- Email: [email protected]
Determination of muscle loading at the hip joint for use in pre-clinical testing
ARTICLE IN PRESS
Journal of Biomechanics 38 (2005) 1155–1163
Determination of muscle loading at the hip joint for use in pre-clinical testing M.O. Hellera, G. Bergmannb, J.-P. Kassia, L. Claesc, N.P. Haasa, G.N. Dudaa,* a
Trauma and Reconstructive Surgery, Charite´, Campus Virchow-Clinic, Humboldt-University of Berlin, Augustenburger Platz 1, Berlin 13353, Germany b Biomechanics Laboratory, Free University of Berlin, Germany c Institute for Orthopaedic Research and Biomechanics, University of Ulm, Germany Accepted 12 May 2004
Abstract The stability of joint endoprostheses depends on the loading conditions to which the implant-bone complex is exposed. Due to a lack of appropriate muscle force data, less complex loading conditions tend to be considered in vitro. The goal of this study was to develop a load profile that better simulates the in vivo loading conditions of a ‘‘typical’’ total hip replacement patient and considers the interdependence of muscle and joint forces. The development of the load profile was based on a computer model of the lower extremities that has been validated against in vivo data. This model was simplified by grouping functionally similar hip muscles. Muscle and joint contact forces were computed for an average data set of up to four patients throughout walking and stair climbing. The calculated hip contact forces were compared to the average of the in vivo measured forces. The final derived load profile included the forces of up to four muscles at the instances of maximum in vivo hip joint loading during both walking and stair climbing. The hip contact forces differed by less than 10% from the peak in vivo value for a ‘‘typical’’ patient. The derived load profile presented here is the first that is based on validated musculoskeletal analyses and seems achievable in an in vitro test set-up. It should therefore form the basis for further standardisation of pre-clinical testing by providing a more realistic approximation of physiological loading conditions. r 2004 Elsevier Ltd. All rights reserved. Keywords: Hip joint; Loading; Joint contact; Muscle forces; Simulation; Pre-clinical testing
1. Introduction The long term success of total joint replacement (THR) depends on a number of factors such as the surgical technique (Zahiri et al., 1999; Cameron et al., 2001), the material (Black, 1989; Young et al., 1999) and the design of the endoprosthesis (Walker, 1989; Kubo et al., 2001). The introduction of a hip joint endoprosthesis into orthopaedic surgery is therefore preceded by tests concerning e.g. the biocompatibility of materials (Davidson et al., 1994; Rhalmi et al., 1999; Schmidt et al., 2001) and the fatigue strength of the components *Corresponding author. Tel.: +49-30-450-559079; fax: +49-30-450559969. E-mail address: [email protected] (G.N. Duda). 0021-9290/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2004.05.022
(Viceconti et al., 1995; Baleani et al., 1999). The ISO 7206 standard is well established for testing and improving the fatigue properties of metal femoral components (Dall et al., 1993; Wroblewski and Siney, 1993). A recent compilation of in vivo contact force amplitudes and activity frequencies (Bergmann, 2001) has shown, however, that the ISO 7206 standard might underestimate the severity of loads acting at the hip joint. It is widely accepted that primary stability is an essential parameter for the long-term performance of joint replacements (Mjoberg et al., 1986; Kim and Kim, 1993; Freeman and Plante-Bordeneuve, 1994; Kobayashi et al., 1997; Claes et al., 2000). The primary stability depends on the geometric design of the implant (Callaghan et al., 1992; Speirs et al., 2000), but also on
ARTICLE IN PRESS 1156
M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
the loading conditions to which the implant-bone complex is exposed (Bergmann et al., 1995). Whilst there has been considerable effort in the past to develop realistic testing protocols for both experimental conditions (Cristofolini et al., 1995; Britton et al., 2003; Cristofolini et al., 2003) as well as for numerical analyses (Stolk et al., 2001, 2002, 2003; Viceconti et al., 2001), to date there is no ISO or CEN standard to evaluate the primary stability during pre-clinical testing. In many in vitro analyses of implant stability, only the hip contact force has been simulated (Gebauer et al., 1989; McKellop et al., 1991; Berzins et al., 1993; Buhler . et al., 1997; Claes et al., 2000). These loading conditions represent a major simplification of the actual in vivo loading situation and result predominantly in bending loads on the shaft of the femoral component. Whilst this causes a critical failure loading of the implant shaft, it is not clear whether such a load is suitable for evaluation of fixation stability in the bone. Only a few studies of primary stability have included additional muscle forces (Burke et al., 1991; Callaghan et al., 1992). In these studies, however, muscle and joint contact forces were treated as mechanically independent parameters, and the interdependence between muscle activity and joint contact forces (Winter, 1990) was not considered. Musculoskeletal loading conditions at the hip are determined by the joint contact force and the forces of over 20 different muscles spanning the joint (Pedersen et al., 1997). Although there is strong evidence that muscles are major contributors to femoral loading (Rohlmann et al., 1982; Lu et al., 1997; Duda et al., 1998), the muscle forces acting in vivo are hardly accessible. Computer models have been used to predict the musculoskeletal loading conditions in the past (Seireg and Arvikar, 1975; Brand et al., 1986; Glitsch and Baumann, 1997; Pedersen et al., 1997), but the validation of the results against in vivo data remains a difficult issue (Brand et al., 1994). In the first direct comparison of musculoskeletal loading conditions computed from gait data (Heller et al., 2001) with simultaneously measured hip contact forces (Bergmann, 2001; Bergmann et al., 2001a), it has been shown that computer models can predict the in vivo data with an average error of 12% to 14% in four patients. Based on the model proposed by Heller et al. (2001), it therefore seems possible to determine physiological-like musculoskeletal loading conditions at the hip that consider the interdependence of muscle and joint contact forces. Whilst the study of Heller and co-workers (2001) employed a complex model of the hip musculature with over 30 different lines of muscle action, the physical space limitations and complexity of in vitro testing require a simplification and reduction of the number of muscle fibres.
The goal of the current study was therefore to develop a load profile for the proximal femur that resembles the in vivo loading conditions of a typical THR patient and at the same time considers the relationship between muscle and joint forces.
2. Material and methods In order to develop a simplified profile of the physiological-like loading conditions of the hip, this study has employed a previously developed computer model of the lower extremity that has been validated against in vivo data from walking and stair climbing (Heller et al., 2001). The model is only briefly summarised here: A computer model of the bones and muscles of the human lower extremities (CT-data, Visible Human, NLM, USA) was scaled to match the anatomies of four THR patients with telemeterized femoral components. Gait analysis data for walking and stair climbing were determined simultaneously with in vivo hip contact forces. The gait data and the individual musculoskeletal models were then used to calculate the intersegmental resultant joint forces at the ankle, hip and knee joint, as well as the muscle and joint contact forces throughout each gait cycle. The calculated hip joint contact forces were finally validated against the contact forces measured in vivo for both activities. 2.1. Modification of the musculoskeletal model In a first step, the complex representation of the hip muscles from the previously described musculoskeletal model was grossly simplified in order to reduce the number of muscle fibres (Fig. 1). The simplification of the musculoskeletal model was mainly focussed to the so-called ‘‘single joint muscles’’, i.e. muscles that solely span the hip joint. In this gross simplification, all fibres of the gluteus muscles with a similar function (gluteus maximus, medius and minimus) were grouped into one simplified ‘‘abductor muscle’’ with a single attachment site. A similar process was applied to the adductor muscles (adductus brevis, magnus and longus), yielding a resultant ‘‘adductor muscle’’. Muscles that span both the hip and knee joint are commonly referred to as bi-articular or ‘‘two joint muscles’’. These two joint muscles play an important functional role regarding the execution of movements (van Ingen Schenau et al., 1987; van Groeningen and Erkelens, 1994; van Ingen Schenau et al., 1995; Heise et al., 1996; van Bolhuis et al., 1998), and hence are important for deriving the physiological-like loading conditions. From these two joint muscles, the long head of the biceps femoris, the semitendinosous, semimembranosous, rectus femoris, gracilis and sartorius, all
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
1157
2.2. Typical patient In order to base the derived load profile on representative data for THR patients, data sets for a ‘‘typical patient’’ (Fig. 2) were computed from the gait data and in vivo hip contact forces of the investigated subjects (Bergmann, 2001; Bergmann et al., 2001a) using an averaging procedure based on Fourier analysis (Bergmann et al., 2001b). These data sets included four patients for walking and three patients for stair climbing. Therefore, the bodyweight (BW) of the ‘‘typical patient’’ differed for the two activities (walking: 836N; stair climbing: 847N). 2.3. Calculation of the musculoskeletal loads Based on the gait data for the ‘‘typical patient’’, the intersegmental resultant joint forces, muscle forces and joint contact forces were then computed for the ankle, knee and hip joint using both the complex model and the different configurations of the simplified models of
Walking
Fig. 1. Comparison of the complex (left) and most simplified (right) models of the hip musculature. Hip muscles with large attachment areas that had been represented by several muscle fibres in the complex model were pooled to obtain a simplified muscle model (see Table 1 for details). The total number of muscle fibres was further reduced by removing muscles that had only a small effect on hip joint loading.
[%BW]
400
0 Stancephase 50
0
100
[% Gaitcycle]
(a)
Stair Climbing 400
[%BW]
contribute to the hip contact force, but they do not have attachment sites at the proximal femur and therefore exert no direct forces on the bone. Whilst these musculoskeletal structures were included in full detail for the calculation of the musculoskeletal loading conditions throughout the gait cycle, their loading effect on the proximal femur can be modelled by ensuring an appropriate joint contact force acted on the femoral head, thereby not increasing the complexity of an in vitro test set-up. In an attempt to further reduce the number of muscles included in the model to the lowest feasible total, joint contact forces were calculated for a series of different configurations in which muscles with small forces (i.e. muscles from the group of the lateral rotators) were progressively removed from the model to a point beyond which unphysiological hip joint loading was calculated. The musculature for the remainder of the lower limb remained as described in the original model (Heller et al., 2001).
Swingphase
0 Stancephase 0 (b)
50 [% Gaitcycle]
Swingphase 100
Fig. 2. Resultant hip contact forces during (a) walking (4 individuals) and (b) stair climbing (3 individuals) shown as a percentage of bodyweight. The forces are shown for the individuals (thin solid lines) and a ‘‘typical patient’’ (thick broken lines).
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
1158
the hip musculature as described above. The whole gait cycles of both walking and stair climbing were analysed in order to establish the conformity of the computed musculoskeletal loading patterns with the hip contact forces of the ‘‘typical patient’’ derived from the in vivo measured data. The final derived load profile for the proximal femur was taken as the most simplified muscle configuration as described in the modification of the musculoskeletal model section that resulted in physiological-like joint loading throughout the gait cycle at the instances of maximum in vivo measured hip joint contact forces for both walking and stair climbing.
3. Results The time course and magnitudes of the hip contact forces calculated with the complex musculoskeletal models compared well with the in vivo forces for the ‘‘typical patient’’ for both activities (Fig. 3). At the Walking
[%BW]
400
0 Stancephase
Swingphase
0
50 [% Gaitcycle]
(a)
100
Stair Climbing
[%BW]
400
0 Stancephase 0 (b)
Swingphase 50
100
[% Gaitcycle]
Fig. 3. Comparison of the hip contact forces computed for the ‘‘typical patient’’ using either the complex (solid grey lines) or the most simplified model of the hip muscles (solid black lines) and the average of the measured in vivo hip contact forces (broken black lines, which also correspond to those in Fig. 2) during (a) walking and (b) stair climbing. The forces are given in percentage of bodyweight.
instance of maximum in vivo loading, the hip contact force calculated using the complex model of the hip muscles differed from the in vivo forces by approximately 1% during walking and 4% during stair climbing. 3.1. Simplification of the hip muscle model The gross grouping of the hip muscles to a single abductor and adductor muscle and the stepwise removal of most muscles with small forces resulted in joint contact forces that remained in good agreement with the in vivo hip loading data. Although the absolute force magnitudes in the ilio-tibial tract and the tensor fascia latae were found to be low (max. 49%BW), the forces in the remaining muscles and the hip joint contact forces were substantially increased when these structures were absent from the model: whilst the in vivo hip forces had peak values of 238%BW and 251%BW during walking and stair climbing respectively, the values of the computed hip contact forces were 427%BW during walking and 556%BW during stair climbing at the same instant. Furthermore, these computed forces reached peak values of 526%BW (walking) and 880%BW (stair climbing), thus leading to differences from the in vivo values in excess of 200%. The developed muscle model (Fig. 1, Table 1) therefore represented the simplest model of the hip muscles resulting in physiological-like hip joint loading throughout the entire gait cycles (Fig. 3). At the instance of maximum loading, the hip contact force calculated using this model differed from the in vivo forces by approximately 7% during walking (Fig. 3a) and 10% during stair climbing (Fig. 3b). During the entire walking cycle, the forces predicted by the developed musculoskeletal model were slightly higher than the in vivo forces (Fig. 4a). The musculoskeletal loading conditions at the hip during the walking cycle were characterised by a dominant inferior–superior joint contact force component (Fig. 4a). The model predicted two force peaks in this component, during early and late stance phase. The largest resultant hip contact force (262%BW) acted at the beginning of the gait cycle. The medio-lateral (max. 84%BW) and the anterior-posterior (max. 28%BW) shear force components were considerably smaller than the inferior–superior forces. At the instant of the measured peak resultant contact force (early stance phase), the largest muscle forces were computed for the abductor (104%BW), the vastus lateralis (95%BW) and the tensor fascia latae (19%BW). At the same instant, the largest contribution to the hip contact force from the group of the two joint hip muscles came from the semimenbranosous (60%BW). A higher activity of the muscles spanning both the hip and knee joint, i.e. the rectus femoris (max. 80%BW), the gracilis (max. 27%BW) and the sartorius (max. 14%BW) was
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163 Walking
Table 1 Summary of simplifications of the hip muscles resulting in the derived muscle model
gluteus medius gluteus minimus gluteus maximum
adductor magnus adductor longus adductor brevis
400
max. loading
Simplified 9 > = pooled > ; 9 > = pooled > ;
[%BW]
Complex
proximal-distal 200
abductor medio-lateral
anteriorposterior
0 Stancephase
adductor 0
unchanged
Ilio-tibial tract
tensor fascia latae
unchanged
tensor fascia latae
Swingphase 50 [% Gaitcycle]
(a)
ilio-tibial tract
9 > > > > > psoas major > > > = pectineus removed gmellus inferior & superior > > > > > obturator externus & internus > > > ; piriformis
1159
100
Stair Climbing
400
max. loading [%BW]
iliacus
proximal-distal 200
– medio-lateral
anteriorposterior
0
Stancephase
The table illustrates how the hip muscles included in the complex model were simplified to create the derived muscle model. In an iterative process, muscles with similar functions were pooled first. Subsequently muscles with smaller forces were removed from the model as long as physiological like hip joint loading was maintained. The modifications were restricted to muscles and structures with either attachments or wrapping points at the proximal femur.
observed, particularly at the end of the stance phase and during the swing phase of the cycle. The largest forces in the gastrocnemi, which attach at the distal femur, were also found towards the end of the stance phase (up to 31%BW). During stair climbing, the calculated hip joint contact forces overestimated the in vivo forces for the ‘‘typical patient’’ during the early stance phase, but underestimated the forces during late stance (Fig. 4b). Similar to the walking cycle, the peak joint contact force at the hip during stair climbing was calculated during the early stance phase (278%BW). Whilst the peak hip contact forces were generally larger than during walking, the most pronounced difference in amplitude was observed in the anterior–posterior joint contact force. This force component was twice as large during stair-climbing than during walking. Analysis of the muscle activity showed that the peak force in the abductors was increased by about 10% compared to walking (114%BW). Most importantly, the anterior–posterior component of the abductor force was increased more than 6-fold. The largest muscle forces during stair climbing were calculated for the vastus lateralis (137%BW) and the vastus
0
(b)
Swingphase 50 [% Gaitcycle]
100
Fig. 4. Comparison of the hip contact force components computed for the ‘‘typical patient’’ (solid lines) and the average of the measured in vivo hip contact forces (broken lines) during (a) walking (BW=836N) and (b) stair climbing (BW=847N). For the load profile for pre-clinical test conditions the instances of maximum in vivo resultant contact force were selected. The three force components -Fx, -Fy, -Fz are given in percentage of bodyweight with respect to a femur fixed coordinate system (Bergman et al., 1993). For a left femur, the positive x-axis of this coordinate system is oriented latero-medially, the positive y-axis posterior–anteriorly and the positive z-axis in an inferior-superior direction. The origin is located in the centre of the femoral head.
medialis (270%BW). Whilst a considerable peak force in the two joint semitendinosous muscle (115%BW) was found during early stance phase, activity of the muscles spanning both the hip and knee joint was observed mainly during the late stance and swing phase, albeit with small force magnitudes (up to 29%BW). 3.2. Derived musculoskeletal load profile Having calculated the musculoskeletal loading conditions throughout the complete cycles of both activities, the single instant of maximum in vivo hip contact force in each of the walking and stair climbing cycles was taken to define the derived load profile (Fig. 4). For these two selected instances, three (walking) and four (stair climbing) muscles of the derived model exerted forces at the proximal femur (Fig. 5). The instance of
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
1160
Fig. 5. Derived load profile developed for the pre-clinical testing of femoral shaft components. The forces (in percentage of bodyweight) and the coordinates (in millimetres) are given in a local coordinate system of the femur (Bergmann et al., 1993). The hip contact force acts at the origin of the coordinate system labelled as P0. The attachments or wrapping points of the muscles are labelled P1 to P3. The intersegmental resultant force is a force that describes the loading of the hip joint due to the weight of all body segments above the joint and their inertial forces. The hip contact force was computed as the sum of the intersegmental resultant joint forces and the sum of all single and two joint muscle forces of the hip. Therefore, when the hip contact force is directly applied at the femoral head, this force does not have to be applied separately. However, if an active simulation of the muscle forces is to be realised, in an experimental set-up for example, this intersegmental resultant force is required in order for the actual load applied to the femoral head to be determined.
maximum coincided (Fig. 4b) maximum
hip contact force during stair climbing with the peak anterior-posterior force and also represented the load profile for torsion acting on the shaft of the implant.
4. Discussion The aim of this study was to develop a load profile for the proximal femur that improves the simulation of in vivo loading conditions of a THR patient and considers the relationship between muscle and joint forces. It is accepted that the physiological loading conditions to which implant-bone complexes are exposed, have, amongst other factors such as the geometric design of the implant (Callaghan et al., 1992; Speirs et al., 2000), a major influence on the stability of joint reconstruction (Bergmann et al., 1995). Primary stability is considered to be a crucial factor for the biological integration of an endoprosthesis and the longevity of a
total joint replacement (Mjoberg et al., 1986; Kim and Kim, 1993; Freeman and Plante-Bordeneuve, 1994; Kobayashi et al., 1997). Therefore, it seems imperative to evaluate the primary stability achievable by each new implant design in pre-clinical testing. Whilst a number of authors have demonstrated the importance of muscle forces for simulating the loading conditions in the proximal femur, using both experimental set-ups with varying complexity (Cristofolini et al., 1995; Britton et al., 2003) as well as numerical analyses (Stolk et al., 2001, 2002), the loading often employed in such tests (Gebauer et al., 1989; McKellop et al., 1991; Buhler . et al., 1997), i.e. the hip contact force only, represents a major simplification of the actual in vivo situation. This is potentially due to a lack of standardised musculoskeletal loading conditions. Whilst loading the implant– bone complex by the hip contact force alone might create critical stresses in the shaft of the femoral component, it does not seem to create similar, critical loading conditions at the implant–bone interface (Stolk et al., 2001, 2002). In fact, the overestimated bending
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
moment (Duda et al., 1998) might cause locking of the prosthesis within the bone and hinder further implant movement. Therefore, the primary stability obtained under such oversimplified loading conditions may well overestimate the stability that will actually be achievable in vivo. The current study has focussed on the musculoskeletal loading conditions during two activities of daily life, i.e. walking and stair climbing. It has been shown that walking is the activity that occurs most frequently in THR patients (Morlock et al., 2001) and results in considerable hip contact forces in vivo (Bergmann, 2001; Bergmann et al., 2001a). In vivo measurements further demonstrate that stair climbing represents the activity during which the highest forces and the highest torsion on the shaft of an implant occur (Bergmann et al., 2001a). At the same time, stair climbing is one of the 5 most frequent activities in the daily life of a THR patient (Morlock et al., 2001). Whilst these two activities can represent only a fraction of the total activity of a typical patient after THR, the selection seems to be appropriate within the context of standardised pre-clinical testing of implant stability, especially considering the load magnitudes, the type of loading, the frequency of the activities and therefore the number of load cycles to which an implant will be exposed during its service life. Previous studies have suggested that there is an intraindividual as well as a considerable inter-individual variability in the musculoskeletal loading conditions at the hip between several repetitions of an activity in a single patient or between different patients (Bergmann et al., 2001a; Heller et al., 2001). In order that the derived load profile avoids only reflecting the loading characteristics of an individual patient, the musculoskeletal loads presented in this study were derived for a ‘‘typical patient’’. The data for the ‘‘typical patient’’ were obtained by averaging the gait data of up to four THR patients, for which in vivo measured hip contact forces were also available, using a dedicated averaging method (Bergmann et al., 2001a, b). In order to obtain the musculoskeletal loads for a preclinical test, two concurring requirements had to be met: on one hand, only a limited complexity in terms of the number of muscles included was acceptable in order to be able to realise those conditions within an in vitro setup. On the other hand, however, the loading had to result in a mechanical environment which closely resembled the physiological loading conditions at the hip. Therefore, a simplified model of the hip muscles was deduced from a complex model of the human lower extremities that had been validated against in vivo data (Heller et al., 2001). This musculoskeletal model was used to determine the muscle and joint contact forces for a ‘‘typical patient’’. A potential limitation of this approach is that the derived testing conditions are obtained from a numerical model which itself was based
1161
upon some assumptions (Heller et al., 2001). A numerical model using a different optimisation criterion, for example, could produce a variation in the distribution of muscle forces. However, the derived musculoskeletal load profile developed for pre-clinical testing using these techniques both achieved a physiologicallike loading environment, shown by an error of less than 10%, but also simplified the muscle activity to a maximum of four muscles. Considering the fact that an even more complex load configuration has already been described in the literature (Cristofolini et al., 1995), it seems likely that the currently proposed scenario would be also achievable in an in vitro test set-up. At the selected time instances of maximum hip contact force, one could further reduce the number of force vectors in the derived load profile by summing all muscle forces acting at the greater trochanter to a single force. This load configuration, however, was not capable of producing a physiological-like loading environment throughout the entire gait cycle. The characteristics of the musculoskeletal loading conditions of the typical patient obtained using a simplified model of the hip muscles were comparable to those found for the individual hip patients using complex muscle models (Heller et al., 2001). The load levels during stair climbing were generally higher than those during walking. The instances of maximum loading during walking and stair climbing were characterised by the strong activity of the abductors and the quadriceps muscles (vasti, rectus femoris). Since the hip was more flexed during stair climbing, the muscle activity in the abductor led to a doubling of the anterior–posterior joint contact force component at the hip. The higher torsional loads acting on the prosthesis shaft during this activity mean that it is considered a critical activity for the fixation of the implant (Bergmann et al., 1995; O’Connor et al., 1996; Shelley et al., 1996; Aamodt et al., 2001; Stolk et al., 2002). The musculoskeletal loads associated with this activity were therefore included in the load profile developed in this study. Activity of the tensor fascia latae and the ilio-tibial tract was also observed. Whilst the tensor fascia latae and the ilio-tibial tract are two joint structures without direct attachments at the proximal femur, they can exert forces as they are wrapped around the greater trochanter (Duda et al., 1998) and thereby act to minimise the bending moments in the femur (Pauwels, 1951; Rohlmann et al., 1980; Rohlmann et al., 1982). The step-wise simplification process of the hip muscles revealed that these structures had to be included to maintain physiological loading conditions throughout the full movement cycles. Although the physiological recruitment of muscles is rather complex, the activity of the muscles as described above is also in agreement with typical muscle activity profiles derived from EMG data (Sutherland, 2001).
ARTICLE IN PRESS 1162
M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
The load profile specified in this study may help to reproduce in vitro loading conditions which are closer to the peak load situations in vivo. This load profile could therefore be a basis for further standardisation of preclinical testing by providing a more realistic approximation of the musculoskeletal loading conditions.
Acknowledgements The authors would like to thank Dr. G. Deuretzbacher for supplying the gait data and Dr. W. R. Taylor for his thorough review of the manuscript. This study was partially supported by grants from the European Commission (SMT-CT96-207), the German Federal Institute for Drugs and Medical Devices (BfArM) and the German Research Society (DFG Du 298/4-1).
References Aamodt, A., Lund-Larsen, J., Eine, J., Andersen, E., Benum, P., Husby, O.S., 2001. Changes in proximal femoral strain after insertion of uncemented standard and customised femoral stems. An experimental study in human femora. Journal of Bone and Joint Surgery British Volume 83, 921–929. Baleani, M., Cristofolini, L., Viceconti, M., 1999. Endurance testing of hip prostheses: a comparison between the load fixed in ISO 7206 standard and the physiological loads. Clinical Biomechanics 14, 339–345. Bergmann, G. (Ed.), 2001. HIP98 - Loading of the Hip Joint. Compact Disc. Free University of Berlin, ISBN 3980784800. Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., Duda, G.N., 2001a. Hip contact forces and gait patterns from routine activities. Journal of Biomechanics 34, 859–871. Bergmann, G., Gandini, R., Ruder, H., 2001b. Averaging of strongly varying signals. Biomedizinische Technik 46, 168–171. Bergmann, G., Graichen, F., Rohlmann, A., 1995. Is staircase walking a risk for the fixation of hip implants? Journal of Biomechanics 28, 535–553. Berzins, A., Sumner, D.R., Andriacchi, T.P., Galante, J.O., 1993. Stem curvature and load angle influence the initial relative bone-implant motion of cementless femoral stems. Journal of Orthopaedic Research 11, 758–769. Black, J., 1989. Requirements for successful total knee replacement. Material considerations. The Orthopaedic Clinics of North America 20, 1–13. Brand, R.A., Pedersen, D.R., Davy, D.T., Kotzar, G.M., Heiple, K.G., Goldberg, V.M., 1994. Comparison of hip force calculations and measurements in the same patient. Journal of Arthroplasty 9, 45–51. Brand, R.A., Pedersen, D.R., Friederich, J.A., 1986. The sensitivity of muscle force predictions to changes in physiologic cross-sectional area. Journal of Biomechanics 19, 589–596. Britton, J.R., Walsh, L.A., Prendergast, P.J., 2003. Mechanical simulation of muscle loading on the proximal femur: analysis of cemented femoral component migration with and without muscle loading. Clinical Biomechanics 18, 637–646. Buhler, . D.W., Oxland, T.R., Nolte, L.P., 1997. Design and evaluation of a device for measuring three-dimensional micromotions of pressfit femoral stem prostheses. Medical Engineering and Physics 19, 187–199.
Burke, D.W., O’Connor, D.O., Zalenski, E.B., Jasty, M., Harris, W.H., 1991. Micromotion of cemented and uncemented femoral components. Journal of Bone and Joint Surgery British Volume 73, 33–37. Callaghan, J.J., Fulghum, C.S., Glisson, R.R., Stranne, S.K., 1992. The effect of femoral stem geometry on interface motion in uncemented porous-coated total hip prostheses. Journal of Bone and Joint Surgery American Volume 74, 839–848. Cameron, B., Galatz, L., Williams Jr., G.R., 2001. Factors affecting the outcome of total shoulder arthroplasty. The American Journal of Orthopedic 30, 613–623. Claes, L., Fiedler, S., Ohnmacht, M., Duda, G.N., 2000. Initial stability of fully and partially cemented femoral stems. Clinical Biomechanics 15, 750–755. Cristofolini, L., Teutonico, A.S., Monti, L., Cappello, A., Toni, A., 2003. Comparative in vitro study on the long term performance of cemented hip stems: validation of a protocol to discriminate between ‘‘good’’ and ‘‘bad’’ designs. Journal of Biomechanics 36, 1603–1615. Cristofolini, L., Viceconti, M., Toni, A., Giunti, A., 1995. Influence of thigh muscles on the axial strains in a proximal femur during early stance in gait. Journal of Biomechanics 28, 617–624. Dall, D.M., Learmonth, I.D., Solomon, M., Davenport, J.M., 1993. 811 charnley hips followed for 3–17 years. Acta orthopaedia Scandinavia 64, 252–256. Davidson, J.A., Mishra, A.K., Kovacs, P., Poggie, R.A., 1994. New surface-hardened, low-modulus, corrosion-resistant Ti-13Nb-13Zr alloy for total hip arthroplasty. Bio-medical Materials Engineering 4, 231–243. Duda, G.N., Heller, M., Albinger, J., Schulz, O., Schneider, E., Claes, L., 1998. Influence of muscle forces on femoral strain distribution. Journal of Biomechanics 31, 841–846. Freeman, M.A., Plante-Bordeneuve, P., 1994. Early migration and late aseptic failure of proximal femoral prostheses. Journal of Bone and Joint Surgery British Volume 76, 432–438. Gebauer, D., Refior, H.J., Haake, M., 1989. Micromotions in the primary fixation of cementless femoral stem prostheses. Archives of Orthopaedic and Trauma Surgery 108, 300–307. Glitsch, U., Baumann, W., 1997. The three-dimensional determination of internal loads in the lower extremity. Journal of Biomechanics 30, 1123–1131. Heise, G.D., Morgan, D.W., Hough, H., Craib, M., 1996. Relationships between running economy and temporal EMG characteristics of bi-articular leg muscles. Internations Journal of Sports Medicine 17, 128–133. Heller, M.O., Bergmann, G., Deuretzbacher, G., Durselen, . L., Pohl, M., Claes, L., Haas, N.P., Duda, G.N., 2001. Musculo-skeletal loading conditions at the hip during walking and stair climbing. Journal of Biomechanics 34, 883–893. Kim, Y.H., Kim, V.E., 1993. Early migration of uncemented porous coated anatomic femoral component related to aseptic loosening. Clinical Orthopaedics 146-155. Kobayashi, A., Donnelly, W.J., Scott, G., Freeman, M.A., 1997. Early radiological observations may predict the long-term survival of femoral hip prostheses. Journal of Bone and Joint Surgery British Volume 79, 583–589. Kubo, T., Inoue, S., Maeda, T., Arai, Y., Hirakawa, K., Wu, Y., Suehara, H., Ogura, T., Hirasawa, Y., 2001. Cementless lord total hip arthroplasty: cup loosening common after minimum 10-year follow-up of 103 hips. Acta orthopaedics Scandinowa 72, 585–590. Lu, T.W., Taylor, S.J., O’Connor, J.J., Walker, P.S., 1997. Influence of muscle activity on the forces in the femur: an in vivo study. Journal of Biomechanics 30, 1101–1106. McKellop, H. A., Ebramzadeh, E., Park, S. H., Wiss, D., Brien, W., Grover, J. S. (Eds.), 1991. Development and clinical performance of a reversible titanium alloy femoral interlocking nail.
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163 Mjoberg, B., Selvik, G., Hansson, L.I., Rosenqvist, R., Onnerfalt, R., 1986. Mechanical loosening of total hip prostheses. A radiographic and roentgen stereophotogrammetric study. Journal of Bone and Joint Surgery British Volume 68, 770–774. Morlock, M., Schneider, E., Bluhm, A., Vollmer, M., Bergmann, G., Muller, . V., Honl, M., 2001. Duration and frequency of every day activities in total hip patients. Journal of Biomechanics 34, 873–881. O’Connor, D.O., Burke, D.W., Jasty, M., Sedlacek, R.C., Harris, W.H., 1996. In vitro measurement of strain in the bone cement surrounding the femoral component of total hip replacements during simulated gait and stair-climbing. Journal of Orthopaedic Research 14, 769–777. . Pauwels, F., 1951. Uber die Bedeutung der Bauprinzipien des St.utz- und Bewegungsapparates f.ur die Beanspruchung des R.ohrenknochens. Acta anatonica 12, 207–227. Pedersen, D.R., Brand, R.A., Davy, D.T., 1997. Pelvic muscle and acetabular contact forces during gait. Journal of Biomechanics 30, 959–965. Rhalmi, S., Odin, M., Assad, M., Tabrizian, M., Rivard, C.H., Yahia, L.H., 1999. Hard, soft tissue and in vitro cell response to porous nickel-titanium: a biocompatibility evaluation. Bio-medical Materials Engineering 9, 151–162. . Rohlmann, A., Bergmann, G., Kolbel, R., 1980. Die Beanspruchung des Femur. I. Intaktes Femur ohne Berucksichtigung . des Tractus iliotibialis. Zeitschrift fur . Orthop.adie und ihre Grenzgebiete 118, 897–904. . . Rohlmann, A., Mossner, U., Bergmann, G., Kolbel, R., 1982. Finiteelement-analysis and experimental investigation of stresses in a femur. Journal of Biomedical Engineering 4, 241–246. Schmidt, C., Ignatius, A.A., Claes, L.E., 2001. Proliferation and differentiation parameters of human osteoblasts on titanium and steel surfaces. Journal of Biomedical Matererials Research 54, 209–215. Seireg, A., Arvikar, R.J., 1975. The prediction of muscular load sharing and joint forces in the lower extremities during walking. Journal of Biomechanics 8, 89–102. Shelley, F.J., Anderson, D.D., Kolar, M.J., Miller, M.C., Rubash, H.E., 1996. Physical modelling of hip joint forces in stair climbing. Proceeding of the Institute of Mechanical Engineereing. Part H: Journal of Engineering in Medicine 210, 65–68. Speirs, A.D., Slomczykowski, M.A., Orr, T.E., Siebenrock, K., Nolte, L.P., 2000. Three-dimensional measurement of cemented femoral stem stability: an in vitro cadaver study. Clinical Biomechanics 15, 248–255. Stolk, J., Maher, S.A., Verdonschot, N., Prendergast, P.J., Huiskes, R., 2003. Can finite element models detect clinically inferior cemented hip implants? Clinical Orthopaedics 409, 138–150.
1163
Stolk, J., Verdonschot, N., Huiskes, R., 2001. Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip reconstruction. Journal of Biomechanics 34, 917–926. Stolk, J., Verdonschot, N., Huiskes, R., 2002. Stair climbing is more detrimental to the cement in hip replacement than walking. Clinical Orthopaedics 405, 294–305. Sutherland, D.H., 2001. The evolution of clinical gait analysis part 1: kinesiological EMG. Gait & Posture 14, 61–70. van Bolhuis, B.M., Gielen, C.C., van Ingen Schenau, G.J., 1998. Activation patterns of mono- and bi-articular arm muscles as a function of force and movement direction of the wrist in humans. Journal of Physiology 508, 313–324. van Groeningen, C.J., Erkelens, C.J., 1994. Task-dependent differences between mono- and bi-articular heads of the triceps brachii muscle. Experimental Brain Research 100, 345–352. van Ingen Schenau, G.J., Bobbert, M.F., Rozendal, R.H., 1987. The unique action of bi-articular muscles in complex movements. Journal of Anatomy 155, 1–5. van Ingen Schenau, G.J., Dorssers, W.M., Welter, T.G., Beelen, A., de Groot, G., Jacobs, R., 1995. The control of mono-articular muscles in multijoint leg extensions in man. Journal of Physiology 484 (Pt 1), 247–254. Viceconti, M., Cristofolini, L., Baleani, M., Toni, A., 2001. Pre-clinical validation of a new partially cemented femoral prosthesis by synergetic use of numerical and experimental methods. Journal of Biomechanics 34, 723–731. Viceconti, M., Toni, A., Giunti, A., 1995. Effects of some technological aspects on the fatigue strength of a cementless hip stem. Journal of Biomedical Materials Research 29, 875–881. Walker, P.S., 1989. Requirements for successful total knee replacements. Design considerations. The Orthopaedic Clinics of North America 20, 15–29. Winter, D. A., 1990. Biomechanics and Motor Control of Human Movement. In: Book Biomechanics and Motor Control of Human Movement. Wiley, New York. p. 78. Wroblewski, B.M., Siney, P.D., 1993. Charnley low-friction arthroplasty of the hip. Long-term results. Clinical Orthopaedics 191–201. Young, T.H., Cheng, C.K., Lee, Y.M., Chen, L.Y., Huang, C.H., 1999. Analysis of ultrahigh molecular weight polyethylene failure in artificial knee joints: thermal effect on longterm performance. Journal of Biomedical Materials Research 48, 159–164. Zahiri, C.A., Schmalzried, T.P., Ebramzadeh, E., Szuszczewicz, E.S., Salib, D., Kim, C., Amstutz, H.C., 1999. Lessons learned from loosening of the McKee-Farrar metal-on-metal total hip replacement. Journal of Arthroplasty 14, 326–332.
Journal of Biomechanics 38 (2005) 1155–1163
Determination of muscle loading at the hip joint for use in pre-clinical testing M.O. Hellera, G. Bergmannb, J.-P. Kassia, L. Claesc, N.P. Haasa, G.N. Dudaa,* a
Trauma and Reconstructive Surgery, Charite´, Campus Virchow-Clinic, Humboldt-University of Berlin, Augustenburger Platz 1, Berlin 13353, Germany b Biomechanics Laboratory, Free University of Berlin, Germany c Institute for Orthopaedic Research and Biomechanics, University of Ulm, Germany Accepted 12 May 2004
Abstract The stability of joint endoprostheses depends on the loading conditions to which the implant-bone complex is exposed. Due to a lack of appropriate muscle force data, less complex loading conditions tend to be considered in vitro. The goal of this study was to develop a load profile that better simulates the in vivo loading conditions of a ‘‘typical’’ total hip replacement patient and considers the interdependence of muscle and joint forces. The development of the load profile was based on a computer model of the lower extremities that has been validated against in vivo data. This model was simplified by grouping functionally similar hip muscles. Muscle and joint contact forces were computed for an average data set of up to four patients throughout walking and stair climbing. The calculated hip contact forces were compared to the average of the in vivo measured forces. The final derived load profile included the forces of up to four muscles at the instances of maximum in vivo hip joint loading during both walking and stair climbing. The hip contact forces differed by less than 10% from the peak in vivo value for a ‘‘typical’’ patient. The derived load profile presented here is the first that is based on validated musculoskeletal analyses and seems achievable in an in vitro test set-up. It should therefore form the basis for further standardisation of pre-clinical testing by providing a more realistic approximation of physiological loading conditions. r 2004 Elsevier Ltd. All rights reserved. Keywords: Hip joint; Loading; Joint contact; Muscle forces; Simulation; Pre-clinical testing
1. Introduction The long term success of total joint replacement (THR) depends on a number of factors such as the surgical technique (Zahiri et al., 1999; Cameron et al., 2001), the material (Black, 1989; Young et al., 1999) and the design of the endoprosthesis (Walker, 1989; Kubo et al., 2001). The introduction of a hip joint endoprosthesis into orthopaedic surgery is therefore preceded by tests concerning e.g. the biocompatibility of materials (Davidson et al., 1994; Rhalmi et al., 1999; Schmidt et al., 2001) and the fatigue strength of the components *Corresponding author. Tel.: +49-30-450-559079; fax: +49-30-450559969. E-mail address: [email protected] (G.N. Duda). 0021-9290/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2004.05.022
(Viceconti et al., 1995; Baleani et al., 1999). The ISO 7206 standard is well established for testing and improving the fatigue properties of metal femoral components (Dall et al., 1993; Wroblewski and Siney, 1993). A recent compilation of in vivo contact force amplitudes and activity frequencies (Bergmann, 2001) has shown, however, that the ISO 7206 standard might underestimate the severity of loads acting at the hip joint. It is widely accepted that primary stability is an essential parameter for the long-term performance of joint replacements (Mjoberg et al., 1986; Kim and Kim, 1993; Freeman and Plante-Bordeneuve, 1994; Kobayashi et al., 1997; Claes et al., 2000). The primary stability depends on the geometric design of the implant (Callaghan et al., 1992; Speirs et al., 2000), but also on
ARTICLE IN PRESS 1156
M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
the loading conditions to which the implant-bone complex is exposed (Bergmann et al., 1995). Whilst there has been considerable effort in the past to develop realistic testing protocols for both experimental conditions (Cristofolini et al., 1995; Britton et al., 2003; Cristofolini et al., 2003) as well as for numerical analyses (Stolk et al., 2001, 2002, 2003; Viceconti et al., 2001), to date there is no ISO or CEN standard to evaluate the primary stability during pre-clinical testing. In many in vitro analyses of implant stability, only the hip contact force has been simulated (Gebauer et al., 1989; McKellop et al., 1991; Berzins et al., 1993; Buhler . et al., 1997; Claes et al., 2000). These loading conditions represent a major simplification of the actual in vivo loading situation and result predominantly in bending loads on the shaft of the femoral component. Whilst this causes a critical failure loading of the implant shaft, it is not clear whether such a load is suitable for evaluation of fixation stability in the bone. Only a few studies of primary stability have included additional muscle forces (Burke et al., 1991; Callaghan et al., 1992). In these studies, however, muscle and joint contact forces were treated as mechanically independent parameters, and the interdependence between muscle activity and joint contact forces (Winter, 1990) was not considered. Musculoskeletal loading conditions at the hip are determined by the joint contact force and the forces of over 20 different muscles spanning the joint (Pedersen et al., 1997). Although there is strong evidence that muscles are major contributors to femoral loading (Rohlmann et al., 1982; Lu et al., 1997; Duda et al., 1998), the muscle forces acting in vivo are hardly accessible. Computer models have been used to predict the musculoskeletal loading conditions in the past (Seireg and Arvikar, 1975; Brand et al., 1986; Glitsch and Baumann, 1997; Pedersen et al., 1997), but the validation of the results against in vivo data remains a difficult issue (Brand et al., 1994). In the first direct comparison of musculoskeletal loading conditions computed from gait data (Heller et al., 2001) with simultaneously measured hip contact forces (Bergmann, 2001; Bergmann et al., 2001a), it has been shown that computer models can predict the in vivo data with an average error of 12% to 14% in four patients. Based on the model proposed by Heller et al. (2001), it therefore seems possible to determine physiological-like musculoskeletal loading conditions at the hip that consider the interdependence of muscle and joint contact forces. Whilst the study of Heller and co-workers (2001) employed a complex model of the hip musculature with over 30 different lines of muscle action, the physical space limitations and complexity of in vitro testing require a simplification and reduction of the number of muscle fibres.
The goal of the current study was therefore to develop a load profile for the proximal femur that resembles the in vivo loading conditions of a typical THR patient and at the same time considers the relationship between muscle and joint forces.
2. Material and methods In order to develop a simplified profile of the physiological-like loading conditions of the hip, this study has employed a previously developed computer model of the lower extremity that has been validated against in vivo data from walking and stair climbing (Heller et al., 2001). The model is only briefly summarised here: A computer model of the bones and muscles of the human lower extremities (CT-data, Visible Human, NLM, USA) was scaled to match the anatomies of four THR patients with telemeterized femoral components. Gait analysis data for walking and stair climbing were determined simultaneously with in vivo hip contact forces. The gait data and the individual musculoskeletal models were then used to calculate the intersegmental resultant joint forces at the ankle, hip and knee joint, as well as the muscle and joint contact forces throughout each gait cycle. The calculated hip joint contact forces were finally validated against the contact forces measured in vivo for both activities. 2.1. Modification of the musculoskeletal model In a first step, the complex representation of the hip muscles from the previously described musculoskeletal model was grossly simplified in order to reduce the number of muscle fibres (Fig. 1). The simplification of the musculoskeletal model was mainly focussed to the so-called ‘‘single joint muscles’’, i.e. muscles that solely span the hip joint. In this gross simplification, all fibres of the gluteus muscles with a similar function (gluteus maximus, medius and minimus) were grouped into one simplified ‘‘abductor muscle’’ with a single attachment site. A similar process was applied to the adductor muscles (adductus brevis, magnus and longus), yielding a resultant ‘‘adductor muscle’’. Muscles that span both the hip and knee joint are commonly referred to as bi-articular or ‘‘two joint muscles’’. These two joint muscles play an important functional role regarding the execution of movements (van Ingen Schenau et al., 1987; van Groeningen and Erkelens, 1994; van Ingen Schenau et al., 1995; Heise et al., 1996; van Bolhuis et al., 1998), and hence are important for deriving the physiological-like loading conditions. From these two joint muscles, the long head of the biceps femoris, the semitendinosous, semimembranosous, rectus femoris, gracilis and sartorius, all
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
1157
2.2. Typical patient In order to base the derived load profile on representative data for THR patients, data sets for a ‘‘typical patient’’ (Fig. 2) were computed from the gait data and in vivo hip contact forces of the investigated subjects (Bergmann, 2001; Bergmann et al., 2001a) using an averaging procedure based on Fourier analysis (Bergmann et al., 2001b). These data sets included four patients for walking and three patients for stair climbing. Therefore, the bodyweight (BW) of the ‘‘typical patient’’ differed for the two activities (walking: 836N; stair climbing: 847N). 2.3. Calculation of the musculoskeletal loads Based on the gait data for the ‘‘typical patient’’, the intersegmental resultant joint forces, muscle forces and joint contact forces were then computed for the ankle, knee and hip joint using both the complex model and the different configurations of the simplified models of
Walking
Fig. 1. Comparison of the complex (left) and most simplified (right) models of the hip musculature. Hip muscles with large attachment areas that had been represented by several muscle fibres in the complex model were pooled to obtain a simplified muscle model (see Table 1 for details). The total number of muscle fibres was further reduced by removing muscles that had only a small effect on hip joint loading.
[%BW]
400
0 Stancephase 50
0
100
[% Gaitcycle]
(a)
Stair Climbing 400
[%BW]
contribute to the hip contact force, but they do not have attachment sites at the proximal femur and therefore exert no direct forces on the bone. Whilst these musculoskeletal structures were included in full detail for the calculation of the musculoskeletal loading conditions throughout the gait cycle, their loading effect on the proximal femur can be modelled by ensuring an appropriate joint contact force acted on the femoral head, thereby not increasing the complexity of an in vitro test set-up. In an attempt to further reduce the number of muscles included in the model to the lowest feasible total, joint contact forces were calculated for a series of different configurations in which muscles with small forces (i.e. muscles from the group of the lateral rotators) were progressively removed from the model to a point beyond which unphysiological hip joint loading was calculated. The musculature for the remainder of the lower limb remained as described in the original model (Heller et al., 2001).
Swingphase
0 Stancephase 0 (b)
50 [% Gaitcycle]
Swingphase 100
Fig. 2. Resultant hip contact forces during (a) walking (4 individuals) and (b) stair climbing (3 individuals) shown as a percentage of bodyweight. The forces are shown for the individuals (thin solid lines) and a ‘‘typical patient’’ (thick broken lines).
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
1158
the hip musculature as described above. The whole gait cycles of both walking and stair climbing were analysed in order to establish the conformity of the computed musculoskeletal loading patterns with the hip contact forces of the ‘‘typical patient’’ derived from the in vivo measured data. The final derived load profile for the proximal femur was taken as the most simplified muscle configuration as described in the modification of the musculoskeletal model section that resulted in physiological-like joint loading throughout the gait cycle at the instances of maximum in vivo measured hip joint contact forces for both walking and stair climbing.
3. Results The time course and magnitudes of the hip contact forces calculated with the complex musculoskeletal models compared well with the in vivo forces for the ‘‘typical patient’’ for both activities (Fig. 3). At the Walking
[%BW]
400
0 Stancephase
Swingphase
0
50 [% Gaitcycle]
(a)
100
Stair Climbing
[%BW]
400
0 Stancephase 0 (b)
Swingphase 50
100
[% Gaitcycle]
Fig. 3. Comparison of the hip contact forces computed for the ‘‘typical patient’’ using either the complex (solid grey lines) or the most simplified model of the hip muscles (solid black lines) and the average of the measured in vivo hip contact forces (broken black lines, which also correspond to those in Fig. 2) during (a) walking and (b) stair climbing. The forces are given in percentage of bodyweight.
instance of maximum in vivo loading, the hip contact force calculated using the complex model of the hip muscles differed from the in vivo forces by approximately 1% during walking and 4% during stair climbing. 3.1. Simplification of the hip muscle model The gross grouping of the hip muscles to a single abductor and adductor muscle and the stepwise removal of most muscles with small forces resulted in joint contact forces that remained in good agreement with the in vivo hip loading data. Although the absolute force magnitudes in the ilio-tibial tract and the tensor fascia latae were found to be low (max. 49%BW), the forces in the remaining muscles and the hip joint contact forces were substantially increased when these structures were absent from the model: whilst the in vivo hip forces had peak values of 238%BW and 251%BW during walking and stair climbing respectively, the values of the computed hip contact forces were 427%BW during walking and 556%BW during stair climbing at the same instant. Furthermore, these computed forces reached peak values of 526%BW (walking) and 880%BW (stair climbing), thus leading to differences from the in vivo values in excess of 200%. The developed muscle model (Fig. 1, Table 1) therefore represented the simplest model of the hip muscles resulting in physiological-like hip joint loading throughout the entire gait cycles (Fig. 3). At the instance of maximum loading, the hip contact force calculated using this model differed from the in vivo forces by approximately 7% during walking (Fig. 3a) and 10% during stair climbing (Fig. 3b). During the entire walking cycle, the forces predicted by the developed musculoskeletal model were slightly higher than the in vivo forces (Fig. 4a). The musculoskeletal loading conditions at the hip during the walking cycle were characterised by a dominant inferior–superior joint contact force component (Fig. 4a). The model predicted two force peaks in this component, during early and late stance phase. The largest resultant hip contact force (262%BW) acted at the beginning of the gait cycle. The medio-lateral (max. 84%BW) and the anterior-posterior (max. 28%BW) shear force components were considerably smaller than the inferior–superior forces. At the instant of the measured peak resultant contact force (early stance phase), the largest muscle forces were computed for the abductor (104%BW), the vastus lateralis (95%BW) and the tensor fascia latae (19%BW). At the same instant, the largest contribution to the hip contact force from the group of the two joint hip muscles came from the semimenbranosous (60%BW). A higher activity of the muscles spanning both the hip and knee joint, i.e. the rectus femoris (max. 80%BW), the gracilis (max. 27%BW) and the sartorius (max. 14%BW) was
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163 Walking
Table 1 Summary of simplifications of the hip muscles resulting in the derived muscle model
gluteus medius gluteus minimus gluteus maximum
adductor magnus adductor longus adductor brevis
400
max. loading
Simplified 9 > = pooled > ; 9 > = pooled > ;
[%BW]
Complex
proximal-distal 200
abductor medio-lateral
anteriorposterior
0 Stancephase
adductor 0
unchanged
Ilio-tibial tract
tensor fascia latae
unchanged
tensor fascia latae
Swingphase 50 [% Gaitcycle]
(a)
ilio-tibial tract
9 > > > > > psoas major > > > = pectineus removed gmellus inferior & superior > > > > > obturator externus & internus > > > ; piriformis
1159
100
Stair Climbing
400
max. loading [%BW]
iliacus
proximal-distal 200
– medio-lateral
anteriorposterior
0
Stancephase
The table illustrates how the hip muscles included in the complex model were simplified to create the derived muscle model. In an iterative process, muscles with similar functions were pooled first. Subsequently muscles with smaller forces were removed from the model as long as physiological like hip joint loading was maintained. The modifications were restricted to muscles and structures with either attachments or wrapping points at the proximal femur.
observed, particularly at the end of the stance phase and during the swing phase of the cycle. The largest forces in the gastrocnemi, which attach at the distal femur, were also found towards the end of the stance phase (up to 31%BW). During stair climbing, the calculated hip joint contact forces overestimated the in vivo forces for the ‘‘typical patient’’ during the early stance phase, but underestimated the forces during late stance (Fig. 4b). Similar to the walking cycle, the peak joint contact force at the hip during stair climbing was calculated during the early stance phase (278%BW). Whilst the peak hip contact forces were generally larger than during walking, the most pronounced difference in amplitude was observed in the anterior–posterior joint contact force. This force component was twice as large during stair-climbing than during walking. Analysis of the muscle activity showed that the peak force in the abductors was increased by about 10% compared to walking (114%BW). Most importantly, the anterior–posterior component of the abductor force was increased more than 6-fold. The largest muscle forces during stair climbing were calculated for the vastus lateralis (137%BW) and the vastus
0
(b)
Swingphase 50 [% Gaitcycle]
100
Fig. 4. Comparison of the hip contact force components computed for the ‘‘typical patient’’ (solid lines) and the average of the measured in vivo hip contact forces (broken lines) during (a) walking (BW=836N) and (b) stair climbing (BW=847N). For the load profile for pre-clinical test conditions the instances of maximum in vivo resultant contact force were selected. The three force components -Fx, -Fy, -Fz are given in percentage of bodyweight with respect to a femur fixed coordinate system (Bergman et al., 1993). For a left femur, the positive x-axis of this coordinate system is oriented latero-medially, the positive y-axis posterior–anteriorly and the positive z-axis in an inferior-superior direction. The origin is located in the centre of the femoral head.
medialis (270%BW). Whilst a considerable peak force in the two joint semitendinosous muscle (115%BW) was found during early stance phase, activity of the muscles spanning both the hip and knee joint was observed mainly during the late stance and swing phase, albeit with small force magnitudes (up to 29%BW). 3.2. Derived musculoskeletal load profile Having calculated the musculoskeletal loading conditions throughout the complete cycles of both activities, the single instant of maximum in vivo hip contact force in each of the walking and stair climbing cycles was taken to define the derived load profile (Fig. 4). For these two selected instances, three (walking) and four (stair climbing) muscles of the derived model exerted forces at the proximal femur (Fig. 5). The instance of
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
1160
Fig. 5. Derived load profile developed for the pre-clinical testing of femoral shaft components. The forces (in percentage of bodyweight) and the coordinates (in millimetres) are given in a local coordinate system of the femur (Bergmann et al., 1993). The hip contact force acts at the origin of the coordinate system labelled as P0. The attachments or wrapping points of the muscles are labelled P1 to P3. The intersegmental resultant force is a force that describes the loading of the hip joint due to the weight of all body segments above the joint and their inertial forces. The hip contact force was computed as the sum of the intersegmental resultant joint forces and the sum of all single and two joint muscle forces of the hip. Therefore, when the hip contact force is directly applied at the femoral head, this force does not have to be applied separately. However, if an active simulation of the muscle forces is to be realised, in an experimental set-up for example, this intersegmental resultant force is required in order for the actual load applied to the femoral head to be determined.
maximum coincided (Fig. 4b) maximum
hip contact force during stair climbing with the peak anterior-posterior force and also represented the load profile for torsion acting on the shaft of the implant.
4. Discussion The aim of this study was to develop a load profile for the proximal femur that improves the simulation of in vivo loading conditions of a THR patient and considers the relationship between muscle and joint forces. It is accepted that the physiological loading conditions to which implant-bone complexes are exposed, have, amongst other factors such as the geometric design of the implant (Callaghan et al., 1992; Speirs et al., 2000), a major influence on the stability of joint reconstruction (Bergmann et al., 1995). Primary stability is considered to be a crucial factor for the biological integration of an endoprosthesis and the longevity of a
total joint replacement (Mjoberg et al., 1986; Kim and Kim, 1993; Freeman and Plante-Bordeneuve, 1994; Kobayashi et al., 1997). Therefore, it seems imperative to evaluate the primary stability achievable by each new implant design in pre-clinical testing. Whilst a number of authors have demonstrated the importance of muscle forces for simulating the loading conditions in the proximal femur, using both experimental set-ups with varying complexity (Cristofolini et al., 1995; Britton et al., 2003) as well as numerical analyses (Stolk et al., 2001, 2002), the loading often employed in such tests (Gebauer et al., 1989; McKellop et al., 1991; Buhler . et al., 1997), i.e. the hip contact force only, represents a major simplification of the actual in vivo situation. This is potentially due to a lack of standardised musculoskeletal loading conditions. Whilst loading the implant– bone complex by the hip contact force alone might create critical stresses in the shaft of the femoral component, it does not seem to create similar, critical loading conditions at the implant–bone interface (Stolk et al., 2001, 2002). In fact, the overestimated bending
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
moment (Duda et al., 1998) might cause locking of the prosthesis within the bone and hinder further implant movement. Therefore, the primary stability obtained under such oversimplified loading conditions may well overestimate the stability that will actually be achievable in vivo. The current study has focussed on the musculoskeletal loading conditions during two activities of daily life, i.e. walking and stair climbing. It has been shown that walking is the activity that occurs most frequently in THR patients (Morlock et al., 2001) and results in considerable hip contact forces in vivo (Bergmann, 2001; Bergmann et al., 2001a). In vivo measurements further demonstrate that stair climbing represents the activity during which the highest forces and the highest torsion on the shaft of an implant occur (Bergmann et al., 2001a). At the same time, stair climbing is one of the 5 most frequent activities in the daily life of a THR patient (Morlock et al., 2001). Whilst these two activities can represent only a fraction of the total activity of a typical patient after THR, the selection seems to be appropriate within the context of standardised pre-clinical testing of implant stability, especially considering the load magnitudes, the type of loading, the frequency of the activities and therefore the number of load cycles to which an implant will be exposed during its service life. Previous studies have suggested that there is an intraindividual as well as a considerable inter-individual variability in the musculoskeletal loading conditions at the hip between several repetitions of an activity in a single patient or between different patients (Bergmann et al., 2001a; Heller et al., 2001). In order that the derived load profile avoids only reflecting the loading characteristics of an individual patient, the musculoskeletal loads presented in this study were derived for a ‘‘typical patient’’. The data for the ‘‘typical patient’’ were obtained by averaging the gait data of up to four THR patients, for which in vivo measured hip contact forces were also available, using a dedicated averaging method (Bergmann et al., 2001a, b). In order to obtain the musculoskeletal loads for a preclinical test, two concurring requirements had to be met: on one hand, only a limited complexity in terms of the number of muscles included was acceptable in order to be able to realise those conditions within an in vitro setup. On the other hand, however, the loading had to result in a mechanical environment which closely resembled the physiological loading conditions at the hip. Therefore, a simplified model of the hip muscles was deduced from a complex model of the human lower extremities that had been validated against in vivo data (Heller et al., 2001). This musculoskeletal model was used to determine the muscle and joint contact forces for a ‘‘typical patient’’. A potential limitation of this approach is that the derived testing conditions are obtained from a numerical model which itself was based
1161
upon some assumptions (Heller et al., 2001). A numerical model using a different optimisation criterion, for example, could produce a variation in the distribution of muscle forces. However, the derived musculoskeletal load profile developed for pre-clinical testing using these techniques both achieved a physiologicallike loading environment, shown by an error of less than 10%, but also simplified the muscle activity to a maximum of four muscles. Considering the fact that an even more complex load configuration has already been described in the literature (Cristofolini et al., 1995), it seems likely that the currently proposed scenario would be also achievable in an in vitro test set-up. At the selected time instances of maximum hip contact force, one could further reduce the number of force vectors in the derived load profile by summing all muscle forces acting at the greater trochanter to a single force. This load configuration, however, was not capable of producing a physiological-like loading environment throughout the entire gait cycle. The characteristics of the musculoskeletal loading conditions of the typical patient obtained using a simplified model of the hip muscles were comparable to those found for the individual hip patients using complex muscle models (Heller et al., 2001). The load levels during stair climbing were generally higher than those during walking. The instances of maximum loading during walking and stair climbing were characterised by the strong activity of the abductors and the quadriceps muscles (vasti, rectus femoris). Since the hip was more flexed during stair climbing, the muscle activity in the abductor led to a doubling of the anterior–posterior joint contact force component at the hip. The higher torsional loads acting on the prosthesis shaft during this activity mean that it is considered a critical activity for the fixation of the implant (Bergmann et al., 1995; O’Connor et al., 1996; Shelley et al., 1996; Aamodt et al., 2001; Stolk et al., 2002). The musculoskeletal loads associated with this activity were therefore included in the load profile developed in this study. Activity of the tensor fascia latae and the ilio-tibial tract was also observed. Whilst the tensor fascia latae and the ilio-tibial tract are two joint structures without direct attachments at the proximal femur, they can exert forces as they are wrapped around the greater trochanter (Duda et al., 1998) and thereby act to minimise the bending moments in the femur (Pauwels, 1951; Rohlmann et al., 1980; Rohlmann et al., 1982). The step-wise simplification process of the hip muscles revealed that these structures had to be included to maintain physiological loading conditions throughout the full movement cycles. Although the physiological recruitment of muscles is rather complex, the activity of the muscles as described above is also in agreement with typical muscle activity profiles derived from EMG data (Sutherland, 2001).
ARTICLE IN PRESS 1162
M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163
The load profile specified in this study may help to reproduce in vitro loading conditions which are closer to the peak load situations in vivo. This load profile could therefore be a basis for further standardisation of preclinical testing by providing a more realistic approximation of the musculoskeletal loading conditions.
Acknowledgements The authors would like to thank Dr. G. Deuretzbacher for supplying the gait data and Dr. W. R. Taylor for his thorough review of the manuscript. This study was partially supported by grants from the European Commission (SMT-CT96-207), the German Federal Institute for Drugs and Medical Devices (BfArM) and the German Research Society (DFG Du 298/4-1).
References Aamodt, A., Lund-Larsen, J., Eine, J., Andersen, E., Benum, P., Husby, O.S., 2001. Changes in proximal femoral strain after insertion of uncemented standard and customised femoral stems. An experimental study in human femora. Journal of Bone and Joint Surgery British Volume 83, 921–929. Baleani, M., Cristofolini, L., Viceconti, M., 1999. Endurance testing of hip prostheses: a comparison between the load fixed in ISO 7206 standard and the physiological loads. Clinical Biomechanics 14, 339–345. Bergmann, G. (Ed.), 2001. HIP98 - Loading of the Hip Joint. Compact Disc. Free University of Berlin, ISBN 3980784800. Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., Duda, G.N., 2001a. Hip contact forces and gait patterns from routine activities. Journal of Biomechanics 34, 859–871. Bergmann, G., Gandini, R., Ruder, H., 2001b. Averaging of strongly varying signals. Biomedizinische Technik 46, 168–171. Bergmann, G., Graichen, F., Rohlmann, A., 1995. Is staircase walking a risk for the fixation of hip implants? Journal of Biomechanics 28, 535–553. Berzins, A., Sumner, D.R., Andriacchi, T.P., Galante, J.O., 1993. Stem curvature and load angle influence the initial relative bone-implant motion of cementless femoral stems. Journal of Orthopaedic Research 11, 758–769. Black, J., 1989. Requirements for successful total knee replacement. Material considerations. The Orthopaedic Clinics of North America 20, 1–13. Brand, R.A., Pedersen, D.R., Davy, D.T., Kotzar, G.M., Heiple, K.G., Goldberg, V.M., 1994. Comparison of hip force calculations and measurements in the same patient. Journal of Arthroplasty 9, 45–51. Brand, R.A., Pedersen, D.R., Friederich, J.A., 1986. The sensitivity of muscle force predictions to changes in physiologic cross-sectional area. Journal of Biomechanics 19, 589–596. Britton, J.R., Walsh, L.A., Prendergast, P.J., 2003. Mechanical simulation of muscle loading on the proximal femur: analysis of cemented femoral component migration with and without muscle loading. Clinical Biomechanics 18, 637–646. Buhler, . D.W., Oxland, T.R., Nolte, L.P., 1997. Design and evaluation of a device for measuring three-dimensional micromotions of pressfit femoral stem prostheses. Medical Engineering and Physics 19, 187–199.
Burke, D.W., O’Connor, D.O., Zalenski, E.B., Jasty, M., Harris, W.H., 1991. Micromotion of cemented and uncemented femoral components. Journal of Bone and Joint Surgery British Volume 73, 33–37. Callaghan, J.J., Fulghum, C.S., Glisson, R.R., Stranne, S.K., 1992. The effect of femoral stem geometry on interface motion in uncemented porous-coated total hip prostheses. Journal of Bone and Joint Surgery American Volume 74, 839–848. Cameron, B., Galatz, L., Williams Jr., G.R., 2001. Factors affecting the outcome of total shoulder arthroplasty. The American Journal of Orthopedic 30, 613–623. Claes, L., Fiedler, S., Ohnmacht, M., Duda, G.N., 2000. Initial stability of fully and partially cemented femoral stems. Clinical Biomechanics 15, 750–755. Cristofolini, L., Teutonico, A.S., Monti, L., Cappello, A., Toni, A., 2003. Comparative in vitro study on the long term performance of cemented hip stems: validation of a protocol to discriminate between ‘‘good’’ and ‘‘bad’’ designs. Journal of Biomechanics 36, 1603–1615. Cristofolini, L., Viceconti, M., Toni, A., Giunti, A., 1995. Influence of thigh muscles on the axial strains in a proximal femur during early stance in gait. Journal of Biomechanics 28, 617–624. Dall, D.M., Learmonth, I.D., Solomon, M., Davenport, J.M., 1993. 811 charnley hips followed for 3–17 years. Acta orthopaedia Scandinavia 64, 252–256. Davidson, J.A., Mishra, A.K., Kovacs, P., Poggie, R.A., 1994. New surface-hardened, low-modulus, corrosion-resistant Ti-13Nb-13Zr alloy for total hip arthroplasty. Bio-medical Materials Engineering 4, 231–243. Duda, G.N., Heller, M., Albinger, J., Schulz, O., Schneider, E., Claes, L., 1998. Influence of muscle forces on femoral strain distribution. Journal of Biomechanics 31, 841–846. Freeman, M.A., Plante-Bordeneuve, P., 1994. Early migration and late aseptic failure of proximal femoral prostheses. Journal of Bone and Joint Surgery British Volume 76, 432–438. Gebauer, D., Refior, H.J., Haake, M., 1989. Micromotions in the primary fixation of cementless femoral stem prostheses. Archives of Orthopaedic and Trauma Surgery 108, 300–307. Glitsch, U., Baumann, W., 1997. The three-dimensional determination of internal loads in the lower extremity. Journal of Biomechanics 30, 1123–1131. Heise, G.D., Morgan, D.W., Hough, H., Craib, M., 1996. Relationships between running economy and temporal EMG characteristics of bi-articular leg muscles. Internations Journal of Sports Medicine 17, 128–133. Heller, M.O., Bergmann, G., Deuretzbacher, G., Durselen, . L., Pohl, M., Claes, L., Haas, N.P., Duda, G.N., 2001. Musculo-skeletal loading conditions at the hip during walking and stair climbing. Journal of Biomechanics 34, 883–893. Kim, Y.H., Kim, V.E., 1993. Early migration of uncemented porous coated anatomic femoral component related to aseptic loosening. Clinical Orthopaedics 146-155. Kobayashi, A., Donnelly, W.J., Scott, G., Freeman, M.A., 1997. Early radiological observations may predict the long-term survival of femoral hip prostheses. Journal of Bone and Joint Surgery British Volume 79, 583–589. Kubo, T., Inoue, S., Maeda, T., Arai, Y., Hirakawa, K., Wu, Y., Suehara, H., Ogura, T., Hirasawa, Y., 2001. Cementless lord total hip arthroplasty: cup loosening common after minimum 10-year follow-up of 103 hips. Acta orthopaedics Scandinowa 72, 585–590. Lu, T.W., Taylor, S.J., O’Connor, J.J., Walker, P.S., 1997. Influence of muscle activity on the forces in the femur: an in vivo study. Journal of Biomechanics 30, 1101–1106. McKellop, H. A., Ebramzadeh, E., Park, S. H., Wiss, D., Brien, W., Grover, J. S. (Eds.), 1991. Development and clinical performance of a reversible titanium alloy femoral interlocking nail.
ARTICLE IN PRESS M.O. Heller et al. / Journal of Biomechanics 38 (2005) 1155–1163 Mjoberg, B., Selvik, G., Hansson, L.I., Rosenqvist, R., Onnerfalt, R., 1986. Mechanical loosening of total hip prostheses. A radiographic and roentgen stereophotogrammetric study. Journal of Bone and Joint Surgery British Volume 68, 770–774. Morlock, M., Schneider, E., Bluhm, A., Vollmer, M., Bergmann, G., Muller, . V., Honl, M., 2001. Duration and frequency of every day activities in total hip patients. Journal of Biomechanics 34, 873–881. O’Connor, D.O., Burke, D.W., Jasty, M., Sedlacek, R.C., Harris, W.H., 1996. In vitro measurement of strain in the bone cement surrounding the femoral component of total hip replacements during simulated gait and stair-climbing. Journal of Orthopaedic Research 14, 769–777. . Pauwels, F., 1951. Uber die Bedeutung der Bauprinzipien des St.utz- und Bewegungsapparates f.ur die Beanspruchung des R.ohrenknochens. Acta anatonica 12, 207–227. Pedersen, D.R., Brand, R.A., Davy, D.T., 1997. Pelvic muscle and acetabular contact forces during gait. Journal of Biomechanics 30, 959–965. Rhalmi, S., Odin, M., Assad, M., Tabrizian, M., Rivard, C.H., Yahia, L.H., 1999. Hard, soft tissue and in vitro cell response to porous nickel-titanium: a biocompatibility evaluation. Bio-medical Materials Engineering 9, 151–162. . Rohlmann, A., Bergmann, G., Kolbel, R., 1980. Die Beanspruchung des Femur. I. Intaktes Femur ohne Berucksichtigung . des Tractus iliotibialis. Zeitschrift fur . Orthop.adie und ihre Grenzgebiete 118, 897–904. . . Rohlmann, A., Mossner, U., Bergmann, G., Kolbel, R., 1982. Finiteelement-analysis and experimental investigation of stresses in a femur. Journal of Biomedical Engineering 4, 241–246. Schmidt, C., Ignatius, A.A., Claes, L.E., 2001. Proliferation and differentiation parameters of human osteoblasts on titanium and steel surfaces. Journal of Biomedical Matererials Research 54, 209–215. Seireg, A., Arvikar, R.J., 1975. The prediction of muscular load sharing and joint forces in the lower extremities during walking. Journal of Biomechanics 8, 89–102. Shelley, F.J., Anderson, D.D., Kolar, M.J., Miller, M.C., Rubash, H.E., 1996. Physical modelling of hip joint forces in stair climbing. Proceeding of the Institute of Mechanical Engineereing. Part H: Journal of Engineering in Medicine 210, 65–68. Speirs, A.D., Slomczykowski, M.A., Orr, T.E., Siebenrock, K., Nolte, L.P., 2000. Three-dimensional measurement of cemented femoral stem stability: an in vitro cadaver study. Clinical Biomechanics 15, 248–255. Stolk, J., Maher, S.A., Verdonschot, N., Prendergast, P.J., Huiskes, R., 2003. Can finite element models detect clinically inferior cemented hip implants? Clinical Orthopaedics 409, 138–150.
1163
Stolk, J., Verdonschot, N., Huiskes, R., 2001. Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip reconstruction. Journal of Biomechanics 34, 917–926. Stolk, J., Verdonschot, N., Huiskes, R., 2002. Stair climbing is more detrimental to the cement in hip replacement than walking. Clinical Orthopaedics 405, 294–305. Sutherland, D.H., 2001. The evolution of clinical gait analysis part 1: kinesiological EMG. Gait & Posture 14, 61–70. van Bolhuis, B.M., Gielen, C.C., van Ingen Schenau, G.J., 1998. Activation patterns of mono- and bi-articular arm muscles as a function of force and movement direction of the wrist in humans. Journal of Physiology 508, 313–324. van Groeningen, C.J., Erkelens, C.J., 1994. Task-dependent differences between mono- and bi-articular heads of the triceps brachii muscle. Experimental Brain Research 100, 345–352. van Ingen Schenau, G.J., Bobbert, M.F., Rozendal, R.H., 1987. The unique action of bi-articular muscles in complex movements. Journal of Anatomy 155, 1–5. van Ingen Schenau, G.J., Dorssers, W.M., Welter, T.G., Beelen, A., de Groot, G., Jacobs, R., 1995. The control of mono-articular muscles in multijoint leg extensions in man. Journal of Physiology 484 (Pt 1), 247–254. Viceconti, M., Cristofolini, L., Baleani, M., Toni, A., 2001. Pre-clinical validation of a new partially cemented femoral prosthesis by synergetic use of numerical and experimental methods. Journal of Biomechanics 34, 723–731. Viceconti, M., Toni, A., Giunti, A., 1995. Effects of some technological aspects on the fatigue strength of a cementless hip stem. Journal of Biomedical Materials Research 29, 875–881. Walker, P.S., 1989. Requirements for successful total knee replacements. Design considerations. The Orthopaedic Clinics of North America 20, 15–29. Winter, D. A., 1990. Biomechanics and Motor Control of Human Movement. In: Book Biomechanics and Motor Control of Human Movement. Wiley, New York. p. 78. Wroblewski, B.M., Siney, P.D., 1993. Charnley low-friction arthroplasty of the hip. Long-term results. Clinical Orthopaedics 191–201. Young, T.H., Cheng, C.K., Lee, Y.M., Chen, L.Y., Huang, C.H., 1999. Analysis of ultrahigh molecular weight polyethylene failure in artificial knee joints: thermal effect on longterm performance. Journal of Biomedical Materials Research 48, 159–164. Zahiri, C.A., Schmalzried, T.P., Ebramzadeh, E., Szuszczewicz, E.S., Salib, D., Kim, C., Amstutz, H.C., 1999. Lessons learned from loosening of the McKee-Farrar metal-on-metal total hip replacement. Journal of Arthroplasty 14, 326–332.