis. J. Biomech. 28, 949–961. Murray, D.W., Rushton, N., 1992. Mediators of bone resorption around implants. Clin. Orthop. Relat. R 281, 295–304. Shih, C.H., Chen, W.P., Tai, C.L., Kuo, R.F., Wu, C.C., Chen, C.H., 1997. New concepts-biomechanical studies of a newly design femoral prosthesis (cevico-trochanter prosthesis). Clin. Biomech. 12, 482–490. Shih, C.H., Wu, C.C., Lee, Z.L., Yang, W.E., 1994. Localized femoral osteolysis in cementless ceramic total hip arthroplasty. Orthop. Rev. 23, 325–328. Testi, D., Viceconti, M., Brauffaldi, F., Cappello, A., 1999. Risk of fracture in elderly patients: a new predictive index based on bone mineral density and finite element analysis. Comput. Meth. Prog. Bio. 60, 23–33. Trentani, C., Vaccarino, F., 1981. Complications in surface replacement arthroplasty of the hip: experience with the paltrinieritrentani prosthesis. Int. Orthop. (SICOT) 4, 247–252. Whiteside, L.A., 1989. The effect of stem fit on bone hypertrophy and pain relief in cementless total hip arthroplasty. Clin. Orthop. Relat. R 247, 138–147.
Finite element analysis of the cervico-trochanteric stemless femoral prosthesis Ching-Lung Tai a
a,b
, Chun-Hsiung Shih b,c, Weng-Pin Chen a,*, Shiuann-Sheng Lee b, Yu-Liang Liu a, Pang-Hsin Hsieh b, Wen-Jer Chen b
Department of Biomedical Engineering, Chung Yuan Christian University, Chungli 320, Taiwan, ROC Department of Orthopedic Surgery, Chang Gung Memorial Hospital, Kweishan 333, Taiwan, ROC c Department of Orthopedic Surgery, Chung Shan Hospital, Taipei 106, Taiwan, ROC
b
Abstract Objective. To investigate the biomechanical performance of a newly designed cervico-trochanteric stemless prosthesis by comparing the stress distribution with that of the traditional stem-type porous-coated anatomic prosthesis. Design. Three-dimensional finite element models were created for the intact femur, cervico-trochanteric implanted femur and porous-coated anatomic implanted femur. The stress distributions on the femur and the implant were compared. The effects of using two or three screws fixation for the cervico-trochanteric implanted femur were also investigated. Background. Local bone loss after implantation of traditional stem-type prostheses remains an unsolved problem during the long-term application of total hip replacement. The stress shielding effect and osteolysis were thought to be the two main factors that result in local bone loss after prosthesis implantation. In order to eliminate the mechanical and the biological causes of bone loss after total hip arthroplasty, a newly designed stemless femoral prosthesis was investigated. Methods. Three-dimensional finite element models were created for the intact, cervico-trochanteric (with two or three fixation screws), and porous-coated anatomic implanted femora with the geometry of a standardized composite femur. Analysis was performed for a loading condition simulating the single-legged stance. The von Mises stress distributions of each model were analyzed and compared. Results. The results can be summarized as follows: (1) Von Mises stress in the proximal, medial femur for the cervico-trochanteric implanted model was higher than that of the intact model and the porous-coated anatomic implanted model; (2) stress-shielding effect of the cervico-trochanteric models (with two or three fixation screws) were eliminated as compared with the porous-coated anatomic model; (3) no obvious difference in von Mises stress distribution for the cervico-trochanteric implanted model with two or three fixation screws. Conclusions. The cervico-trochanteric femoral prosthesis may reduce the stress-shielding effect of the proximal femur and achieve a more physiological stress distribution on the proximal femur than that of the porous-coated anatomic prosthesis. Relevance The new concept of cervico-trochanteric stemless prosthesis has proven to possess several advantages based on the current results, and may be an alternative for traditional stem-type prostheses in future clinical applications. Ó 2003 Elsevier Science Ltd. All rights reserved. Keywords: Cervico-trochanteric stemless femoral prosthesis; Finite element analysis; Stress-shielding effect; Composite femur; Osteolysis
1. Introduction Since the introduction of Charnley hip prosthesis in the early 1960s, total hip arthroplasty (THA) has proven
*
Corresponding author. E-mail address: [email protected] (W.-P. Chen).
to be a successful surgical procedure due to the improvements of prosthetic design, biomaterials and surgical technique (Harris, 1984; Callaghan, 1992). However, there is still a great concern of bone loss associated with stress shielding and osteolysis (Kroger et al., 1997; McCarthy et al., 1991; Chao and Coventry, 1981). Stress shielding is a mechanical cause of bone loss and is characterized by adaptive remodeling changes in the proximal femoral cortical bone following stem
0268-0033/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(03)00085-8
S54
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
implantation (Marchetti et al., 1996; Lewis et al., 1984; Engh and Bobyn, 1988). The proposed mechanism of stress shielding is based on WolffÕs law of remodeling. The redistribution of stress results in a decrease in the bone mineral density around the proximal femur, which may influence the longevity of the prosthesis. Osteolysis is another cause of bone loss following THA. This biological process of periprosthetic bone resorption is associated with the generation of particulate debris, especially the polyethylene wear particles (Shih et al., 1994; Murray and Rushton, 1992). To eliminate these two well-known complications of THA, the surface replacement arthroplasty was once widely used in the 1980s (Trentani and Vaccarino, 1981; Head, 1981; Bassett et al., 1982; Buechel et al., 1994; Herbeterts et al., 1983). However, it was proved a failure in design because of significant interface stress concentration (Huiskes et al., 1985). New types of stemless prostheses were designed in the 1990s to solve the clinical problems of stress shielding and osteolysis with reduced interface stress that ensures secure fixation (Munting and Verhelpen, 1995; Shih et al., 1997). In the current study, we used the finite element method to investigate the biomechanical performance of a newly designed cervico-trochanteric (C-T) stemless prosthesis. Three-dimensional finite element models (FEMs) were created for the intact femur, C-T implanted femur, and the traditional stem-type porouscoated anatomic (PCA) implanted femur. The stress distributions on the femur were analyzed and compared among each models. The effects of using two or three fixation screws on the C-T prosthesis were also investigated.
2. Methods The newly designed C-T femoral prosthesis and the fixation screws were made of Titanium and are shown in Fig. 1. The three holes are made with inner threads for initial locking after fastening of the screws. The traditional PCA femoral prosthesis (Howmedica, Rutherford, NJ, USA) used in this investigation is a stem-type design made of cobalt–chrome (Co–Cr) alloy. The femora used for the current study were commercially available synthetic products (Pacific Research Laboratory Inc., Vashon Island, WA, USA). They were made from a composite glass fiber/epoxy resin material to form the cortices with the internal cavity being filled with polyurethane foam. Standard surgical procedures for the implantation of the PCA and the C-T (with three screws) prostheses into the synthetic femora were performed respectively by an orthopedic surgeon (one of the authors). Since the scattering effect of the Co–Cr alloy may cause an undistinguishable contour in the computed tomography (CT) image, a methacrylate-made prosthesis with identical geometry with the Co–Cr PCA prosthesis was made by molding procedure for the purpose of CT scanning. Due to the sharp difference in density between the composite cortex, polyurethane medullary canal, titanium alloy (C-T prosthesis and fixation screws), and methacryate (PCA prosthesis), accurate definition of the boundaries was facilitated. CT scan images of the intact femur, C-T implanted femur, and PCA implanted femur were obtained at 1.25 mm intervals in the transverse planes from the femoral head down to the mid femoral shaft using a GE Hispeed
Fig. 1. The newly designed C-T prosthesis: (a) inner side, (b) outer side, and (c) configuration with three screws.
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
scanner (General Electric, Milwaukee, WI, USA). The resolution for each of the CT scans was 512 by 512 pixels, the field of view was 330 mm, and the pixel size was 0.625 mm/pixel. The cross-sectional image files of the femur and the implant were transferred to a customwritten automatic contouring program for the detection of the contours for the cortical bone, cancellous bone, and prosthesis. The three sets of parallel contours were then input into the Solidworks 99 CAD program (Solidworks Corp., Boston, MA, USA) for the reconstruction of 3-D solid models. In the Solidworks program, the C-T (with three screws) implanted femur was modified by removing the third screw (distal one) to create the C-T (with two screws) implanted femur. The four solid models were then transferred to a finite element pre- and post-processing program––Mentat 2000 (MSC Software Corp., Los Angeles, CA, USA) for the generation of 3-D meshes. The 10-node second-order tetrahedral elements were automatically generated in the Mentat 2000 program. The threads and the tips of the fixation screws were not modeled in the FEM in order to simplify the model setup. The geometry of the screw was assumed cylindrical with 6 mm in diameter and 36 mm in length as obtained from the measurement of the actual screw. Four different FEMs were created, namely, the intact femur, C-T (with two screws) implanted femur, C-T (with three screws) implanted femur, and PCA implanted femur. All materials were assumed to be linear, elastic, homogeneous and isotropic (Mann and Bartel, 1995; Mann et al., 1997; McNamara et al., 1997). According to the specifications provided by the manufacturer, the PoissonÕs ratios for all materials were set to be 0.3 and the modulus of elasticity used for the cortical bone (composite glass fiber/epoxy resin), cancellous bone (polyurethane foam), C-T (titanium) and PCA (Co–Cr alloy) prosthesis were set to be 14.2 GPa, 50 MPa, 110 GPa and 220 GPa, respectively. The numbers
S55
Fig. 2. The loading configuration of generated 3-D finite meshes for the (a) intact, (b) C-T, and (c) PCA implanted femora.
of elements used for each material in the FEMs are listed in Table 1. Analysis was performed for the loading condition simulating the single-legged stance as adopted from the literature (Mann and Bartel, 1995). Fig. 2 illustrates the three finite element meshes of the intact femur, the C-T (with three screws) implanted femur, and the PCA implanted femur with the loading condition indicated. The distal ends of the femoral models were constrained as boundary condition. The detailed meshes for the C-T prosthesis, two fixation screws, three fixation screws, and the PCA prosthesis are shown in Fig. 3. The bone/implant and screw/bone/implant interfaces in the implanted models were assumed to be fully bonded without taking into account the micromotions. The finite element analysis for each of the four models were performed using the M A R C 2000 finite element software (MSC Software Corp., Los Angeles, CA, USA) on an HP SPP/2000 supercomputer at the National Center for High-Performance Computing (NCHC, Hsinchu, Taiwan). The analysis results were retrieved back to a local workstation for post-processing.
Table 1 The number of elements for each material in the FEMs Model
Material
No. of elements
Total no. of elements
Intact
Cortical bone Cancellous bone
10,400 15,104
25,504
PCA
Cortical bone Cancellous bone PCA prosthesis (Co–Cr)
5325 5131 1198
11,654
C-T (two screws)
Cortical bone Cancellous bone C-T prosthesis (Ti) Screws (2)
8322 10,484 2831 1030
22,667
C-T (three screws)
Cortical bone Cancellous bone C-T prosthesis (Ti) Screws (3)
8378 11,746 3029 1676
24,829
S56
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58 Table 2 Maximum and minimum principal stresses of the fixation screws in the C-T implanted femur (MPa)
Maximum Minimum
Fig. 3. The 3-D FEMs of (a) C-T prosthesis, (b) two fixation screws, (c) three fixation screws, and (d) PCA prosthesis.
3. Results The von Mises equivalent stress distributions on the medial side for each of the femoral models: intact femur, C-T (with two screws) implanted femur, C-T (with three screws) implanted femur, and the PCA implanted femur were obtained from the analyses and shown in Fig. 4 with gray-scale fringes. As shown in Fig. 4, there was a local high stress on the region of the proximal femur adjacent to the distal edge of the C-T prosthesis. This was caused by the stress-concentration effect near the edge of the C-T prosthesis. The maximum and minimum principal stresses of the fixation screws of the C-T implanted femur is shown in Table 2. For the three fixation screws in the C-T implanted femur, the proximal-posterior screw was subjected to the highest and also the lowest principal stresses simultaneously under loading, and the absolute value of the maximum and minimum principal stresses were almost equal.
Fig. 4. The overall von Mises stress distribution on the medial side of the femur for (a) intact, (b) C-T (two screws), (c) C-T (three screws), and (d) PCA implanted models. (unit: MPa).
Distal
Proximalposterior
Proximalanterior
5.70 )18.28
24.57 )24.48
13.25 )9.37
In summary, the stress results can be summarized as follows: (1) von Mises stress in the proximal, medial femur for the C-T model was higher than that of the intact model and the PCA model; (2) stress-shielding effect of the C-T models (with two or three fixation screws) were obviously eliminated as compared to that of the PCA model; (3) no obvious difference in von Mises stress distribution for the C-T model with two or three fixation screws.
4. Discussion The finite element method has become a useful tool in analyzing the stresses in structures of complex shapes, loading and material behavior. Numerous applications in orthopedics have been presented and proven to be a successful tool in predicting the mechanical characteristics of skeletal parts in interesting circumstances (Cody et al., 1999; Testi et al., 1999). The convergence of FEM model plays an important role on the reliability of the final results, whereas the method used in this study to prove the convergence of model is to calculate the total strain energy of the structure. Three models of different numbers of element and node were created to perform the convergence test, and the results of the total strain energy for the three models were all within a 5% difference. The convergence test demonstrated the validity of the models established. In the C-T implanted model, the C-T prosthesis tended to move anteriorly and distally due to the moment caused by the applied forces. As shown in Table 2, it was easily understood that the distal screw was subjected to a major compressive mode while the proximal screws were subjected to a major tensile mode. In addition, both the principal tensile and compressive stresses on the proximal-posterior screw were larger than those on the proximal-anterior screw due to the anteversion of the femoral neck. Therefore, the proximal-posterior screw was subjected to the largest bending stress as compared to the other two screws under loading. Based on the result, it is reasonable to believe that the proximal-posterior screw might be a dominant component that contributed to the resistance against the major pullout force in the C-T implanted model with three screws fixation. Besides this, since there is no obvious difference in von Mises stress distribution between the two and three screws fixation conditions for the C-T
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
implanted femur (Fig. 4), the addition of the distal screw might not contribute to the elimination of stress shielding. However, it deserves to be further investigated by experimental testing to demonstrate if the addition of the distal screw improves the maximal bearing load. As shown in Fig. 4, a local high stress was found at the region of the medial proximal femur adjacent to the distal edge of the C-T prosthesis, this phenomenon was induced from the oppression of the femoral cortex by the distal edge of the C-T prosthesis under inclined resultant force acting on the prosthesis, which leads to the local high stress concentration. On the viewpoint of long-term application, the risk of bone fracture due to the existence of the local high stress concentration is of concern. A further experimental testing to measure the maximum bearing load and to observe the fracture site of the C-T implanted femur deserves to be performed and is now being carried out. Besides, the stress values at the medial proximal femur are apparently lower in the PCA model as compared with those in the intact and the C-T models (Fig. 4), which revealed significant stress-shielding phenomenon occurring after implantation of the PCA prosthesis. On the other hand, the stress distributions in the C-T model were higher than those in the intact model. The radiograph showing the offset of the C-T (53.9 mm), intact (47.7 mm) and the PCA (47.8 mm) implanted femora is shown in Fig. 5. By using the known value of the metal ball diameter (32 mm), the offset could be calculated for these models. Since the difference of the offset between the PCA (47.8 mm) and intact (47.7 mm) models was not obvious, the phenomenon of stressshielding effect induced by the implantation of traditional stem-type prosthesis was thus observed from the low von Mises stress distributions on the proximal femur of the PCA model in the current study (Fig. 4). On the other hand, the offset of the C-T model (53.9 mm) was obviously larger than that of the intact model (47.7
Fig. 5. The radiograph showing the offset of the (a) C-T (53.9 mm), (b) intact (47.7 mm), and (c) PCA (47.8 mm) implanted femora.
S57
mm), from a mechanical viewpoint, the larger offset of the C-T implanted femur might cause higher moment as compared with that of the intact femur. This might be the reason why the proximal stress of the C-T model was larger than that of the intact model (Fig. 4). Theoretically, since the geometry and constitution of the femur distal to the edge of the C-T prosthesis are identical between the C-T and intact models, both the moment of inertia (I) and the modulus of elasticity (E) should be identical, which leads to an identical flexural rigidity (EI) for these two models. Under circumstances of the same bending moment (joint force times offset), the stress or strain distribution of the femur distal to the edge of the C-T prosthesis will be identical for the C-T and intact models. In the current study, although the offset of the C-T model was larger than that of the intact model, the authors postulate that if in the future when the C-T prosthesis is fabricated with strict dimension control, under circumstances of identical offset between the intact and the C-T implanted femur, the stress or strain distribution of the C-T model will be more consistent with that of the intact model. Based on the results, we believe this stemless-type prosthesis would reduce the incidence of the stressshielding effect in the clinical application. Furthermore, the new C-T prosthesis covers all of the femoral neck and the greater trochanter area, allowing no space for wear debris to access the host bone, thus preventing the periprosthetic osteolysis. Besides this, it is noted that poor distal fixation of the traditional stem-type prosthesis may produce excessive micromotion between stem/bone interface. This leads to the failed bone-ingrowth scenario with the side effects of thigh pain and the development of a fibrous interface between the stem and bone (Whiteside, 1989). Therefore, the introduction of the C-T stemless prosthesis was considered to be able to prevent the complications of thigh pain and growth of fibrous tissue. Although the geometry of the C-T prosthesis was an enormous change as compared with that of traditional stem-type prostheses, it made no change on other components such as the metal ball, the insert, and the cup. The C-T prosthesis thus complies with the traditional prosthesis, and may alternatively replace the stem component of other modular type prostheses. In our future work, the porous coating on the inner surface of the C-T prosthesis is proposed to promote bony ingrowth for long-term consideration. In the present study, the interface between the C-T prosthesis and the underlying bone was assumed fully bonded in the FEM. This may be appropriate when the bone has fully ingrown into the porous coated surface of the C-T prosthesis. Referring to our previous study (Shih et al., 1997), the C-T stemless prosthesis was implanted with three penetrated screws and fixed with metal plate and nuts on the lateral side. The results also demonstrated that the stress-shielding effect was significantly eliminated after
S58
C.-L. Tai et al. / Clinical Biomechanics 18 (2003) S53–S58
implantation of the C-T prosthesis. However, the potential drawbacks caused by the penetrated screw tips and high stress concentration caused by the fixation nuts were of concerns. In the current study, the C-T prosthesis was fixed with two or three screws without penetration of the opposite cortex. The potential drawbacks caused by the screws tips and nuts, thus, no longer existed. However, since single-cortex purchase of the screws may lead to inadequate fixation of the prosthesis on the cancellous bones, the capacity of load bearing for the C-T implanted femur deserves further investigation. By using cadaveric femora, an experimental testing is now being carried out to measure the surface strains and the maximal bearing load of the C-T implanted femur.
5. Conclusions The present study demonstrates that the stress shielding of the C-T implanted femur was significantly eliminated as compared to that of the traditional PCA implanted femur. Moreover, the C-T prosthesis possesses the following features, and might be an alternative for traditional stem-type prostheses for clinical application after future clinical trials: (1) prevention of the stress-shielding effect (decrease bone atrophy); (2) prevention of endosteal osteolysis; (3) preservation of bone stock for easy revision; (4) complied with conventional THR system.
Acknowledgements This study was supported by the grant from the National Science Council of the Republic of China (grant no. NSC-89-2320-B-033-001-M08). The computing facilities provided by the National Center for HighPerformance Computing are greatly appreciated.
References Bassett, L.W., Gold, R.H., Hedley, A.K., 1982. Radiology of failed surface replacement total hip arthroplasty. Am. J. Roentgenol. 139, 1083–1088. Buechel, F.F., Drucker, D., Jasty, M., Jiranek, W., Harris, W.H., 1994. Osteolysis around uncemented acetabular components of cobalt–chrome surface replacement hip arthroplasty. Clin. Orthop. Relat. R 298, 202–211. Callaghan, J.J., 1992. Total hip arthroplasty: clinical perspective. Clin. Orthop. Relat. R 276, 33–40. Chao, E.Y., Coventry, M.B., 1981. Fracture of femoral component after total hip replacement. An analysis of 58 cases. J. Bone Joint Surg. Am. 63A, 1078–1094.
Cody, D.D., Gross, G.J., Hou, J.H., Spencer, H.J., Glodstein, S., Fyhrie, D.P., 1999. Femoral strength is better predicted by finite element models than QCT and DXA. J. Biomech. 32, 1013–1020. Engh, C.A., Bobyn, J.D., 1988. The influence of stem size and extent of porous coating on femoral bone resorption after primary cementless hip arthroplasty. Clin. Orthop. Relat. R 231, 7–28. Harris, W.H., 1984. Will stress shielding limit the longevity of cemented femoral components of total hip replacement. Clin. Orthop. Relat. R 274, 120–123. Head, W.C., 1981. Wagner surface replacement arthoplasty of the hip. J. Bone Joint Surg. Am. 63A, 420–427. Herbeterts, P., Lansinger, O., Romanus, B., 1983. Surface replacement arthroplasty of the hip. Acta. Orthop. Scand. 54, 884–890. Huiskes, R., Strens, P., Heck, J.V., Sloof, T., 1985. Interface stress in the resurfaced hip. Acta Orthop. Scand. 56, 474–478. Kroger, H., Vanninen, E., Overmyer, M., Miettinen, H., Rushton, N., Suomalainen, O., 1997. Periprosthetic bone loss and regional bone turnover in uncemented total hip arthroplasty: a prospective study using high resolution single photon emission tomography and dualenergy X-ray absorptiometry. J. Bone Miner. Res. 12 (3), 487–492. Lewis, J.L., Askew, M.J., Wixson, R.L., Kramer, G.M., Tarr, R.R., 1984. The influence of prosthetic stem stiffness and of a calcar collar on stresses in the proximal end of the femur with a cemented femoral components. J. Bone Joint Surg. Am. 66A, 280–286. Mann, K.A., Bartel, D.L., 1995. Coulomb frictional interfaces in modeling cemented total hip replacement: a more realistic model. J. Biomech. 9, 1067–1078. Mann, K.A., Bartel, D.L., Ayers, D.C., 1997. Influence of stem geometry on mechanics of cemented femoral hip components with a proximal bond. J. Orthop. Res. 15, 700–706. Marchetti, M.E., Steinberg, G.G., Greene, J.M., Jenis, L.G., Baran, D.T., 1996. A prospective study of proximal femur bone mass following cemented and uncemented hip arthroplasty. J. Bone Miner. Res. 11 (7), 1033–1039. McCarthy, C.K., Steinberg, G.G., Agren, M., Leahey, D., Wyman, E., Baran, D.T., 1991. Quantifying bone loss from the proximal femur after total hip arthroplasty. J. Bone Joint Surg. [Br] 73B, 774–778. McNamara, B.P., Cristofolini, L., Toni, A., Taylor, D., 1997. Relationship between bone-prosthesis bonding and load transfer in total hip reconstruction. J. Biomech. 30, 621–630. Munting, E., Verhelpen, M., 1995. Fixation and effect on bone strain pattern of a stemless hip prosthesis. J. Biomech. 28, 949–961. Murray, D.W., Rushton, N., 1992. Mediators of bone resorption around implants. Clin. Orthop. Relat. R 281, 295–304. Shih, C.H., Chen, W.P., Tai, C.L., Kuo, R.F., Wu, C.C., Chen, C.H., 1997. New concepts-biomechanical studies of a newly design femoral prosthesis (cevico-trochanter prosthesis). Clin. Biomech. 12, 482–490. Shih, C.H., Wu, C.C., Lee, Z.L., Yang, W.E., 1994. Localized femoral osteolysis in cementless ceramic total hip arthroplasty. Orthop. Rev. 23, 325–328. Testi, D., Viceconti, M., Brauffaldi, F., Cappello, A., 1999. Risk of fracture in elderly patients: a new predictive index based on bone mineral density and finite element analysis. Comput. Meth. Prog. Bio. 60, 23–33. Trentani, C., Vaccarino, F., 1981. Complications in surface replacement arthroplasty of the hip: experience with the paltrinieritrentani prosthesis. Int. Orthop. (SICOT) 4, 247–252. Whiteside, L.A., 1989. The effect of stem fit on bone hypertrophy and pain relief in cementless total hip arthroplasty. Clin. Orthop. Relat. R 247, 138–147.