Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip...

Journal of Biomechanics 34 (2001) 917–926

Hip-joint and abductor-muscle forces adequately represent in vivoloading of a cemented total hip reconstruction

J. Stolk, N. Verdonschot*, R. Huiskes

Orthopaedic Research Laboratory, Institute of Orthopaedics, University of Nijmegen P.O. Box 9101,

6500 HB Nijmegen, Netherlands

Accepted 21 November 2000

Abstract

Using finite element analyses, we investigated which muscle groups acting around the hip-joint most prominently affected the loaddistributions in cemented total hip reconstructions with a bonded and debonded femoral stem. The purpose was to determine whichmuscle groups should be included in pre-clinical tests, predicting bone adaptation and mechanical failure of cementedreconstructions, ensuring an adequate representation of in vivo loading of the reconstruction. Loads were applied as occurring

during heel-strike, mid-stance and push-off phases of gait. The stress/strain distibutions within the reconstruction, produced by thehip-joint contact force, were compared to ones produced after sequentially including the abductors, the iliotibial tract and theadductors and vastii. Inclusion of the abductors had the most pronounced effect. They neutralized lateral bending of the

reconstruction at heel-strike and increased medial bending at mid-stance and push-off. Bone strains and stem stresses were changedaccordingly. Peak tensile cement stresses were reduced during all gait phases by amounts up to 50% around a bonded stem and 11%around a debonded one. Additional inclusion of the iliotibial tract, the adductors and the vastii produced relatively small effects

during all gait phases. Their most prominent effect was a slight reduction of bone strains at the level of the stem tip during heel-strike. These results suggest that a loading configuration including the hip-joint contact force and the abductor forces canadequately reproduce in vivo loading of cemented total hip reconstructions in pre-clinical tests. # 2001 Elsevier Science Ltd. All

rights reserved.

Keywords: Muscle forces; Pre-clinical testing; Total hip replacement; Implant loosening; Femoral loading; Finite element analysis

1. Introduction

Pre-clinical testing of hip replacement implants mayhelp to reduce the failure probability of cemented totalhip arthroplasty (THA) reconstructions. A good exam-ple is the standardized test ISO 7206 (Semlitsch andPanic, 1983), which is an experimental test for fatiguefailure of the femoral stem. Since this test wasintroduced the incidence of stem breakage has beenreduced considerably and is now only a minor reasonfor revision (Malchau and Herberts, 1998). In the caseof cemented hip replacement implants other mechanicalfailure processes leading to aseptic loosening, such ascement creep, cement cracking, failure of interfaces,particle formation and adverse bone remodeling havebecome more apparent (Huiskes, 1993a). Experimental

and numerical non-standardized tests are performed totest implants against these failure scenario’s (Huiskes,1993b). These pre-clinical tests have to test thereconstruction as a whole, rather than the isolatedimplant. It is important that the loading configuration insuch tests accurately represents in vivo loading of thereconstruction, to ensure that failure occurs at locationssimilar to in situ failure sites. This enables comparisonof test results to clinical findings and allows the testresults to be used to optimize implant designs. However,a complex loading configuration including all musclesthat act around the hip joint is infeasible, particularly inexperimental tests. A choice should be made to includeonly those muscles that have a major influence on theload-transfer in a THA reconstruction, thereby affectingmechanical failure of the reconstruction.There seems to be no general agreement about which

muscle groups should be included in experimental andnumerical tests to ensure that they produce realisticinformation about the failure probability of THA

*Corresponding author. Tel.: +31-24-36-17080; fax: +31-24-35-

40555.

E-mail address: [email protected] (N. Verdonschot).

0021-9290/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.

PII: S 0 0 2 1 - 9 2 9 0 ( 0 0 ) 0 0 2 2 5 - 6

reconstructions. Cristofolini (1997) and Colgan et al.(1994) listed a large number of numerical and experi-mental studies in which hip-joint reconstructions werepre-clinically tested, also listing the muscle forces thatwere included in the test setups. In most cases a hip-jointforce was applied to the prosthetic head without anyadditional muscle forces or only a set of abductor forces.In some cases an additional iliotibial tract was included.Only a few numerical studies included a more extensiveset of muscles in the model. It is largely unknown towhat extent muscle forces affected the load distributionswithin the reconstructions tested and, hence, thepredictions of these tests. This information is essentialwhen developing a loading configuration that ensuresrealistic predictions from pre-clinical tests.The purpose of this study was to determine which

muscle forces should be included in pre-clinical testspredicting bone adaptation and mechanical failure ofcemented THA reconstructions. To address this pur-pose, we investigated which muscle groups actingaround the hip joint were the principle ones affectingthe load distributions in a cemented THA reconstruc-tion during walking. Finite element analyses (FEA) wereperformed to analyze the load-transfer around a bondedand a debonded Exeter stem (Howmedica, Staines, UK)at three phases of the gait cycle. The deflection of thereconstruction and the stress/strain distributions in thestem, cement and bone, as produced by the hip jointcontact force alone, were compared to the onesproduced after additionally including the abductors,the iliotibial tract and the adductors and vastii.

2. Materials and methods

A three-dimensional FEA model (Fig. 1) was created,representing a THA reconstruction with a cementedexeter stem (Howmedica, Staines, UK). The geometry ofthe femur was obtained from 27 CT-scans of theproximal part of a right human femur (Huiskes et al.,1992). The cement mantle had a minimal thickness of2mm. Distal to the stem tip a void was present, ascreated by a centralizer. The resulting model consistedof 2226 8-node isoparametric brick-elements and 3438nodes. Both a bonded and a debonded stem weresimulated. In the latter case, a total of 272 gap-elementswere positioned at the stem–cement interface to simu-late the non-linear behavior of a debonded interface(MARC Analysis Research Corporation, Palo Alto,CA). A friction coefficient of 0.25 was used, which isrealistic for polished stainless steel surfaces (Mann et al.,1991; Verdonschot and Huiskes, 1996).All materials were assumed to be isotropic and linear

elastic. A Young’s modulus of 200GPa and a Poisson’sratio of 0.28 were assigned to the stem. The correspond-ing values for the cement were 2.2GPa and 0.3. An

individual Young’s modulus was assigned to each boneelement, based on the average apparent density r forthat specific element as determined from the CT-scans.The apparent density was related to the Young’smodulus by (Carter and Hayes, 1977)

E ¼ Cr3; ð1Þwith C=3790MPa/(g/cm3). The Poisson’s ratio forbone was 0.35.To define the muscle forces acting around the

proximal femur we used a data-set by Brand et al.(1982, 1986, 1994), modified by Duda (1996) to allowwrapping of muscles around bony structures. Brandet al. (1986) split all glutei muscles in three parts, eachexerting its own force on the femur. In our model, thethree parts of the gluteus minimus were combined intoone single force, as was also done with parts 2 and 3 ofthe gluteus medius (Duda, 1996). The small externalrotators (gemelli, obturatori and piriformis) wereassumed to be dissected during surgery and were notincluded in our model. The iliotibial tract was repre-sented by the tensor fasciae latae and part 3 of thegluteus maximus, both running from the pelvis to thetibia, wrapped around the greater trochanter. To modelthe wrapping of these muscles, Duda (1996) split theminto parts a and b running from the greater trochanter tothe pelvis and the tibia, respectively. For both parts hedefined a pseudo-insertion at the greater trochanter. Aresulting total of 19 muscle forces were included in ourFEA model (Fig. 2, Table 1).

Fig. 1. In this study we used a CT-based, 3-D FEA model representing

the femoral part of a THA reconstruction with a cemented exeter stem

(Howmedica, Staines, UK). Only the anterior half of the model is

shown.

J. Stolk et al. / Journal of Biomechanics 34 (2001) 917–926918

Insertions and origins of the muscles were obtainedfrom Duda (1996). The coordinates of the femoralinsertions were transferred to our model by an osteo-metric scaling process using coordinates of bony land-marks (Sommer et al., 1982). This resulted in the points

of load application of all muscles on the femur. Todetermine the force directions, the muscles wereassumed to act along a straight line from their femoralinsertion to the insertion on either the pelvis, the tibia orthe patella. Knowing the hip and knee-joint anglesthroughout the gait cycle (Brand et al., 1986), theposition of the patellar, tibial and pelvic insertions couldbe determined relative to the femoral ones. The resultinglines of action were transferred to our FEA model byrepeating the osteometric scaling process. The forcemagnitudes throughout the gait cycle were taken fromBrand et al. (1986). They were scaled to a bodyweightof 735N.A hip-joint force was applied to the center of the

prosthetic head. Its magnitude and direction wereobtained from Duda et al. (1997). After scaling to abodyweight of 735N, the hip-joint force was rotated tothe coordinate system used in our model (Fig. 3).Additional boundary conditions implied that the modelwas fixed at its distal end.The load-transfer in the THA reconstruction was

investigated at 10, 30 and 45% of the gait cycle (Fig. 3).At 10% of the gait cycle, just after heel strike, thetorsional component of the joint force has a high valueand muscle activity is high as well (Brand et al., 1986).The mid-stance phase is reached at 30% of the gaitcycle. At 45% of the gait cycle, just before toe off, theaxial component of the hip-joint force reaches itsmaximum.To study the influence of various muscle groups

on the load-transfer in the reconstruction with a

Fig. 2. A total of 19 muscle forces were included in the FEA model in

loading configuration 4. The arrows indicate the attachment locations

of these 19 muscles on the anterior (left) and posterior (right) side of the

model. Their direction corresponds to the muscle lines of action at 10%

gait. The numbers correspond to the muscle names listed in Table 1.

Table 1

A total of 19 muscle forces were included in this studya

Sl. no. Muscle name Muscle force included in Force magnitude at phase

LC 1 LC 2 LC 3 LC 4 10% 30% 45 %

1 Gluteus maximus part 1 } x x x 0.183 0.198 0.191

2 Gluteus maximus part 2 } x x x 0.145 0.161 0.133

3 Gluteus maximus part 3a } } x x 0.136 0.0 0.0

4 Gluteus maximus part 3b } } x x 0.136 0.0 0.0

5 Gluteus medius part 1 } x x x 0.140 0.125 0.147

6 Gluteus medius part 2–3 } x x x 0.232 0.307 0.309

7 Gluteus minimus part 1–3 } x x x 0.089 0.173 0.197

8 Psoas major } } } x 0.0 0.038 0.112

9 Iliacus } } } x 0.0 0.068 0.143

10 Tensor fasciae latae a } } x x 0.062 0.074 0.130

11 Tensor fasciae latae b } } x x 0.062 0.074 0.130

12 Quadratus femoris } } } x 0.065 0.0 0.0

13 Pectineus } } } x 0.0 0.0 0.003

14 Adductor brevis } } } x 0.0 0.0 0.0

15 Adductor longus } } } x 0.0 0.0 0.006

16 Adductor magnus cran. } } } x 0.126 0.0 0.000

17 Vastus medialis } } } x 0.396 0.0 0.013

18 Vastus intermedius } } } x 0.0 0.0 0.092

19 Vastus lateralis } } } x 0.0 0.0 0.332

aThis table lists which muscles were included in the four loading configurations. In addition, the magnitudes of these forces (in BW) are listed, as

occurring during the three phases of the gait cycle analyzed.

J. Stolk et al. / Journal of Biomechanics 34 (2001) 917–926 919

bonded and a debonded stem, each phase of the gaitcycle was analyzed with four loading configurations(Table 1):LC 1: hip-joint contact force, no muscles active;LC 2: hip-joint contact force, abductor muscles active;LC 3: hip-joint contact force, abductors and muscles

inserting into iliotibial tract active;LC 4: hip-joint contact force, all muscles active.To comprehend the action of the various muscle

groups and their effect on the loading of the reconstruc-tion as a whole, we compared the deflection of thereconstruction between the four loading cases. Subse-quently, the effect of muscle forces on the individualparts of the reconstruction was analyzed by comparingstem and cement stresses between the four loading cases,as well as bone strain and strain energy density (SED)distributions.

3. Results

Inclusion of abductor-muscle forces in addition to thehip-joint force had a major impact on the deflection ofthe reconstruction, whereas additional inclusion of theiliotibial tract, the vastii and the adductors hadrelatively small effects. Displacements of the head centrewere similar for cases with a bonded and a debondedstem (differences within 0.05mm) and, hence, the effectsof muscle forces on the deflection of the reconstructionwere similar for a bonded and a debonded stem. Theabductors considerably affected the medio-lateral de-flection of the reconstruction, but hardly affected theposterior deflection (Table 2). At 10% gait, they largelyneutralized the lateral component of the displacement ofthe head centre, as produced by the hip joint force alone.At 30 and 45% gait, the lateral component wasconverted to a medial one after the activation ofabductors. Additional inclusion of the iliotibial tract,the vastii and the adductors caused the prosthetic headto move slightly antero-medially during all phases of thegait cycle.For both a bonded and a debonded stem the effects of

muscle forces on stem stresses were small proximal tothe lesser trochanter, whereas further distally stemstresses were mainly affected by inclusion of theabductor forces in the loading configuration (Fig. 4).The abductors reduced lateral bending of the stem at10% gait and increased medial bending of the stem at 30and 45% gait. As a result, the peak tensile stem stressesas produced by the hip-joint force alone (LC1), weredecreased at 10% gait, and increased at 30 and 45% gaitby the inclusion of the abductor forces (Table 2).Activation of additional muscle groups had relatively

Fig. 3. The three components of the hip-joint contact force through-

out the gait cycle in the coordinate system as given in Fig. 1 (adapted

from Duda et al., 1997). Load transfer in the THA reconstruction was

analyzed at 10, 30 and 45% of the gait cycle.

Table 2

Effect of inclusion of muscle forces on several variables

Gait

phase

Loading

conf.

Displacement of head center (mm) of debonded stema Peak axial stem stress (MPa)b Peak tensile cement stress (MPa)c

(%) DX DY DZ Bonded stem Debonded stem Bonded stem Debonded stem

10 LC 1 1.78 �1.40 0.35 42.1 47.4 4.61 3.89

LC 2 0.52 �1.05 0.038 30.6 33.3 2.35 3.83

LC 3 0.21 �0.93 �0.039 30.8 34.1 1.95 4.30

LC 4 �0.022 �0.97 �0.095 30.9 34.2 1.81 4.26

30 LC 1 0.81 �0.86 0.059 40.8 51.6 3.52 6.02

LC 2 �0.81 �0.75 �0.35 53.1 59.0 2.05 5.65

LC 3 �0.90 �0.66 �0.37 52.2 59.5 2.04 5.55

LC 4 �0.94 �0.55 �0.37 51.1 58.3 2.04 5.51

45 LC 1 0.010 �1.66 �0.20 67.4 85.3 3.61 9.85

LC 2 �1.70 �1.70 �0.63 87.6 99.5 3.49 8.77

LC 3 �1.85 �1.56 �0.66 86.5 98.7 3.45 8.81

LC 4 �1.84 �1.26 �0.66 82.6 95.8 3.30 8.78

aDisplacements of the centre of the prosthetic head are given for the case of a debonded stem. The coordinate system is as indicated in Fig. 1.

Displacements in the case of a bonded stem were within 0.05mm difference and are not listed.bPeak values of the axial stem stress (stress in z-direction as defined in Fig. 1) are listed both bonded and debonded stems.cPeak values of the maximal principal stress in the cement mantle are listed as occurring around both bonded and debonded stems.


small effects, the most pronounced being a reduction ofpeak stem stresses by roughly 5% at 45% of the gaitcycle. Although stress levels in a debonded stem weregenerally higher than those in a bonded stem, the effectsof muscle forces on stem stresses were very similar inboth cases.The cement stresses around a bonded and a debonded

stem were most prominently affected by the activationof abductors, whereas additional activation of othermuscle groups had relatively small effects. Inclusion ofthe abductors had the most pronounced effect oncement stresses around a bonded stem, where theyreduced tensile stresses proximally–laterally and aroundthe tip, but increased tensile stresses at mid-stem levels(Fig. 5). This effect was noted during all phases of thegait cycle. Peak tensile cement stresses around a bondedstem were decreased by as much as 50% at 10% gait(Table 2). In the case of a debonded stem, the tensilestress levels in the cement mantle were higher andappeared to be mainly determined by the hip-joint force.The only effect of the abductors was a slight reduction oftensile stresses around the stem tip during all phasesof the gait cycle. The peak tensile cement stresses arounda debonded stem changed less than 11% when abductormuscle forces were included in the model (Table 2).Peak tensile cement stresses were maximal at 45% of thegait cycle.Both in the case of a bonded and a debonded stem,

bone surface strains and SED values were mostprominently affected by the inclusion of the abductorforces in the loading configuration, in addition to the

hip-joint contact force. At intertrochanteric levels theyamplified medial bending of the femur during all phasesof the gait cycle, leading to increased SED values(Fig. 6) and bending strains (Fig. 7) in these regions. Inthe diaphysis they neutralized bone loading at 10% gait,but amplified bone loading at 30 and 45% gait.Additional inclusion of the iliotibial tract, the adductorsand the vastii had the most pronounced influence at10% of the gait cycle around the level of the stem tip.Locally, they reduced bending strains by approximately30% of the strain reduction caused by the inclusion ofabductors in the loading configuration (Fig. 7). Theirinfluence on bone loading was limited at 30 and 45% ofthe gait cycle.

4. Discussion

In this study, we investigated which muscle groupsacting around the hip-joint were the main ones affectingthe load distribution in a cemented THA reconstructionwith a femoral prosthetic stem, as produced duringwalking. The aim was to determine which muscle forcesshould be included in FEA and experimental pre-clinicaltests, predicting mechanical failure and peri-prostheticbone adaptation in cemented THA reconstructions.The stress/strain distributions within the reconstruc-

tion were primarily affected by inclusion of the abductorforces in the loading configuration, in addition the hip-joint contact force. By pulling the reconstructionmedially, they had a major impact on the stress/strain

Fig. 4. Stem stresses in axial direction (z-direction as defined in Fig. 1) along the lateral and the medial side of a debonded stem, as occurring at 45%

gait. Stresses are plotted for all four loading configurations. Negative stresses are compressive, positive ones tensile. Stem stresses were mainly

affected by inclusion of the abductors in addition to the hip-joint contact force (LC 1 vs. LC 2). Inclusion of additional muscles hardly affected the

stem stresses in any of the gait phases analyzed.


distributions in all parts of the reconstruction duringall phases of the gait cycle. Although stem–cementdebonding completely altered the load-transfer in the

reconstruction, causing higher cement stresses in parti-cular, the abductors remained the most influencialmuscle group. Additional inclusion of the iliotibial

Fig. 5. Maximal principal stresses in the cement mantle around a bonded stem at three levels along the stem, as occurring at 30% gait. Stresses are

shown for all four loading configurations. The levels for which the stresses are plotted are: just distally to the resection level (top), mid-stem level

(middle) and stem-tip level (bottom). Activation of the abductors (LC1 vs. LC2) reduced cement stresses around the stem tip at all three gait phases

analyzed. Activation of additional muscle groups hardly affected the cement stresses.

Fig. 6. Bone SED distributions along the pereosteal surface at 45% gait around a debonded stem, plotted for three levels along the femur: an

intertrochanteric level (a, top left), a mid-stem level (b, top right) and a stem-tip level (c, bottom right). No muscles had their attachment at any of

these three levels. Each graph shows the SED value versus the position along the circumference of the femur. A cross-section of the model at each of

the three levels is shown in the corresponding graph, to indicate the position along the circumference. The lateral, posterior, medial and anterior side

of the cross-sections are indicated by L, P, M and A, respectively. The abductors were the main muscles affecting bone SED values. Inclusion of

additional muscle had much less effect.


tract, the adductors and the vastii caused relatively smallchanges to the stress/strain distributions. Their mostprominent effect was a reduction of bone loading in thediaphysis at 10% gait. These results suggest that for pre-clinical tests investigating mechanical failure of acemented THA reconstruction, a loading configurationincluding the hip-joint contact force and the abductorforces would be sufficient.Bone loading in the proximal femur was predomi-

nantly affected by the activation of the abductormuscles. Additional activation of the iliotibial tract,the adductors and the vastii affected the bone strain andSED distributions in the diaphysis early in the gait cycleonly. This suggests that a loading configuration includ-ing the hip-joint force and abductor forces adequatelyreproduces in vivo bone loading in pre-clinical testsinvestigating bone adaptation and stress shieldingaround cemented THA implants. It must be noted,however, that this study only analyzed bone loading inthe reconstructed femur, whereas stress shielding and

bone adaptation analyses are usually based on a changein bone loading in the reconstructed femur relative to theintact one (Cristofolini, 1997; Huiskes, 1992). It is notobvious that a loading configuration including a hip-joint force and abductor forces also adequately repro-duces in vivo loading in the intact femur. In literaturethere is much debate about this subject, particularlyabout the functioning of the iliotibial tract and its effectson bone loading. Ling et al. (1996) and Taylor et al.(1996) claim that the iliotibial tract could considerablyreduce medio-lateral bending of the diaphysis and that itshould therefore be included in pre-clinical testspredicting peri-prosthetic bone adaptation. However,the loading situation in the diaphysis is not of maininterest in pre-clinical tests predicting bone adaptationand stress shielding, since these processes are mostprominent at intertrochanteric levels (Maloney et al.,1996). In a study similar to the current one, Duda et al.(1998) sequentially activated several muscle groups andanalyzed their effects on bone loading in the intact femur

Fig. 7. Bone surface strains in axial direction (z-direction as indicated in Fig. 1) at three locations along the lateral and the medial side of the femur.

Strain values are given as occurring around a bonded stem (top table) and as occurring around a debonded stem (bottom table). Abductor muscles

were the main muscles affecting bone surface strains.


during walking. Their results showed that reasonablyaccurate bone loading in the proximal femur could beobtained using a loading configuration including thehip-joint contact force, abductor forces and the iliotibialtract. However, they did not compare loading config-urations with and without the iliotibial tract, as wasdone in the current study (loading configuration 2 vs. 3)and cannot conclude whether the iliotibial tract shouldbe included or not. Cristofolini et al. (1995) experimen-tally measured strain distributions in the intact proximalfemur alternatingly activating 10 different muscles. As inthe current study, they found that the abductors werethe main muscles affecting the load distribution in theproximal femur. Rohlmann et al. (1982) showed thatincluding an additional iliotibial tract hardly influencedbone loading in the intact proximal femur. From theseresults and those found in the current study, a loadingconfiguration including the hip-joint contact force andthe abductor forces seems sufficient in pre-clinical testspredicting bone adaptation and stress shielding aroundcemented THA implants.In this study, the iliotibial tract was modeled as a

combination of the tensor fasciae latae and a part of thegluteus maximus, both running from the pelvis to thetibia, wrapped around the greater trochanter. Inliterature there is no consensus about the activity andthe magnitude of the force exerted by the iliotibial tract,nor about how it should be modeled. In the data-setused in this study (Brand et al., 1986) the forces in theiliotibial tract ranged from 7.4 to 19.2% bodyweight,depending on the phase of the gait cycle. It was undertension during all three phases analyzed. In contrast,according to Jacob et al. (1982) it is only loaded duringheel strike, whereas according to McLeish and Charnley(1970) it is slack during the entire stance phase of gait.Force magnitudes in different studies vary between 4and 76% bodyweight (Cristofolini, 1997). In addition tothe disagreement about the loading of the iliotibial tract,there seems to be no general agreement about how theiliotibial tract should be modeled. Rohlmann et al.(1982) and Taylor et al. (1996) modeled the iliotibialtract as a tension band, running from the greatertrochanter to the tibia (Fig. 8A). They applied a tensileforce to the greater trochanter to account for its effect,which makes the iliotibial tract behave more like thevastus lateralis (Ling et al., 1996). A second type ofapproach is used by Rybicki et al. (1972) and Finlayet al. (1991), modeling the iliotibial tract as runningfrom the pelvis to the tibia, not wrapping around thegreater trochanter, hence, not exerting direct forces tothe femur (Fig. 8B). In such an approach, the iliotibialtract could just as well be accounted for by adapting thehip-joint contact force, which would seem moreconvenient in the experimental loading configurations.A third type of approach was used in this study,following Duda et al. (1996) and Ling et al. (1996),

modeling the iliotibial tract as a tension band runningfrom the pelvis to the tibia, wrapping around the greatertrochanter (Fig. 8C). Its resultant force pushes theproximal part of the femur medially. It is difficult tosay which way of modeling is most realistic, mainlybecause it is unknown how the iliotibial tract is loadedand how it exerts force on the femur under in vivoconditions (Jacob et al., 1982). Based on the lack ofconsensus about the action of the iliotibial tract and onthe small effect it had on the load-transfer in theproximal femur before and after THA, as shown by ourresults and those of other authors (Rohlmann et al.,1982; Cristofolini, 1995), it is justified to exclude theiliotibial tract from the loading configuration.A considerable number of muscle force data sets can

be found in literature, mostly originating from mathe-matical muscle models. According to Colgan et al.(1994), it is questionable whether the muscle forcespredicted by these models are all representative of invivo muscle loading around the hip joint. The muscleand joint forces that were used in our study originatedfrom a muscle model by Brand et al. (1982, 1986),modified by Duda (1996) to allow wrapping of musclesaround bony structures. One reason to use their muscle-force data was that it is detailed and well documented.More important, however, is that Brand et al. (1994)validated their muscle model by comparing the hip-jointforce, as predicted by their muscle model, to measurewith an instrumented hip. The hip-joint forces showedreasonable agreement.In the current study, we only modeled the proximal

half of the femur, including only those muscles thatattached to that particular part of the femur. The model

Fig. 8. In literature three ways of modeling the iliotibial tract (ITB)

can be found: (a) as a lateral tension band running from the greater

trochanter to the tibia, (b) as a lateral tension band running from the

pelvis to the tibia, (c) as a lateral tension band running from the pelvis

to the tibia, wrapped around the greater trochanter.


was fixed distally. All muscles that acted around thedistal part of the femur, as well as the knee joint andpatellar forces were accounted for in the reaction forcesat the distal fixation of the model. The distance from thestem tip to the fixation was approximately twice thediameter of the diaphysis, which, according to the SaintVenant’s principle, ensured that boundary effects didnot influence the load distribution around the stem.In a separate study, we extensively validated two FE

models similar to the one currently analyzed, byexperimental strain gauge measurements (Stolk et al.,2000). Numerical and experimental bone and cementstrains showed an overall agreement within 10%. Thefemur geometry in the two validated FE models wasobtained from a CAD model of a femur, whereas thefemur geometry in the current model was obtained fromCT-scans. However, the validated models were built inthe same way as the model used in this study. Thevalidation was performed for a loading configurationincluding only hip-joint and abductor forces.This study was confined to analyzing a cemented

THA reconstruction with one particular femur and oneparticular implant design. The specific femur we used tobuild the FE model was selected from a total of 166cadaver femurs in stock. After measuring the dimen-sions of all femurs, this femur was considered to havethe dimensions closest to average. Hence, the conclu-sions drawn in this study may not be valid for odlyshaped femurs. However, using a typical femur is aninherent aspect of developing a standard pre-clinicaltest. The implant, which we analyzed, was a double-tapered collarless stem. For this type of stem, debondingof the stem–cement interface drastically changes theload-transfer as produced by a bonded stem (Huiskes,1990). This is particularly well represented by thecompletely different stress patterns in the cement mantleproduced in the cases of a bonded and a debonded stem.However, although the load-transfer patterns werechanged, the abductors were still the main muscle groupaffecting the stress/strain distributions in all parts of thereconstruction when the stem had completely debonded.Relative to the effect produced by stem–cement debond-ing, adding a collar or making the stem more flexible willprobably have a less pronounced effect on the overallload-transfer pattern. Collared and more flexible im-plants will cause a more proximal transfer of load fromstem to bone compared to the implant modeled in thisstudy (Huiskes, 1990, Crowninshield et al., 1980). Thisgenerally increases the stress/strain levels in bone andcement in the proximal region of the reconstruction and,conversely, reduces these in the distal regions of thereconstruction. This makes the distal region of lowerinterest when predicting mechanical failure of thereconstruction. Since the iliotibial tract, the adductorsand the vastii were noted to affect the stress/strainpatterns primarily distally at the level of the stem tip, the

effects of these muscles on mechanical failure of thereconstruction are probably small for collared and moreflexible stems. However, for implants other thanstemmed ones, such as surface replacement or trans-epiphysial prostheses, this may not be the case. Hence,the conclusions as drawn in this study are valid forstemmed implants only.This study showed that the load distribution in a

cemented THA reconstruction is primarily affected bythe inclusion of abductors in the loading configuration,in addition to the hip-joint contact force. Additionalinclusion of the iliotibial tract, the adductors and thevastii had relatively small effects on stem, cement andbone loading. This would lead to the conclusion that aloading configuration including the hip-joint contactforce and the abductor forces can adequately reproducein vivo loading of a cemented THA reconstruction inpre-clinical tests predicting mechanical failure of thereconstruction and peri-prosthetic bone adaptation.

Acknowledgements

The authors thank Prof. Brand (Department ofOrthopaedic Surgery, University of Iowa Hospitalsand Clinics, Iowa City, IA, USA) for providinginformation about hip- and knee-joint angles duringwalking. This research was performed in the frameworkof European Project SMT4–CT96–2076: ‘‘Pre-clinicaltesting of cemented hip replacement implants: pre-normative research for a European standard’’.

References

Brand, R.A., Crowninshield, R.D., Wittstock, C.E., Pedersen, D.R.,

Clark, C.R., Van Krieken, F.M., 1982. A model of lower extremity

muscular anatomy. Journal of Biomechanical Engineering 104,

304–310.

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., Friedrich, J.A., 1986. The sensitivity of

muscle force predictions to changes in physiologic cross-sectional

area. Journal of Biomechanics 19, 589–596.

Carter, D.R., Hayes, W.C., 1977. The behavior of bone as a two-phase

porous structure. Journal of Bone and Joint surgery A 59, 954–962.

Colgan, D., Trench, P., Slemon, D., McTague, D., Finlay, J.B.,

O’Donnell, P., Little, E.G., 1994. A review of joint and muscle load

simulation relevant to in-vitro stress analysis of the hip. Strain 30,

47–60.

Cristofolini, L., 1997. A critical analysis of stress shielding evaluation

of hip prostheses. Critical Reviews in Biomedical Engineering 25,

409–483.

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.


Crowninshield, R.D., Brand, R.A., Johnston, R.C., Milroy, J.C., 1980.

An analysis of femoral component design in total hip arthroplasty.

Journal of Bone and Joint surgery A 62, 68–78.

Duda, G.N., 1996. Influence of muscle forces on the internal loads in

the femur during gait. Ph.D. Thesis, Springer, Aachen.

Duda, G.N., Heller, M., Albinger, L., Schulz, O., Schneider, E.,

Claes, L., 1998. Influence of muscle forces on femoral strain distri-

bution. Journal of Biomechanics 31, 841–846.

Duda, G.N., Schneider, E., Chao, E.Y., 1997. Internal forces and

moments in the femur during walking. Journal of Biomechanics 30,

933–941.

Finlay, J.B., Chess, D.G., Hardie, W.R., Rorabeck, C.H., Bourne, R.B.,

1991. An evaluatation of three loading configurations for the in vitro

testing of femoral strains in total hip arthroplasty. Journal of

Orthopaedic Research 9, 749–759.

Huiskes, R., 1990. The various stress patterns of press-fit, ingrown,

and cemented femoral stems. Clinical Orthopaedics and Related

Research 261, 27–38.

Huiskes, R., 1993a. Mechanical failure in total hip arthroplasty with

cement. Current Orthopaedics 7, 239–247.

Huiskes, R., 1993b. Failed innovation in total hip replacement. Acta

Orthopaedica Scandinavica 64, 699–716.

Huiskes, R., Weinans, H., Van Rietbergen, B., 1992. The relationship

between stress shielding and bone resorption around total hip

stems and the effect of flexible materials. Clinical Orthopaedics 274,

124–134.

Jacob, H.A.C., Huggler, A.H., Ruttiman, B., 1982. In-vivo investiga-

tions on the mechanical function of the tractus iliotibialis. In:

Huiskes, R., Van Campen, D., De Wijn, J. (Eds.), Biomechanics:

Principles and Applications. Martinus Nijhoff Publishers, The

Hague, pp. 161–167.

Ling, R.S.M., O’Connor, J.J., Lu, T.-W., Lee, A.J.C., 1996. Muscular

activity and the biomechanics of the hip. Hip International 6,

91–105.

Malchau, H., Herberts, P., 1998. Prognosis of total hip replacement.

65th AAOS, New Orleans, USA.

Maloney, W.J., Sychterz, C., Bragdon, C., McGovern, T., Jasty, M.,

Engh, C.A., Harris, W.H., 1996. The Otto Aufranc Award.

Skeletal response to well fixed femoral components inserted with

and without cement. Clinical Orthopaedics 333, 15–26.

Mann, K.A., Bartel, D.L., Wright, T.M., Ingraffea, A.R., 1991.

Mechanical characteristics of the stem-cement interface. Journal of

Orthopaedic Research 9, 798–808.

McLeish, R.D., Charnley, J., 1970. Abduction forces in the one-legged

stance. Journal of Biomechanics 3, 191–209.

Rohlmann, A., Mossner, U., Bergmann, G., Kolbel, R., 1982. Finite-

element-analysis and experimental investigation of stresses in a

femur. Journal of Biomedical Engineering 4, 241–246.

Rybicki, E.F., Simonen, F.A., Weis, E.B.jr., 1972. On the mathema-

tical analysis of stress in the human femur. Journal of Biomecha-

nics 5, 203–215.

Semlitsch, M., Panic, B., 1983. Ten years of experience with test

criteria for fracture-proof anchorage stems of artificial hip joints.

Engineering in Medicine 12, 185–198.

Sommer, H.J., Miller, N.R., Pijanowksi, G.J., 1982. Three-dimen-

sional osteometric scaling and normative modelling of skeletal

segments. Journal of Biomechanics 15, 171–180.

Stolk, J., Verdonschot, N., Cristofolini, L., Firmati, L., Toni, A.,

Huiskes, R., 2000. Strains in a composite hip joint reconstruction

obtained through FEA and experiments correspond closely.

Transactions of the 46th Annual Meeting of the Orthopaedic

Research Society, p. 515.

Taylor, M.E., Tanner, K.E., Freeman, M.A.R., Yettram, A.L., 1996.

Stress and strain distribution within the intact femur: compression

or bending? Medical Engineering and Physics 18, 122–131.

Verdonschot, N., Huiskes, R., 1996. Subsidence of THA stems due to

acrylic cement creep is extremely sensitive to interface friction.

Journal of Biomechanics 29, 1569–1575.


Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip...

Documents

Transcript of Hip-joint and abductor-muscle forces adequately represent in vivo loading of a cemented total hip...