|
Presented at NACOB 98:
North American Congress
on Biomechanics
Canadian Society for Biomechanics -
American Society of Biomechanics
University of Waterloo
Waterloo, Ontario, Canada
August 14-18, 1998
|
|
AN ADAPTATION SIMULATION TO PREDICT BONE
REMODELING AROUND
IMPLANT STEMS FOLLOWING HIP
REPLACEMENT SURGERY
S.J. Hazelwood(1), R.B. Martin(1), J.J.
Rodrigo(2), and M.M. Rashid(3)
(1) Orthopaedic Research Labs, University of
California Davis, Sacramento, CA 95817
(2) Dept. of Orthopaedic Surgery, UC Davis Medical
Center, Sacramento, CA 95817
(3) Dept. of Civil and Environmental Engineering, UC
Davis, Davis, CA 95616
INTRODUCTION
Hip replacement surgery is one of the most common and
successful alternatives for patients with degenerative
diseases affecting the hip joint. One concern with
this surgery, though, is the possibility that the
inserted prosthesis alters the mechanical environment
of the host femur, leading to remodeling and cortical
bone resorption around the implant stem. Although
the consequences of this bone loss are not well known,
it is believed that remodeling may play a role in the
long-term success of the surgery. In addition,
revision surgery for loose or failed prostheses is
becoming more common as recipients and their implants
age. Cortical bone loss in the diaphysis may limit
the available surgical options if a revision is
necessary.
A bone adaptation simulation based on disuse and
damage repair was utilized to investigate remodeling
around implant stems. This simulation accounted for
the cellular response of bone to mechanical stimuli
and also for the temporal sequence of bone remodeling
by basic multicellular units (BMUs). The implants
used for this study were a conventional, long-stem,
press-fit prosthesis and a surface replacement.
REVIEW AND THEORY
A modified version of the model of Hazelwood et al.
(1997) was used for this study. The rate at which
damage was formed in bone (D'F) was assumed to be
related to the loading rate (RL) and the strain range
(s) from a mixture of n daily activities:
D'F =
kD*(RL1*s1**q+RL2*s2**q+...+RLi*si**q+...+RLn*sn**q) =
kD*phiD
where the damage coefficient, kD, and the exponent q
were estimated to be 40.2 and 2.89, respectively. The
quantity phiD is defined here as the damage potential.
Minimum principal strain was used as the predictor (s)
for damage formation. The assumed rate of damage
repair (D'R) was:
D'R = D*fa*A*Fs
where D is the existing damage, fa is the BMU
activation frequency, and Fs is a damage repair
specificity factor (set to 5 for this study). The
area, A, was assumed to be equivalent to the area of
an osteonal cross-section (0.095mm radius) for
cortical bone or half that value for trabecular bone.
The damage coefficient was chosen so a damage
potential value of 0.000000865cpd produced the
condition where bone formation and resorption were in
equilibrium. Disuse was defined as values of phiD
below this equilibrium damage potential value.
The daily BMU activation frequency was assumed to be a
function of disuse, the existing damage, and the
internal surface area available for bone remodeling.
The number of resorbing and refilling BMUs active each
day were calculated from the activation frequency
history over the remodeling period: 25 days for
resorption, 5 days for reversal, and 64 days for
refilling. Daily porosity changes were then
calculated at the integration points based on the net
amount of bone removed or added by each resorbing or
refilling BMU, respectively. The area for a haversian
canal (0.020mm radius) was accounted for when
predicting bone formation in cortical regions.
Elastic modulus and porosity relationships were
determined from data by Currey (1988) for cortical
bone and Rho et al. (1993) for trabecular bone.
PROCEDURES
A two-dimensional finite element model (linearly
elastic, isotropic), consisting of 4216 4-node
quadrilateral elements, was created from a radiograph
of a representative femur. A bony side plate (Weinans
et al., 1992) was added to the model to account for
the out-of-plane cortical bone. An initial porosity
of 5% and Poisson's ratio of 0.3 were assumed for all
elements and the porosity of the bony side plate was
kept constant during the simulation. Three load
cases, each consisting of joint reaction and abductor
muscle forces, were used to simulate the daily loading
history for normal activity (Carter et al., 1989).
For this simulation, the first load condition
(single-leg stance) was applied for 3000cpd while the
second and third conditions (simulating abduction and
adduction) were applied for 500cpd.
An intact (unimplanted) femur model simulation was run
using ABAQUS 5.6 (HKS, Pawtucket, RI) until the
porosity distribution achieved steady state. The bone
adaptation algorithm was integrated into the analysis
through a UMAT subroutine. Using these results as the
preoperative condition, implants (modulus of
210,000MPa and Poisson's ratio of 0.33) were
introduced and the analysis was continued for an
additional 1200 days of simulated remodeling.
RESULTS AND DISCUSSION
Predicted porosity distribution contour plots for (a)
the preoperative condition and at 1200 days
postoperative for simulations with (b) a conventional
implant and (c) a surface replacement are shown in Figure 1. In addition, Figures
1b and 1c indicate the percent changes in porosity
between the preoperative and postoperative results in
six regions of the femur. For the model with the
conventional implant, bone loss was observed in all
regions surrounding the implant, with the largest loss
(43.7%) exhibited in the calcar region of the inferior
neck. Percent bone loss values were consistent with
results from clinical measurements (Engh et al., 1993
and Kilgus et al., 1993). The zone exhibiting the
lowest bone loss (2.4%) was in the lateral region of
the proximal femur and possibly resulted from the use
of identical muscle forces for both the preoperative
and postoperative models. BMU activation was observed
to increase following insertion of the conventional
implant, especially in the medial region of the
proximal femur and along the lateral cortex. Bone
loss values with the conventional implant were
substantially larger than those with the surface
replacement, which produced a porosity distribution
similar to the preoperative condition When the
surface replacement was modeled, BMU activation
increased directly under the implant head and in the
calcar region. Damage for both models was generally
higher in the cortices than in the trabecular areas.
Although these results indicate that bone loss
following hip replacement surgery may be reduced with
the use of surface replacements, many complications
still exist with their implantation.
REFERENCES
Carter et al. J Biomech, 22:231, 1989.
Currey J Biomech, 21:131, 1988.
Engh et al. Clin Orthop, 292:177, 1993.
Hazelwood et al. Proc 21st ASB, 256, 1997.
Kilgus et al. J Bone Joint Surg, 75B:279, 1993.
Rho et al. J Biomech, 26:111, 1993.
Weinans et al. J Orthop Res, 10:845, 1992.
ACKNOWLEDGMENTS
Simulations were run in part on computer facilities
provided by a gift from Miss Lorna Talbot to the VORL,
University of California, Davis.
