Pre-operative Surgical Simulation of an Acetabular Press Fit:
Assembly Strains in Bone
Branislav Jaramaz, Ph.D., Anthony M. DiGioia, III, M.D.
Center for Orthopaedic Research, Shadyside Medical Center, Pittsburgh, PA
and
Center for Medical Robotics and Computer Assisted Surgery,
Robotics Institute, Carnegie Mellon University, Pittsburgh, PA
ABSTRACT
An idealized axisymmetric finite element model of the
acetabulum and acetabular implant was analyzed. The
assembly strains induced in the bone by the press fit
insertion of an oversized cementless acetabular implant were
examined. The interaction between implant and bone was
modeled as a nonlinear contact-coupled interface to simulate
the actual process of press fitting the oversized implant into
the prepared acetabular bone bed.
INTRODUCTION
Initial implant stability of the cementless acetabular
implant is commonly achieved by press fitting oversized
acetabular components. Implicit in this concept is
"pre-straining" of the bone which must occur at the time of
implantation. Assembly strains are expected to be very
different in both magnitude and direction from the strains
induced by normal joint forces, and can greatly change the
mechanical environment and load transfer patterns.
The mechanical consequences of an interference fit and
the development of these assembly strains in bone are critical
to our understanding of both the biologic response of bone to
mechanical stimuli and the way in which the bone/implant
construct may be modeled. The purpose of this study was to
determine the effect of the amount of press fit and the size of
the implant on the magnitude of resultant assembly strains.
METHODS
Axisymmetric nonlinear, contact-coupled models of a
cementless acetabular system were developed to examine the
complex interaction between the bone and implant at the
time of implantation. The geometry was parametrically
described to accommodate for different sizes of the
acetabulum and the implant. The resulting finite element
mesh was constructed of up to 958 quadrilateral axisymmetric
solid elements for the bone and up to 201 solid elements for
the implant. The computational analysis was performed on a
DECstation 5000/240 using the software package ANSYS
5.0 (Swanson Analysis Systems, Inc., Houston, PA). The
finite element code allows for a solution of generally
formulated contact between two deformable bodies. This
type of formulation permitted the simulation of implanting
an acetabular component in a manner that mimics the actual
surgery. The implant is initially positioned above the
acetabulum and incrementally driven into prepared
hemispherical cavity. In its final position, the implant is
held in place only by contact with the bone. The forces
which tend to eject the implant are equilibrated only by the
development of friction at the contact surface.
The material model assumed for bone was elastic-
perfectly plastic. Material properties were assigned to bone
as: for cortical bone, modulus of elasticity E = 16 GPa and
yield stress σY= 175 MPa; for cancellous bone E = 1.0 GPa
and σY= 5 MPa. The implant was constructed of a titanium
shell with overlying titanium mesh. Geometries were
detailed from the Harris-Galante II (HGP II) acetabular
implant. The titanium component was modeled as a linear
elastic material (E=110 GPA for the shell and E=55 GPa for
the mesh). The Poisson‘s ratio was 0.3 for all materials. A
value of 0.48 was used for the coefficient of friction at the
contact between the bone and the implant. Using this model,
the effect of parameters such as the amount of press fit, the
implant geometry and the coefficient of friction were
examined. The amount of implant oversizing (or nominal
press fit) was varied from 1, 2, and 4 millimeters for bone
size of 40, 50, and 60 millimeters. Resultant strains in the
bone were calculated for each case.
RESULTS AND DISCUSSION
The total equivalent strains at the final position (after the
constraints are fully released on the implant) are presented for
several cases. These are the strains induced in bone prior to
the application of joint loads. Figure 1 shows the results for
the same size bone cavity (60 mm) and three different
amounts of implant oversizing (1, 2 and 4 mm). Figure 2
shows the effect of the same amount of oversizing (2 mm)
on several different sizes of the bone and implant. In all cases
the strains predicted by the finite element model were much
higher than the yield stress in cancellous bone, and in many
cases higher than the yield stress in the cortical shell.
All cases showed similar behavior and could not fully
bottom out, i.e., the contact between the implant and the
bone could not be established over the entire surface. Instead,
at the implant’s lowest position, with the implant still held
by applied impact forces, the contact was achieved over
approximately 1/3 to 1/2 of its arch length, and as expected,
was at the periphery of the acetabulum. After the forces on
the implant were released, the contact zone was reduced to 1/5
to 1/4 of the arch due to elastic recoil of the implant/bone
system. These results are in very good agreement with
recently reported experimental results by McKenzie et al. [2].
They noted that only peripheral contact was achieved in all
tested cases and that the contact area does not exceed 20% of
all area potential for contact. Although the results of this
