Biomedical Engineering Reference
In-Depth Information
Computational simulations can also be used to predict tissue behavior in
response to biomechanical factors.
Nowadays, finite element models are normally used to design implants, analyzing
the influence of different factors such as implant variables, like size, shape, position,
material stiffness, coating and ingrowth conditions, debonding, or lack of fixation;
bone variables, such as geometry, bone density and anisotropy distributions, and
response to different loading conditions. In this section, we perform a brief revision
of some examples of the application of bone remodeling models to analyze the
effect of different design factors in implants.
One of the first works in which a computational theory of bone remodeling
was applied to study the influence of a prosthesis implantation was developed by
Huiskes
et al
. [48]. Although 2D and axisymmetric models were used, the isotropic
theory proposed by Huiskes in this work was used to simulate bone response
(elastic modulus adaptation) due to prosthesis implantation. In fact, Huiskes and
coworkers carried out the most extensive development of these finite element
simulations combined with bone remodeling theories. For example, Huiskes
et al
.
[61] studied the dependence of the stress-shielding on the stiffness of the stem
analyzing a 3D model of a titanium stem (100 GPa) versus a more flexible stem
(20 GPa). They used a 3D FE mesh where the initial apparent density distribution
was determined from CT-density values; subsequently, they used an isotropic
bone remodeling theory to predict the apparent density distribution after total
hip replacement (THR). They concluded that stiffer stems produce more bone
resorption than flexible ones. This conclusion has also been shown in animal
experiments [62, 63] and in clinical radiographic studies [64].
Weinans
et al
. [65] used 3D FEA for studying the effect of bonded non-cemented
total hip arthroplasties in dogs, comparing computational and experimental results
through cross-sectional measurements of the canine femurs after two years of
follow-up. This comparison showed that long-term changes in the bone around
femoral components of THRs can be fully explained with bone remodeling theories.
Weinans
et al
. [66] also analyzed non-cemented prosthesis and various situations
of prosthesis-bone bonding with finite elements, using apparent density as the
variable that quantifies bone remodeling. They checked the influence of the coating
conditions (fully, partial, or noncoated) and the fitting characteristics (press fitted
or overreamed), concluding that partial coating can reduce bone atrophy relative to
fully coated stems. For smooth press-fit stems, computer predictions showed that
the amount of bone loss was lower than that for one-third proximally coated or
fully coated stems.
van Rietbergen
et al
. [67] also performed 3D finite element simulations for
press-fit prostheses implanted into the femurs of dogs, for which CT scan ge-
ometric data were available. This model included not only internal remodeling
by modification of apparent density but also external remodeling by modifying
the boundary geometry. The results of this work confirmed the value of such
approaches for preclinical testing of implants.
Kuiper and Huiskes [68] predicted the eventual loss of bone mass from the
initial patterns of elastic energy deviation around hip replacements, combining
