Sie sind auf Seite 1von 7

Medical Engineering & Physics 22 (2000) 453–459

www.elsevier.com/locate/medengphy

Influence of head constraint and muscle forces on the strain


distribution within the intact femur
a,*
J.A. Simões , M.A. Vaz b, S. Blatcher c, M. Taylor d

a
Department of Mechanical Engineering, University of Aveiro, 3810-193 Aveiro, Portugal
b
Department of Mechanical Engineering and Industrial Management, University of Porto, 4200-465 Porto, Portugal
c
IRC in Biomedical Materials, Queen Mary and Westfield College, Mile End Road, London E1 4NS, UK
d
Bioengineering Research Group, School of Engineering Sciences, The University of Southampton, Highfield, Southampton SO17 1BJ, UK

Received 7 March 2000; received in revised form 25 July 2000; accepted 7 September 2000

Abstract

The aim of this study was to analyse the influence of muscle action and a horizontally constrained femoral head on the strain
distribution within the intact femur. The strain distribution was measured for three loading configurations: joint reaction force only,
joint reaction force plus abductors, and joint reaction force plus the abductors, vastus lateralis and iliopsoas. In each case the strains
were recorded from 20 uniaxial strain gauges placed on the medial, lateral, anterior and posterior aspects of the proximal femur.
Application of the abductor muscle force produced a marginal decrease in the strain levels on all aspects of the femur as compared
with the joint reaction force alone. This is in contrast with previous studies which have simulated an unconstrained femoral head.
The inclusion of vastus lateralis and iliopsoas further reduced the strain levels. A horizontally constrained femoral head produces
smaller variation in the strain levels when muscle forces are applied. In vivo data, demonstrating negligible movement of the
femoral head in one-legged stance, support the results of this study and suggest that in the absence of comprehensive muscle force
data, a constrained femoral head may provide a more physiologically relevant loading condition.  2001 IPEM. Published by
Elsevier Science Ltd. All rights reserved.

Keywords: Strain distribution; Loading; Intact composite femur; Strain gauge

1. Introduction mechanisms have been comprehensively reviewed by


Pauwels [11], Frost [7] and Currey [6].
Simulation of physiological loading of the hip is of Many combinations of muscle forces have been mod-
considerable importance to improve prostheses design, elled in vitro in an attempt to simulate femoral loading
bone remodelling simulations and mechanical testing of [2]. Telemetry studies of the implanted hip have given
implants. There has been considerable debate in the us reliable data for the magnitude and direction of the
literature as to how the femur is loaded. The majority joint reaction force. Although the direction of the muscle
of experimental and analytical studies assume the femur forces can be estimated, their true magnitude during gait
to be simplistically loaded, with the application of just or any other activity has yet to be properly defined. The
a joint reaction force or the joint reaction force plus the best estimates of the muscle forces have been supplied
abductors. This generates a characteristic bending from muscle optimisation studies, such as those of
stress/strain pattern within the diaphyseal femur. There Crowninshield et al. [5], Seireg et al. [15] and Röhrle
is growing evidence to suggest that mechanisms exist et al. [14]. All of the optimisation studies predict some
which act to minimise bending to produce a predomi- abductor activity. Röhrle et al. [14] predict significant
nately compressive stress/strain distribution. The various abductor muscle activity, of approximately two times
body weight. Apart from the abductors, there is little
consensus on the other major muscles that are active dur-
* Corresponding author. Tel.: + 351-234-370830; fax: +351-234- ing gait. However, other muscles which have been pre-
370953. dicted as active somewhere in the gait cycle by at least
E-mail address: simoes@mec.ua.pt (J.A. Simões). two studies are: tensor fasciae latae, iliacus, semitend-

1350-4533/01/$ - see front matter  2001 IPEM. Published by Elsevier Science Ltd. All rights reserved.
PII: S 1 3 5 0 - 4 5 3 3 ( 0 0 ) 0 0 0 5 6 - 4
454 J.A. Simões et al. / Medical Engineering & Physics 22 (2000) 453–459

inosus, semimembranosus, sartorius, rectus femoris, diaphysis (Fig. 1). All gauges had the axis of the grid
adductor longus, adductor magnus, biceps femoris and aligned with the longitudinal axis of the femur and were
vastus. However, prediction of muscle forces should be positioned in an identical way to those in the studies of
treated with caution due to the numerous assumptions McNamara et al. [9] and Cristofolini et al. [3]. The strain
made in their calculation. Röhrle et al. [14] reported that gauges were bonded with an adhesive after preparing
a 5 mm displacement of the origin of the rectus muscle the surface of the femur model and connected to a data-
resulted in a 55% decrease in the predicted force of the acquisition system (Solartron SI 35951B IMP, Solartron
adductors. Other sources of error include the location of Instruments). The signal stability of each gauge was
the centre of mass of skeletal segments, the use of skin measured for 2 h, and very small variations were
markers to track the motion of underlying bone during observed. The tests were performed at room temperature,
kinematic studies and the assumption of no antagonistic 22±1°C, and a relative humidity of 65%.
or synergistic muscle activity. Thus, there is a degree of Cristofolini et al. [4] have performed a detailed mech-
uncertainty in the selection of which muscles are anical validation of composite femur models. An exten-
important, and their relevance to mechanical testing. sive experimental validation of the mechanical behaviour
The applied boundary conditions are of equal impor- of 15 composite femurs (Pacific Research Labs) and
tance when modelling the femur as an isolated structure, comparison with four dried and rehydrated and four
but this has been rarely considered. The femoral head fresh-frozen human femurs was carried out. The behav-
can be simulated as being either unconstrained or hori- iour under axial loading and the characterisation of the
zontally constrained. To the authors’ knowledge, all bending and torsional stiffness showed that composite
finite element analyses to date have assumed the femoral femurs fall well within the range for cadaveric femurs,
head to be unconstrained [1,8,12,16,17,19,18,20]. In with no significant differences being detected between
contrast, the majority of experimental studies have the synthetic femurs and the two groups of cadaveric
assumed the femoral head to be horizontally constrained femurs. The interfemur variability for the composite
[2] except for Christofolini et al. [3] and Rohlmann et femurs was 20 to 200 times lower than for the
al. [13], who simulated the femoral head as uncon- cadaveric specimens.
strained. The reason for this division between finite A jig, shown in Fig. 2, was designed and built to apply
element studies and experimental studies, using uncon- the joint reaction force and muscle forces to the femur.
strained and horizontally constrained femoral heads, The load on the femoral head was applied through a
respectively, is not clear in the literature. Yet the effect screw connected to a polyethylene socket and monitored
on the strain distribution could be pronounced.
An in vivo radiological study performed by Taylor et
al. [16] suggested that there is minimal medial deflection
of the loaded femoral head during one-legged stance
(approximately 1.5 mm). Therefore, in an attempt to
simplify experimental set-ups and in the absence of com-
prehensive muscle force data, the question arises
whether it is reasonable to simulate the femoral head as
horizontally constrained. The aim of this study was to
analyse the influence muscle action and a horizontally
constrained femoral head have on the strain distribution
within the intact femur. The effect of three principal
muscle groups, the abductors, the iliopsoas and the
vastus lateralis, was investigated. The results were com-
pared with a experimental study performed by Christo-
folini et al. [3], who conducted a similar study but with
an unconstrained femoral head.

2. Materials and methods

For the purpose of this study, a synthetic composite


femur (model 3103, Pacific Research Labs) was used.
The composite femur was prepared with 20 uniaxial
gauges (CEA-06-125UN-350, Measurements Group,
Inc., Raleigh, NC) placed at five levels, namely on the Fig. 1. Composite femur bone model prepared with 20 uniaxial
lateral, medial, anterior and posterior sides of the strain gauges.
J.A. Simões et al. / Medical Engineering & Physics 22 (2000) 453–459 455

Fig. 2. Experimental jig to apply the hip joint reaction and muscle forces.

by a load cell. The femur was distally fixed by the fem- to load the muscles in the required direction. Fig. 3 also
oral condyles on a platform, which allowed us to pos- shows how the femoral head movement was constrained.
ition the femur in the desired position. The muscle forces Three different loading conditions were analysed; the
were exerted by gravity through weights fixed to metallic magnitudes and directions of the forces are shown in
straps that were connected to metallic plates on the Table 1 (Fig. 4). In the first load configuration, load case
femur as shown in Fig. 3. The jig was surrounded by a 1, the femur was loaded with a joint reaction force pos-
frame structure where pulleys were positioned in order itioning the femur 13° in adduction and 7° in flexion.

(a) (b)
Fig. 3. (a) Plates and cables attached to the femur head for muscle force application and (b) device used to constrain the femur head horizontally.
456 J.A. Simões et al. / Medical Engineering & Physics 22 (2000) 453–459

Table 1
Muscle and joint reaction forces for the three load cases

Load case Resultant Angle (degree)


force (N)

f g

1
Joint reaction force 700 167 21
2.
Joint reaction force 700 167 16
Abductors 300 20 180
3.
Joint reaction force 730 159 7
Abductors 300 20 180
Iliopsoas 188 47 262
Vastus lateralis 292 180 –

Load case 2 replicates the loading configuration used by microstrain. The addition of iliopsoas and vastus lat-
several authors, consisting of a joint reaction force eralis, load case 3, resulted in a further reduction in the
applied to the femoral head and an abductor muscle force strains observed. In general, there was a 50–150
applied at the great trochanter [8,12,16–18]. In load case microstrain reduction on the medial aspect of the femur
3, the angle of the joint reaction force was increased to and a 100–250 microstrain reduction on the lateral aspect
20° in the transverse plane to assess its effect on the of the femur as compared with load case 2. On the
overall strain distribution, and the action of the three anterior and posterior aspects of the femur there was
major muscle groups, the abductors, the vastus lateralis little change in the strain magnitudes observed in the
and the iliopsoas, was applied [16]. The simulations proximal femur. However, there was a progressive
were performed with the distal end of the femur rigidly reduction in the strain to a peak reduction of 200
constrained and, for each case, the axial strain distri- microstrain at the most distal location, as compared with
bution was obtained on the medial, lateral, anterior and load cases 1 and 2. There was a peak difference in strain
posterior regions of the diaphyseal femur. of approximately 1500 microstrain, a reduction of nearly
50% as compared with load case 1.

3. Results
4. Discussion
The axial strains on the medial, lateral, anterior and
posterior aspects of the femur for the three load cases The purpose of this study was to examine the influ-
are shown in Fig. 5. Load case 1 displays a typical bend- ence of a horizontally constrained femoral head and
ing strain distribution, with high medial and lateral muscle forces on the strain distribution within the intact
strains (⫺1500 and 1200 microstrain, respectively), and femur. Cristofolini et al. [3] performed a similar study,
significantly lower strain values on the anterior and pos- except that they simulated the femoral head as uncon-
terior surfaces of the femur (⫺850 and 500 microstrain, strained and studied the action of 10 muscles instead of
respectively). This resulted in a peak difference in strain three as in this study. Direct comparison of the strain
of 2700 microstrain. The addition of the abductors’ mus- magnitudes is difficult, as Cristofolini et al. used a
cle force in load case 2 produced a decrease of approxi- cadaveric femur instead of the composite femur used in
mately 300 microstrain over the entire medial and lateral this study. The cadeveric femur was also 50 mm smaller
aspects of the femur. However, there was only a mar- than the composite femur that we used in our study. For
ginal decrease in the strains observed on the anterior and similar load case 1, Cristofolini reported peak medial,
posterior aspects of the femur. For load case 2 there was lateral, anterior and posterior strains of ⫺900, 450, 250
a peak difference in strain of approximately 1800 and ⫺400 microstrain, respectively. In this study peak
J.A. Simões et al. / Medical Engineering & Physics 22 (2000) 453–459 457


f
JRF JRF
Fabd Fabd
20º

FIp

FIp
Fvl

47º

y JRF - Joint Reaction Force


Fabd - Abductors y
FIp - Iliopsoas
x Fvl - Vastus lateralis
z
Fig. 4. The applied joint reaction and muscle forces.

strains of ⫺1500, 1200, 500 and ⫺850 microstrain were the medial and lateral aspects of the femur; i.e., for a
observed on the medial, lateral, anterior and posterior unconstrained femoral head simulation, it would appear
aspects of the femur, respectively, for an identical load that the strain distribution is dominated by the action of
case. Hence the strains observed in the composite femur the abductors. In this study the application of iliopsoas
used in this study were 1.7 to 2.5 times higher than those and vastus lateralis produced a further decrease of 25%
observed in the cadaveric femur used by Cristofolini et in the medial strain levels and a reduction of up to 50%
al. In Cristofolini’s study, the application of an abductor in the lateral strain levels as compared with load case 2.
muscle force produced an increase of over 300% in the Similar reductions in the strain values were seen in the
peak medial strains and an increase of over 400% in the distal femur for the anterior and posterior aspects of the
lateral strain as compared with the femur loaded without femur. Therefore it would appear that the tension band-
muscles. This finding is contrary to those reported here, ing effect of vastus lateralis is still effective in a horizon-
where there was a decrease of approximately 30% in the tally constrained femoral head experimental set-up.
axial strains on both the medial and lateral aspects of In vivo data suggest that there is negligible movement
the femur. Since only the vertical motion of the femoral of the femoral head during one-legged stance [16] as
head was allowed in this study, the action of the abductor compared with the unloaded femur. It has been proposed
force pulling the femoral head medially was resisted by that muscles act to minimise bending during all activities
the loading mechanism. In effect, the horizontal femoral to produce a predominantly compressive strain distri-
head constraint increases the horizontal component of bution within the intact femur [6,7,11,10]. If this is true,
the joint reaction force and minimises femoral bending. then it is also reasonable to assume that muscles also act
However, it should be noted there are fundamental dif- to minimise the deflection of the femoral head during
ferences between the loads applied in Cristofolini’s normal daily activities. If this is the case, then it is
study and those used here. In Cristofolini’s study, the reasonable to simulate the femoral head as horizontally
abductor muscle force consisted of a resultant force of constrained. This study has shown that such a constraint
nearly four times body weight as opposed to just 1.3 acts to resist the horizontal loads applied by the abduc-
times body weight used in this study. tors, thus reducing the degree of bending present in the
In Cristofolini’s study the addition of further muscles femur. However, the use of a constrained head does not
had only a small influence on the strain distribution on produce a predominately compressive strain distribution
458 J.A. Simões et al. / Medical Engineering & Physics 22 (2000) 453–459

Fig. 5. Strain distributions within the medial, lateral, anterior and posterior aspects of the femur.

alone and the application of further muscle forces, such femoral head results in the abductor muscles signifi-
as iliopsoas and vastus lateralis, are need to produce such cantly increasing the femoral bending strains. In vivo
a strain state. However, further work is required to see if data appear to contradict the validity of this type of
the application other muscle forces will produce a more experimental set-up. In the absence of comprehensive
uniform strain distribution. The advantage of the con- muscle force data, a constrained femoral head may pro-
strained femoral head set-up is its simplicity. Complex vide a more physiologically relevant loading condition.
experimental set-ups are required to simulate an uncon-
strained femoral head.
References
[1] Bergmann G, Craichen F, Rohlmann A. Hip joint loading during
5. Conclusions walking and running measured in two patients. J Biomech
1993;26:969–90.
The objective of this study was to analyse the influ- [2] Colgan D, Trench P, Slemon D, McTague D, Finlay JB, O’Don-
ence of muscle simulation on the strain distribution nell P. et al. A review of joint and muscle load simulation relevant
to in-vitro stress analysis of the hip. Strain 1994;May:47–61.
within the diaphyseal of the femur. Femoral head bound- [3] Cristofolini L, Viceconti M, Toni A, Giunti A. Influence of thigh
ary conditions have been shown to influence the strain muscles on the axial strains in a proximal femur during early
distribution within the intact femur. An unconstrained stance in gait. J Biomech 1995;28(5):617–24.
J.A. Simões et al. / Medical Engineering & Physics 22 (2000) 453–459 459

[4] Cristofolini L, Viceconti M, Cappello A, Toni A. Mechanical [13] Rohlmann A, Mossner U, Bergmann G, Kolbel R. Finite-element-
validation of whole bone composite femur models. J Biomech analysis and experimental investigation of stresses in a femur. J
1996;29(4):525–35. Biomed Eng 1982;4:241–6.
[5] Crowninshield RD, Johnston RC, Andrews JG, Brand RA. A [14] Röhrle H, Scholten R, Sigolotto C, Sollbach W, Kellner H. Joint
biomechanical investigation of the human hip. J Biomech forces in the human pelvis–leg skeleton during walking. J
1978;11:75–85. Biomech 1984;17:409–24.
[6] Currey JD. The mechanical adaption of bones. Princeton, NJ: [15] Seireg A, Kempke W. Behaviour of in vivo bone under cyclic
Princeton University Press, 1984. loading. J Biomech 1975;2:455.
[7] Frost HM. Bone remodelling and skeletal modelling errors. [16] Taylor M, Tanner E, Freeman MAR, Yettram AL. Stress and
Springfield (IL): Orthopaedics Lecturers, 1973. strain distribution within the intact femur: compression or bend-
[8] Huiskes R. The various stress patterns of press-fit, ingrown and ing? Med Eng Phys 1996;18(2):122–31.
cemented femoral stems. Clin Orthop Rel Res 1990;261:27–38. [17] Tensi HM, Gese H, Aschrel R. Non-linear three dimensional
[9] McNamara BP, Viceconti M, Cristofolini L, Toni A, Taylor D. finite element analysis of a cementless hip prosthesis. Proc Instn
Experimental and numerical pre-clinical evaluation relating to Mech Engrs 1989;203H:215–22.
total hip arthroplasty. In: Middleton J, editor. Proceedings of 2nd [18] Verdonschot NJJ, Huiskes R, Freeman MAR. Pre-clinical testing
International Symposium on Computer Methods in Biomechanics of hip prosthetic designs: a comparison of finite element calcu-
and Biomedical Engineering. The Netherlands: Gordon & Beach, lations and laboratory tests. Proc Instn Mech Engrs
1996:1–0. 1993;207H:149–54.
[10] Munih M, Kralj A, Bajd T. Bending moments in lower extremity [19] Weinans H, Huiskes R, Grootenboer HJ. Trends of mechanical
bones for two standing postures. J Biomed Eng 1992;14:293–301. consequences and modeling of a fibrous membrane around fem-
[11] Pauwels F. Biomechanics of the locomotor apparatus. oral hip prostheses. J Biomech 1990;23:991–1000.
Berlin/Heidelberg/New York: Springer Verlag, 1980. [20] Yettram AL. Effect of interface conditions on the behaviour of
[12] Prendergast PJ, Taylor D. Stress analysis of the proximo-medial a Freeman hip endoprosthesis. J Biomed Eng 1989;11:520.
femur after total hip replacement. J Biomed Eng 1990;12:379–82.

Das könnte Ihnen auch gefallen