Post on 24-Jan-2023
1
The relationship between cement fatigue damage and implant surface finish in proximal femoral prostheses A. B. Lennona, B. A. O. McCormackb, and P. J. Prendergasta,* a Centre for Bioengineering, Department of Mechanical Engineering, Trinity College,
Dublin, Ireland. b School of Engineering, Institute of Technology, Sligo, Ballinode, Co. Sligo, Ireland.
First submitted February, 2003 Revised version, June 2003. *Corresponding Author Patrick J. Prendergast, Ph.D. Centre for Bioengineering Department of Mechanical Engineering, Trinity College, Dublin, Ireland Tel: +353 (0)1 608 1383 Fax: +353 (0)1 679 5554 E-mail: pprender@tcd.ie www.biomechanics.ie
2
Abstract
The majority of cemented femoral hip replacements fail as a consequence of
loosening. One design feature that may affect loosening rates is implant surface finish.
To determine whether or not surface finish affects fatigue damage accumulation in a
bone cement mantle, we developed an experimental model of the implanted proximal
femur that allows visualization of damage growth in the cement layer. Five matt
surface and five polished surface stems were tested. Pre-load damage and damage
after two million cycles was measured. Levels of pre-load (shrinkage) damage were
the same for both matt and polished stems; furthermore damage for matt vs. polished
stems was not significantly different after two million cycles. This was due to the
large variability in damage accumulation rates. Finite element analysis showed that
the stress is higher for the polished (assumed debonded) stem, and therefore we must
conclude that either the magnitude of the stress increase is not enough to appreciably
increase the damage accumulation rate or, alternatively, the polished stem does not
debond immediately from the cement. Significantly (p=0.05) more damage was
initiated in the lateral cement compared to the medial cement for both kinds of surface
finish. It was concluded that, despite the higher cement stresses with debonded stems,
polished prostheses do not provoke the damage accumulation failure scenario.
Keywords: Bone cement, hip replacement, Exeter prosthesis, porosity, damage
accumulation, arthroplasty
3
Nomenclature σi ith principal stress component
σi max maximum of ith principal stress around a pore
σ max maximum stress around a pore
ν Poisson’s ratio
A, B, C constants used in the expressions for evaluating stress around a pore
FI interface failure index
St interface tensile strength
Sc interface compressive strength
Ss interface shear strength
σ interface normal stress
τ interface shear stress
1
1 Introduction
The success or otherwise of a total hip replacement operation is influenced by the
surgeon’s choice of hip prosthesis1. One design feature that differs considerably between
prostheses is surface finish2,3,4. Ling4 observed that a polished surface of a particular
geometry (i.e., the Exeter prosthesis) subsided within the cement mantle. Therefore,
subsidence of polished implants allows “self-tightening”5; this phenomenon is now taken
to be the basis of how implants with polished surfaces can achieve greater longevity
compared to implants with a matt surface. Huiskes et al.6 have proposed that polished
hip prosthesis stems may be termed “force-closed designs” because they achieve stability
when the load applied is balanced by the frictional force on the debonded stem, and that
matt surface prostheses may be termed “shape-closed designs” because they aim to
remain bonded with the cement mantle.
It is not a simple matter to say what event begins the loosening process of an
orthopaedic implant prosthesis. If debonding of the stem/cement interface begins the
process then the force-closed (polished) designs should commence the loosening process
more rapidly than shape-closed (matt) designs and it would follow that polished stems
should fail faster – however this does not accord with clinical findings. For example,
follow-up studies reported by the Swedish hip register1 show that the Exeter polished
stem outperforms the Exeter Matt as well as many other matt stems. The retrievals of
Jasty et al.7 of 16 specimens found that some debonding was present in every case but
only one case showed radiographic loosening. Similarly, Fornasier and Cameron8 found
all five stems of an autopsy retrieval study to be debonded at the cement/bone interface
but none to be clinically loose. Therefore, it would seem that mechanical loosening by
damage accumulation is not provoked by debonding. In fact, the evidence of the superior
2
performance of the Exeter polished to the Exeter matt prosthesis points to the opposite
conclusion9. Investigations using both physical and numerical models led Verdonschot
and Huiskes10 to conclude that increased roughness for straight-tapered stem designs
could reduce subsidence but would inevitably lead to increased cement damage and wear.
The issue of subsidence and surface finish was again addressed by Speirs et al.11 who
demonstrated that a matt stem could exhibit greater cyclic motion than a polished stem in
spite of lower overall migration. However, their matt and polished stems had different
geometries so the effect of surface finish alone could not be separated from other factors.
In this paper, we aim to find out whether or not a matt stem (i.e. one that tends to
stay bonded) reduces damage accumulation in the cement mantle relative to polished
stems. Since computational approaches have been tried before, it was decided to take an
experimental approach by using a physical model of the implanted proximal femur. The
physical model is a development of the one used by McCormack and Prendergast12 which
allows exposure of the cement so that quantitative measurements of damage
accumulation can be made.
2 Methods
The physical model of the femoral-side of T.H.R.
The physical model consists of a curved implant fixated into the space formed between two
aluminium ‘covers’ (side plates). The resulting rectangular structure has a second moment of area
of 19,026 mm4 at the mid section of the stem, which compares to a value of 25,591 mm4
estimated from cross sectional data of Noble et al. 13 for the isthmus of the femur. The inside of
the covers are layered with cancellous bone strips; cancellous bone margins have been predicted
3
to affect prosthesis subsidence rates14. These were formed from bovine rib bone, cut and pressed
to conform to the curved inside walls of the aluminium, and were fixed using epoxy resin once
they had conformed. ‘Windows’ in the side allowed direct observation of microdamage
accumulation in the cement. Physiological features include interdigitated cement-bone interfaces
(achieved due to the cancellous bone strips) and a trochanteric process enabling the attachment of
an abductor load through a lever (Fig. 1a). Polyethylene inserts were used to prevent cement
escaping during preparation and to keep the cement surface continuous with the exposed stem and
bone surfaces. This also allowed pressure to build up in the specimen ensuring cement
interdigitation with the cancellous bone strips. Simplex Rapid cement was hand mixed for
approximately 60 s at 1 beat/s and then introduced into the cavity. This cement has the same
chemical composition as Simplex P without radio-opaque filler or antibiotics to give a translucent
cement that is suitable for damage visualisation by light transmission. Cement flowed around the
stem and ~5 mm beneath the tip where it was occluded by a plug. The specimen was allowed to
cure for 24 h. Specimens were stored for at least 1 week after preparation. Five polished and five
matt surface stems, i.e. 10 stems in total, were tested (Fig. 1b). The surface fabrication of the
stems was: (i) grit blasted, with mean Ra = 2.12 µm and (ii) polished, with mean Ra = 0.04 µm.
The surface finish was measured using a white light interferometer.
Crack measurement procedure
Dye penetrant was applied to the cement layers of each specimen before testing. An optical
comparator with a ×20 lens was used to project a magnified image of the cement surface onto a
screen. All observed cracks were traced onto acetate transparencies; these had markings placed on
them to allow referral to the comparator measurement system. If a crack was seen to extend
below the surface the focus was changed to assess the full projected length of the crack. The
transparencies were scanned and thresholded to remove any background greyscale from the
4
scanning operation. Image analysis was used to calculate the position, length, and slope of each
crack for each physical model specimen.
2.3 Fatigue testing
Specimens were clamped in 12° adduction and a load was applied to the lever. The load
applied produced a joint reaction force of 2.9 kN through the implant head-centre at 20°
to the long axis of the stem and an abductor load of 1.6 kN at 15° to vertical (Fig. 1a). The
joint load corresponds to 4.2 BW, assuming a 70 kg body mass. The abductor load (55%
of joint load) is similar to that used by Burke et al.15 in a rig designed to simulate single
leg stance. Specimens were tested at 5 Hz for 2×106 cycles in air at room temperature;
higher frequencies of loading were not used to avoid the possibility of cement heating. In
addition, cyclic actuator displacements were recorded.
2.4 Finite element stress analysis
Only half the specimen was modelled because it has mirror symmetry about the frontal
plane (Fig. 1c). Eight-node hexahedral finite elements were used (ANSYS type
SOLID45). Nodes corresponding to the clamped surfaces of the specimen (as shown in
the loading schematic, Fig. 1b) were restrained normal to each clamped surface. Linear
elastic material properties were used (Table 1). Loads were applied to the centres of the
head and muscle attachment, which were attached to the mesh using truss elements
(ANSYS type LINK8). Two interfacial conditions were simulated:
(i) a bonded interface modelled by coupling degrees of freedom for coincident
nodes and
5
(ii) debonding with a low coefficient of friction (µ = 0.25) modelled using
surface-surface contact elements (ANSYS types CONTA173 and
TARGE170).
To investigate the possible effect of porosity on stress, an approximation for the
maximum stress occurring around a pore was made. The analytical solution for stress at
the pole and equator of a pore due to a remote uniaxial stress16 can be superposed to give
the stress components at the pole and equator with respect to each principal stress of a
multiaxial stress state. The maximum of these superposed stresses can then be found
according to Eqn. 1:
�������
�
�
�������
�
�
��
���
�
��
���
�
��
���
�
321
3213max
321
3212max
321
3211max
max
Max =
Max =
Max =
Max =
+ Bσ + AσCσ + Bσ + CσAσ + Aσ + BσCσ + Cσ + BσAσ + Aσ + CσBσ + Cσ + AσBσ
σ
σ
σ
σ , (1)
where
( )( ) ( ) ( ) ,
572315,
5721527,
572153
ν-ν- = C
ν-ν-= B
ν-ν+-= A
and ν is Poisson's ratio. This was done for the centroid of every cement finite element in
order that regions that were more susceptible to damage from pores could be identified.
To assess the likelihood of debonding, a failure index was calculated at the
cement-metal interface using a Hoffman criterion17 with constants taken from
Verdonschot and Huiskes18,
22sct
2
ct
1111 τσσSSSSS
FI +���
����
�−+= (2)
6
where σ and τ are interface normal and shear stress respectively, St = 8 MPa is the tensile
strength of the interface, Sc = 70 MPa is the compressive strength, and Ss = 6 MPa is the
shear strength. Higher positive values imply greater tendency for failure and failure is
assumed to occur at FI = 1.
3 Results
Examination of the specimens before loading identified pre-load cracks in both medial
and lateral cement layers of both matt and polished stems, see Fig. 2. If damage is
quantified as the sum-of-crack-lengths (i.e., �=
n
ia
1i where ia is the length of the ith crack
and there are n cracks) then, as expected, there is no significant difference in the amount
of pre-load (shrinkage) damage for the matt versus polished stems. The polished stems
tended to have more damage than the matt stems after 2 m cycles of loading, see Fig. 2
and Table 2, but this was not statistically significant.
Some considerable variation was found in both quantity of damage (Table 2) and
spatial distribution of damage (Fig. 2). Accumulation of damage in the lateral cement was
higher in the polished specimens compared with their matt counterparts (p = 0.05, Table
2). This can be explained by noting that debonding raised the tensile stress level and
caused a significant change in the stress transfer pattern, with relief of stress in the distal-
lateral and mid-medial regions and elevation of the stress in the mid-lateral, proximo-
lateral, and disto-medial regions (Fig. 3). Inclusion of porosity according to Eqn. (1)
raised the stresses to failure levels for both bonded and debonded stems over much larger
regions (Fig. 3). Calculation of volumes of cement stressed to different tensile stress
ranges illustrates, quantitatively, how debonding influences the stress distribution in the
7
cement layers (Fig. 4) — only a small region on the lateral side of the bonded model
experienced stress above 3 MPa while both lateral and medial sides of the debonded
model had a relatively large proportion of the cement volume raised to higher stresses.
Specimens with high amounts of pre-load (shrinkage) damage had highest
accumulated damage after two million cycles (Fig. 5). However, one of the specimens for
both the matt and polished groups still did not fit the observed trend and a good
correlation could only be achieved by excluding one specimen from each group, see Fig.
5. We propose that an explanation for this might be that the polished stem outlier (P3)
had not debonded and that the matt stem outlier (M5) had debonded.
Cyclic actuator displacement (or “inducible displacement”) was of a similar
magnitude for both matt and polished stems (Fig. 6). A regression analysis was
performed to determine if inducible displacement was correlated with total damage. It
was found that much of the increase in the measured inducible displacement could be
attributed to an increase in the total damage present (Fig. 6). In contrast to this, a similar
analysis for the displacement under peak or minimum load (i.e., migration) showed poor
correlation with damage (R2=0.11 and R2=0.19, respectively).
4 Discussion
This study attempted to quantitatively assess the effect of implant surface finish on
cement damage accumulation. There were several limitations in the experimental
approach. The main one was that there are no hoop stresses since a layered cement
mantle, and not an annular one, was formed; hoop stresses are known to be significant in
supporting polished prostheses19. This simplification of the real in vivo structure was
8
necessary to facilitate visualization and counting of the microcracks. Other limitations of
the model were
(i) Only cracks that intersect the surface can be stained and therefore cracks
wholly within the mantle were not counted. This leads to an underestimate of
the number of cracks. Furthermore, the measurement is based on an
orthogonal projection of each crack and this will cause an underestimation of
crack length for those cracks which have their longest axis at an angle to the
viewing plane.
(ii) Blood and fat inclusions have been shown to reduce properties of both the
cement-metal interface as well as the bulk cement itself20. Damage
accumulation could therefore be higher for in vivo reconstructions.
(iii) Only an abductor muscle load was included in the model. In support of this
simplification, a recent study has shown that inclusion of more muscles does
not greatly affect cement stresses during walking21. Also stumbling loads are
not included and therefore damage accumulation rates measured are likely to
be conservative. A further simplification with respect to loading is that
frictional torques generated at the head are neglected; a complex mechanical
system such as that recently published by Lui et al.22 would be required to
implement this.
(iv) The solution for the stress in a porous cement assumes an isolated pore,
without any nearby boundaries or neighbouring pores, and a remotely applied
stress. However, the stress raising effect of the pore is very localised (for a
uniaxial case the stress drops back to within 5% of the remote value at twice
9
the radius from the pore centre16), so we believe this is an acceptable
simplification.
(v) To more accurately test a force-closed concept for the polished stems, it
would be better to remove cement distal to the stem tip. However, the intent of
the study was to keep as many factors as possible common between the two
groups. Future work could investigate this issue more fully.
(vi) Due to the load controlled nature of the test, creep of the cement would
therefore be expected in the experimental model; however viscoelasticity was
not included in the finite element analysis.
Despite these limitations with respect to the physical model, it does succeed in allowing a
view of damage accumulation in the cement mantle. Visualization of crack growth would
not be possible with anything other than a layered system, although more complex
methods using acoustic emission could have been used; if used in combination with some
form of internal strain measurement, such as that developed by Cristofolini and
Viceconti23, a system for monitoring both damage accumulation and strain redistribution
due to debonding could be envisaged for a physiological reconstruction.
We did not confirm the hypothesis that a matt stem reduces damage accumulation
in the cement mantle compared to a polished stem. Even if the clear trend was that the
polished stem had more damage accumulation, the difference was not significant for the
cement mantle taken as a whole; this was because the variation in damage accumulation
between specimens was considerable. Since it is likely that an even greater variation
exists clinically, the conclusion of indistinguishable damage accumulation for matt and
polished stems is likely to hold in the clinical setting.
10
Taking the lateral cement mantle on its own, there was significantly less damage
for the matt stem compared to the polished stem. An increase in volumes of highly
stressed cement for the case of complete debonding (Fig. 4) offered evidence that high
stress causes the trend towards higher damage accumulation observed in the polished
group. However, when comparing contour stresses for both models with the observed
spatial distributions of cracking (compare Fig. 3 with Fig. 2), cracks appeared in regions
where stress was predicted to be low, or even compressive. This could be explained by
including porosity in the stress calculation (Fig. 3), which showed that tensile stress could
be experienced over the entire exposed cement layers.
The main result of our research on this topic is that damage accumulation is so
variable that the damage accumulation failure scenario is not easily controlled by
prosthesis design parameters. The variability is both spatial (Fig. 2) and in terms of the
total quantity of damage (Table 2). Similar variability has been found under fatigue
loading of standard specimens in uniaxial tension24,25; therefore it may be concluded that
the variability is not primarily due to features of this physical model but rather is a
characteristic of the cement itself — most probably porosity and the random nature of
pre-load damage.
Pre-load cracking was one feature common to both matt and polished stems.
Lennon and Prendergast26 have previously predicted that residual stress due to
polymerisation-induced shrinkage can be sufficient to cause cracking in the presence of
pores. The lack of a significant difference in the levels of pre-load damage between matt
and polished stems suggests that the polished stems did not debond before testing, i.e. did
not debond because of shrinkage.
Because the trend in damage accumulation is different for matt versus polished
stems, it can be surmised that similar interfacial conditions did not persist after cyclic
11
loading commenced. Two specimens exhibited aberrant behaviour; to attempt to explain
this we calculated a Hoffmann failure index for the bone-cement/metal interface. It is
highest on the medial side (Fig. 7) and indicating that a weak matt interface would
debond on that side first. Similarly it is lowest on the lateral side (Fig. 7) and, therefore, if
a polished stem had not debonded on one of its interfaces it is likely to be the lateral one
— and it is on the lateral side that one of the polished specimens (P3) behaves as part of
the matt group. Therefore it is concluded that, at two million cycles, some polished
specimens may not debond much whereas some matt specimens may debond
significantly. Similar variability in debonding may also occur in vivo with cemented hip
prostheses, although polished stems such as the Exeter are more likely to debond more
completely and rapidly due to the lack of distal support from the cement.
Cyclic displacement (often referred to as 'inducible' displacement27) showed a
good correlation with damage (see Fig. 6 and section 4). This supports the use of
inducible displacement as a measure for discriminating between different prosthesis
designs27. The lack of correlation between permanent migration and damage suggest that
other phenomena determine migration, such as cement creep and debonding of the stem.
It should also be noted that residual stresses normal to the stem-cement interface may
affect the load-displacement behaviour, as shown by Mann et al28 in tapered and straight
stem models.
5 Conclusions
1) Analysis of cement cracking in an experimental model of the femoral side of a
total hip replacement has shown that, despite generally higher stresses in cement
mantles of polished stems, the microdamage does not apparently accumulate at a
12
faster rate for those stems. Therefore the hypothesis that overall damage
accumulation was greater around polished stems than around matt stems was not
confirmed (p>0.05).
2) The reason seemed to be that the variability in the strength of the bone cement
due to (a) porosity and (b) pre-load (shrinkage) damage masks the effect of the
increase in stress that occurs with a polished surface finish.
3) A consequence of damage accumulation is increased “inducible displacement” of
the stem relative to the bone. Therefore inducible displacements may be a
measure of the extent of damage accumulation in a cemented joint replacement,
and could be used in a pre-clinical test.
Acknowledgements
The MediLink project of the PRTLI programme provided financial support of one of us
(Dr A.B. Lennon) for the computational part of this research. Further financial support
was provided by the Standards Measurements and Testing Program of the European
Commission (Contract No. SMT4-CT96-2076). Sulzer Medica, Winterthur, Switzerland,
are thanked for stem polishing and the Department of Mechanical Engineering,
University College Dublin for use of their optical comparator.
13
References
[1] Malchau, H., Herberts, P., Söderman, P. and Odén, A., Prognosis of total hip
replacement: Update and validation from the Swedish national hip arthroplasty registry
1979–1998. In Scientific Exhibition presented at the 67th Annual Meeting of the American
Academy of Orthopaedic Surgeons (Orlando, USA), 2000, pp. 1–16, [
www.jru.orthop.gu.se ].
[2] Karrholm, J., Nivbrant, B., Thanner, J., Anderberg, C., Borlin, N., Herberts, P.,
Malchau, H., Radiostereometric evaluation of hip implant design and surface finish.
Micromotion of cemented femoral stems. In Scientific Exhibition presented at the 67th
Annual Meeting of the American Academy of Orthopaedic Surgeons (Orlando, USA), 2000
pp. 1–16, [ www.jru.orthop.gu.se ]
[3] Harris, W. H., Is it advantageous to strengthen the cement-metal interface and use a collar
for cemented femoral components of total hip replacements? Clinical Orthopaedics and
Related Research, 1992, 285, 67–72.
[4] Ling, R.S.M., The use of a collar and precoating on cemented femoral stems is
unnecessary and detrimental. Clinical Orthopaedics and Related Research, 1992, 285, 73–
83.
[5] Lee, A.J.C., Rough or polished surface on femoral anchorage stems?, In Technical
Principles, Design, and Safety of Joint Implants, Eds. G.H. Buchhorn and H.G. Willert,
Toronto, Hogrefe and Hubber, 1994, p.p. 209–211.
[6] Huiskes, R., Verdonschot, N., Nivbrant, B., Migration, stem shape, and surface finish in
cemented total hip arthroplasty. Clinical Orthopaedics and Related Research, 1998, 355,
103–112.
[7] Jasty, M., Maloney, W. J., Bragdon, C. R., O'Connor, D., Haire, T. and Harris, W. H.,
The initiation of failure in cemented femoral components of hip arthroplasties. Journal of
Bone and Joint Surgery, 1991, 73-B, 551–558.
[8] Fornasier, V. L. and Cameron, H. U., The femoral stem/cement interface in total hip
replacement. Clinical Orthopaedics and Related Research, 1976, 116, 248–252.
[9] Crawford, R.W., Gie, G.A. and Ling, R.S.M. An 8-10 year clinical review of comparing
matt and polished Exeter stems. Orthopaedic Transactions 1998, 22, p 40.
14
[10] Verdonschot, N. and Huiskes, R., Surface roughness of debonded straight-tapered stems
in cemented THA reduces subsidence but not cement damage. Biomaterials, 1998, 19,
1773–1779.
[11] Speirs, A. D., Slomczykowski, M. A., Orr, T. E., Siebenrock, K. and Nolte, L. P.,
Three-dimensional measurement of cemented femoral stem stability: an in vitro cadaver
study. Clinical Biomechanics, 2000, 15, 248–255.
[12] McCormack, B.A.O., Prendergast, P.J., Microdamage in the cement layer if hip
replacements under flexural loading. Journal of Biomechanics, 1999, 32, 467–476.
[13] Noble, P. C., Alexander, J., Lindahl, L. J., Nalty, T., and Tullos, H. S., The geometry of
the proximal femur and the design of hip endoprosthes. In Technical Principles, Design,
and Safety of Joint Implants, Eds. G.H. Buchhorn and H.G. Willert, Toronto, Hogrefe and
Hubber, 1994, p.p. 209–211.
[14] Taylor, M., Tanner, K.E., Freeman, M.A.R. and Yettram, A.L., Cancellous bone
stresses surrounding the femoral component of a hip prosthesis – an elastic-pastic finite
element analysis. Medical Engineering and Physics 1995, 17, 544-550.
[15] Burke, D. W., O'Connor, D. O., Zalenski, E. B., Jasty, M. and Harris, W. H.,
Micromotion of cemented and uncemented femoral components. Journal of Bone and Joint
Surgery [Br], 1991, 73-B, 33–37.
[16] Timoshenko, S. P. and Goodier, J. N., Theory of Elasticity (McGraw Hill), 3rd ed., 1970,
396–398.
[17] Hoffman, O., The brittle strength of orthotropic materials. Journal of Composite Materials,
1967, 1, 200–206.
[18] Verdonschot, N. and Huiskes, R., Cement debonding process of total hip arthroplasty
stems. Clinical Orthopaedics and Related Research, 1997, 336, 297–307.
[19] Lennon, A.B., Prendergast, P.J., Evaluation of cement stresses in finite element analyses
of cemented orthopaedic implants. Journal of Biomechanical Engineering, 2001, 123, 623–
628.
[20] Lee, A.J.C., Ling, R.S.M, and Vangala, S.S., Some clinically-relevant variables affecting
the mechanical behaviour of bone-cement. Arch. Orthop. Trauma Surg., 1978, 92, 1-18.
15
[21] Stolk, J., Verdonschot, N. and Huiskes, R., Hip-joint and abductor-muscle forces
adequately represent in vivo loading of a cemented total hip reconstruction. Journal of
Biomechanics, 2001, 34, 917–926.
[22] Liu, C., Green, S.M., Watkins, N.D., Greeg, P.J. and McCaskie, A.W., A preliminary
hip joint simulator study of the migration of a cemented hip prosthesis. Proceedings of the
Institution of Mechanical Engineers, Part H, 2003, 217, 127- 135.
[23] Cristofolini, L. and Viceconti, M., Development and validation of a technique for strain
measurement inside polymethylmethacrylate. Journal of Strain Analysis for Engineering
Design, 2000, 35:1, 21–33.
[24] Murphy, B. P. and Prendergast, P. J., On the magnitude and variability of fatigue
strength in acrylic bone cement. International Journal of Fatigue, 2000, 22, 855–864.
[25] Murphy, B. P. and Prendergast, P. J., The relationship between stress, porosity, and
nonlinear damage accumulation in acrylic bone cement. Journal of Biomedical Materials
Research, 2001, 59, 646–654.
[26] Lennon, A. B. and Prendergast, P. J., Residual stress due to curing can initiate damage in
porous bone cement: experimental and theoretical evidence. Journal of Biomechanics,
2001, 35, 311–321.
[27] Maher, S. A. and Prendergast, P. J., Discriminating cemented hip prostheses based on
migration and inducible displacement. Journal of Biomechanics, 2002, 35, 257–265.
[28] Mann, K. A., Bartel, D. L., Wright, T. W., Ingraffea, A. R., Mechanical characteristics
of the stem-cement interface. Journal of Orthopaedic Research, 1991, 9, 798–808.
List of Captions Tables:
Table 1: Material properties used in the finite element model.
Table 2: Mean sum-of-crack-lengths for pre-test, post-test, and growth during testing
for medial and lateral observable cement for each group. Significance values are
for a one tail Student’s t-test at 95% confidence.
Figures:
Figure 1: (a) Schematic of assembled model, including lever used to apply joint
reaction force and muscle loading and clamp for attachment to material testing
machine, (b), photograph of assembled rig mounted on materials testing
machine, and (c) mesh used for finite element analysis.
Figure 2: Spatial distribution of observed cracks for all specimens and mean sum-of-
crack-lengths (95% confidence) for each specimen type before and after testing.
Figure 3: Maximum principal stress for cement exposed by windows in bonded and
debonded models for cases of completely poreless cement and cement with a
uniform distribution of pores (i.e. a pore at the centroid of each cement element).
Figure 4: Distribution of volume of observable cement experiencing discrete stress
ranges. Black indicates volumes stressed to above 5 MPa while the total height
of each bar indicates the total volume experiencing tensile stress—note that there
is a decrease in the overall volume experiencing tensile stress for the medial side
in spite of the increase in volumes experiencing higher stresses.
Figure 5: Growth in summed crack length vs pre- test summed crack length for (a)
lateral and (b) medial cement layers. Data labels refer to specimen no. within a
population sample (e. g. M1 = matt specimen no. 1). Two trend lines (lateral
polished and medial matt datasets) each correspond to the best fit with an outlier
removed (P3 and M5 respectively).
Figure 6: Inducible displacement of actuator head at end of test vs summed crack
length at end of test.
Figure 7: Failure index for medial and lateral cement-metal interfaces. The lateral
interface is longer and so extends farther on the graph.
Table 1: Material properties used in the finite element model.
Material E (GPa) ν Cement 2.4 0.33
Stem 210.0 0.33 Cancellous Bone 2.0 0.30
Aluminium 73.0 0.33
Table 2: Mean sum-of-crack-lengths for pre-test, post-test, and growth during testing
for medial and lateral observable cement for each group. Significance values are for a
one tail Student’s t-test at 95% confidence.
Matt Polished (mm)� ia (mm)� ia
p
Pre 9.933 ± 11.025 5.500 ± 2.330 0.24 Medial Post 23.459 ± 15.993 24.149 ± 12.266 0.47 Growth 13.527 ± 8.725 18.649 ± 9.958 0.24 Pre 17.841 ± 9.224 16.316 ± 16.229 0.44 Lateral Post 35.626 ± 16.769 61.537 ± 32.186 0.11 Growth 17.785 ± 8.370 45.221 ± 25.871 0.05
Lever
Clamp
12 o
15 o 55.9 mm
M J
A
20.2 o
(a) (b) (c)
Figure 1: (a) Schematic of assembled model, including lever used to apply joint
reaction force and muscle loading and clamp for attachment to material testing
machine, (b), photograph of assembled rig mounted on materials testing machine, and
(c) mesh used for finite element analysis.
Figure 2: Spatial distribution of observed cracks for all specimens and mean sum-of-
crack-lengths (95% confidence) for each specimen type before and after testing.
0 2.5 5 7.5 10 +< 0(MPa)
Porous Porous
No Pores No Pores
Bonded Debonded
Figure 3: Maximum principal stress for cement exposed by windows in bonded and
debonded models for cases of completely poreless cement and cement with a uniform
distribution of pores (i.e. a pore at the centroid of each cement element).
0
1000
2000
3000
4000
MedialBonded
MedialDebonded
LateralBonded
LateralDebonded
>3 MPa0-3 MPa
Figure 4: Distribution of volume of observable cement experiencing discrete stress
ranges. Black indicates volumes stressed to above 5 MPa while the total height of
each bar indicates the total volume experiencing tensile stress—note that there is a
decrease in the overall volume experiencing tensile stress for the medial side in spite
of the increase in volumes experiencing higher stresses.
.
20406080100
1020
3040
50Pr
e-te
st c
rack
leng
th (m
m)
Growth (mm)
Mat
t (de
note
d M
1-M
5)
Polis
hed
(den
oted
P1-
P5)
P1
P2P3
P4
P5
M1
M4
M3
M2
M5
00
(a) L
ater
al
20406080100
Growth (mm)
Mat
t (de
note
d M
1-M
5)
Polis
hed
(den
oted
P1-
P5)
P1
P2P3
P4P5
M1
M4
M3
M2
M5
00
1020
3040
50Pr
e-te
st c
rack
leng
th (m
m)
(b) M
edia
l
Figu
re 5
: Gro
wth
in su
mm
ed c
rack
leng
th v
s pre
- tes
t sum
med
cra
ck le
ngth
for (
a) la
tera
l and
(b) m
edia
l cem
ent l
ayer
s. D
ata
labe
ls re
fer t
o
spec
imen
no.
with
in a
pop
ulat
ion
sam
ple
(e. g
. M1
= m
att s
peci
men
no.
1).
Two
trend
line
s (la
tera
l pol
ishe
d an
d m
edia
l mat
t dat
aset
s) e
ach
corr
espo
nd to
the
best
fit w
ith a
n ou
tlier
rem
oved
(P3
and
M5
resp
ectiv
ely)
.
R2 =
0.7
2
R2 =
0.6
7R
2 = 0
.98
R2 =
0.9
4
R2 = 0.7375
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 50 100 150 200Damage (mm)
Indu
cibl
e di
spla
cem
ent (
mm
)
Figure 6: Inducible displacement of actuator head at end of test vs summed crack
length at end of test.