Current
Principles of Design for Cemented and Cementless
Knees
David F.
Bindelglass, M.D., Jondy L. Cohen, M.D.,
and Lawrence
Dorr, M.D.
Twenty years of experience with modern knee arthroplasty
have shown certain designs of prosthesis to be extremely
successful. Modifications have been made to earlier designs in
an attempt to improve fixation and decrease the potential for
excessive wear, but new problems sometimes develop. Based on
this experience, certain principles become evident which
should guide surgeon/designers in the development of the next
generation of knee replacement implants.
Total knee arthroplasty is at the present time probably the
most durable of all joint replacements that can be performed.
14 Lessons learned over past 20 years have refined surgical
techniques and design. It is interesting to note the
similarity between Charnley's early results with total hip
replacement and those of Insall with the total condylar knee.
In both instances the results with the pioneer design have not
been exceeded by later designs. This is true because the early
designs were based on the fundamental principles necessary for
successful arthroplasty and because patient selection in early
studies was limited to older people. Throughout the 1980s,
surgical technique and design were modified in an attempt to
improve on those already excellent results.
A number of technical issues now have general acceptance.
Overall limb alignment must be in valgus so that medial
compartment bone is not overloaded.22 Minimal thickness bone
cuts on the tibia are desirable so that optimal bone stock
remains to support the tibial component.11 Lastly, joint line
height should be preserved within 1 cm for proper
patellofemoral mechanics15 and ideally the patella should
track centrally without tilting.
While technical points are
agreed upon by most surgeons, questions of design and fixation
remain. Recent designs have improved the patellofemoral
articulation and tracking but the optimal kinematics of the
tibial-femoral articulation are still debated. This article
will outline successful and unsuccessful design features of
implants that have been used. Design and fixation will also be
discussed in relationship to cost, which is a significant
factor in the decade of the nineties.
ARTICULAR
SURFACE DESIGN
The design of the femur is significant for its role both in
the femoral-tibial and patellofemoral articulations. At the
femoral-tibial articulation, controversy remains as to whether
or not to sacrifice the posterior cruciate ligament (PCL).
Early designs were cruciate sacrificing. Conformity between
the components provided anteroposterior (AP) stability and the
femur rolled on the tibia during flexion. Maximum flexion
achieved in this fashion averaged 90°, but this flexion was a
function of both design and the use of a 90° AP bone cut.35
PCL-retaining designs were developed for increased stability
but also to recreate femoral rollback, the natural posterior
sliding of the femur on the tibia.18 The goal was to gain
increased flexion (Fig 1). Without femoral rollback, the only
mechanism for increasing flexion is for the anterior femur to
lift off the tibia.16 Current PCL-sacrificing or substituting
designs demonstrate increased flexion by promoting rollback
within the design. Results of the two general designs appear
to be comparable in the first decade of fup.6 Ewald summarized
the projected long-term differences in 1990.13 The
PCL-sacrificing knees would rely on conformity between the
components for stability. This degree of constrain would lead
to increased interface shear and eventual loosening.
PCL-retaining knees relied on a sliding motion for flexion and
used flat polyethylene surfaces. Both factors led to increased
wear. Thus each design had weaknesses. One question must be
whether these weaknesses can be compensated for by combining
the best features of both designs into a common polyethylene
articular surface.
The design of PCL-retaining prostheses has been shaped by
the principle of 'kinematic conflict.'20 If the PCL is
retained the articular surfaces must be flat so that the femur
will slide more easily over the tibia. If the tibial surface
is dished then the femur will impinge on the posterior lip,
restricting the desired rollback (Fig 2). This impingement
will also generate tensile and shear forces at the tibial
component-bone interface. Therefore, most modern PCL-retaining
devices have flat surfaces,32 or at least large radii of
curvature, when viewed in the sagittal plane. Two major
problems occur with flat surfaces. The contact area between
the femur and plastic is smaller so that higher stresses are
generated, resulting in accelerated wear.39 Second, at the
extremes of motion the femoral component may be contacting
only the edge of the tibia. This 'edge loading' concentrates
high stress over a small area, which leads to wear and causes
tension at the tibial interface of the opposite condyle. This
will result in earlier loosening (Fig 3).
The authors believe that a compromise is possible.
Conforming surfaces can be constructed so that posterior
femoral displacement can occur with a dished tibia and minimal
sliding. The center of rotation of the tibial dish is moved 3
mm posteriorly from the midline (Fig 4). As the femur rotates
into flexion it is displaced posteriorly. This creates femoral
rollback with minimal sliding in a prosthesis that is more
conforming. Conformity leads to higher contact areas and makes
edge loading less likely. Wear is therefore reduced.
Femoral design can also affect the contact area of
articulation. In the total condylar knee the femoral condyles
are round when viewed from the front and conform fairly
closely with the tibial articular surface. The intercondylar
eminence, which is concave on both sides, articulates with the
central surfaces of both condyles, providing stability in the
mediolateral plane and increasing contact areas (Fig 5).38 The
tibial spines produce a similar effect in the normal knee.
Once criticism of the total condylar knee was that with
compressive loads the conformity of the components provided
resistance to the tibial rotation that occurs during normal
walking.34 Werner and colleagues were concerned that the
conformity necessary to resist rotational forces would lead to
the excellent long-term results of the total condylar knee
suggest that this is not the case.35 The explanation for this
difference between theory and clinical results is found in
gait studies by us that demonstrate that normal tibial
rotation does not occur during walking.11
In the coronal plane the authors feel that at least the
central surface of the condyles should be rounded to
articulate with a rounded surface of the tibial intercondylar
eminence, increasing contact area and stability against
mediolateral shear. Perhaps round surfaces both sagitally and
coronally will ultimately return, but presently the femoral
design must be rounded at least on the mediolateral surface to
maximize contact areas and minimize wear. Conformity between
the femur and tibia must be such that contact forces in both
extension and flexion are less than 18 megpasq/mm.8 Buechel
prefers contact forces of less than 10 megpasq/mm6 without
question, the femoral component design is critically important
in minimizing wear, which has become the primary cause of
failure of total knee replacement. the total condylar knee has
yet to show any reported failure rate from wear,35 whereas
wear has been a serious problem for the Porous Coated Anatomic
(PCA) knee, which is prototype of the flat femur/flat tibia
design.
TIBIAL DESIGN
The design principles for the tibial articular surface have
been partially described above. The articular surface is
curved both in the anteroposterior and mediolateral planes to
provide partial conformity and therefore increased contact
area without being overly constrained. In the sagittal plane
the center of rotation is moved 3 mm posterior to the midline
to accommodate posterior femoral displacement in flexion
(femoral rollback) with minimal sliding. In addition, the
central notch is deepened to minimize impingement of the PCL
in the notch. The eminence of the plastic provides a 5°
hyperextension stop.
Cemented design
For cement fixation the authors feel that an
all-polyethylene tibia is as good as, if not better than, a
metal-reinforced tibial tray. Bartel suggests that a
metal-backed tibial component would distribute forces more
evenly throughout the tibial surface, especially with
asymmetric loading, so that cancellous bone would be
protected.2 However, in a properly aligned knee, when forces
are distributed more equally, metal backing is not necessary.
Clinical results confirm the suggestion that all polyethylene
tibias may have equal or greater durability.1,21 L'Insalata,
Stern and Insall reported equally good results in two groups
of elderly patients, one with all polyethylene tibias and one
with metal-backed components.21 Apel, Tozzi, and Dorr
similarly found that there was no difference between
metal-backed and all-poly components in their series of 131
knees.1 Interestingly, this clinical history parallels the
evolution of cemented acetabular components. Charnley first
popularized cemented acetabular components. Charnley first
popularized cemented polyethylene acetabular prostheses.
Metal-backed acetabuli were developed because of the theory
that they would distribute stress better and reduce the
incidence of loosening. However, clinical studies have clearly
shown that a cemented all-polyethylene acetabulum is superior
to a metal-backed one.28
Three reasons are apparent for
the possible superiority of cemented all-poly tibial
components. One is the thickness of the plastic, which should
be 8 mm. Burstein has shown that the longevity of polyethylene
doubles when the thickness is increased from 6 mm to 8 mm.8
This, therefore, could be the difference between a 10-year and
20-year total knee replacement result. Metal trays are about 3
mm thick. This means that, at a minimum, the thickest tibial
component that should be used with a metal tray is 11 mm. This
thickness requires excessive bone resection in a routine,
primary knee.
A second reason to use an all-poly tibia is
the teeter-totter effect of a metal tray (Fig 6). When one
side of a metal tray begins to loosen, this teeter-totter
effect with progress rapidly by creating tensile forces at the
cement-bone interface on the opposite side and loosening the
stem (The fulcrum). All-poly components do not teeter around
the stem, and, in fact, if the plateau fixation loosens the
stem may remain intact for a long period of time. A third
reason to use an all-poly tibia is cost. A metal tray
increases cost by a factor of 2, and a factor of 3 if the
metal is porous coated. If all-polyethylene tibial components
were to replace metal-reinforced ones for almost all cemented
fixations, the saving would be $150 million.12 This is a
significant figure in the face of fixed fees being paid to
hospitals for patient care. Metal-reinforced tibial
components are still favored when a severe bony defect is
present and the choice is made to use a metal wedge to manage
a defect when the tibial component thickness is 15 mm or more,
metal reinforcement is superior because of reduced stiffness
of the implant. An all-poly implant of 15 mm is stiffer than a
15 mm metal-reinforced implant.
Another factor in tibial component design is symmetry. The
Natural Knee (Intermedics Orthopedics, Austin, Texas)
introduced an asymmetric tibial component to gain better
coverage of the tibial plateau. Barter argued that success of
metal-backed tibias depended on support by the cortical rim
with the metal tray acting as a bridge over the weaker
cancellous bone.2 Therefore, maximum coverage of the proximal
tibia, which is only possible with asymmetric right- and
left-sided components, would maximize support from the
cortical shell. As asymmetric design provides its greatest
benefits in noncemented fixation. For cemented fixation Walker
has shown that 85% coverage is acceptable.37 This can easily
be achieved with a symmetric component.
Cementless design
Bone ingrowth into the tibial component remains a challenge
to surgeon/designers. For bone ingrowth a porous surface on
metal is required. Optimal fixation by bony ingrowth requires
early stabilization to prevent motion and adequate surface
area for stable ingrowth fixation. Prostheses that fit into
bone, as into the femur and acetabulum of a total hip
prosthesis, have the best record of bone ingrowth on
retrieval. With total knee replacement, the most consistent
bone ingrowth has been on the PCA porous coated pegs that are
driven into the proximal tibia.26 Current tibial designs rely
on porous surfaces that sit on the bone and achieve initial
stability and lift-off protection by screw fixation. Data from
Volz36 and Whitesides41 support the use of screws for initial
fixation. However, some micromotion is inevitable,4,9 which
can lead to fretting at the screw plate junction, even if the
component eventually becomes well fixed. Polishing at this
junction has been observed by the authors in well-fixed
components, which raises the possibility that particulate
metal debris is being released in the joint. Peters, Engh, and
Dwyer have observed osteolysis along the tracks of the
screws.25
A recent study confirmed many surgeons' findings
that it is nearly impossible to create a perfectly flat tibial
bone cut. Thus, a rigid metal plate is placed on an uneven
surface. Loading subjects it to rocking on 'peaks' of the
proximal tibia, possible uneven settling, and uneven final
surface contact. Sulberg and Manley have shown how fluid is
pumped in and out from underneath the component with loading
when any differential motion between metal and bone is
present.30 The final mechanism for motion between the tibia
and a metal tray is that the tibia spreads radially 100
microns when loaded.8 Therefore, motion at the interface
results from applying a stiff flat metal plate to a more
flexible, if somewhat uneven surface. Cement evens out the
tibial surface and prevents motion. For noncemented fixation
there must be a way to prevent this motion so that ingrowth
can occur and later loosening is prevented.
One solution to lift-off of the tray has been the use of
pegs and screws. To protect against breakage of the tray in
case fixation is better on one condyle (causing bending of the
tray), a stem is used. A long stem also gives lift-off
protection. Another solution for the problem of a solid metal
tray is to avoid a one-piece metal plate. This is the
rationale behind the bicondylar tibia design, which is two
condylar plates connected by a solid tibial polyethylene piece
(Fig 7). Because flexible polyethylene is the bridge between
the metal trays, compression on one plateau will not generate
harmful tensile forces on the opposite one. A second feature
of the bicondylar design is the 5 mm-deep boss, which is
porous coated and press-fit into the bone. This box gives a
large surface area of porous coating for fixation that is
'into bone.' It also gives lift-off protection to the condylar
plate. The boss is press-fit into an undersized cylinder
reamed into each tibial plateau. Each condyle has four smooth
sharp pegs that provide rotational and shear stability. A
single-piece plastic plateau snaps into both condyles. The
single piece polyethylene connector gives the construct the
advantages of a more flexible all-polyethylene implant,
negating the teeter-totter effect while maintaining titanium
surfaces for bone ingrowth. The condylar pieces are well fixed
at surgery by the pegs and the press-fit of the box. The box
provides a relatively large surface area of porous coating
within the bone, greater than the porous coated pegs of the
PCA, which have achieved good ingrowth.26 One potential
drawback of the bosses is that they require reaming away good
proximal tibial bone to implant them. When revision becomes
necessary the tibial surface is usually recut at a depth of a
few millimeters.34 This would negate the defect reamed away
for the boss.
THE PATELLOFEMORAL JOINT
Patellar problems represent the greatest source of
complication of modern total knee replacement.5,17,24,27
Modern design must attempt to improve the results with
cemented patellae and especially with noncemented patellae,
whicih generally have had poor results.
The authors' experience indicates that domed patellae,
while producing good clinical results, have a high incidence
of radiolucencies. In our series 21% of domed patellae
demonstrate at least partial radiolucencies. In addition, the
deformation and wear of all-polyethylene domed replacements is
well documented.19 A conforming design provides larger contact
areas with less wear and deformation. The additional
constraint created by conformity of the patellofemoral joint
has not been shown to lead to increased patellar loosening.
Therefore, the recommended patella is a conforming
all-polyethylene implant.
While a cementless patella would be a desirable option, the
catastrophic failure of metal-backed patellae is well
documented.3,31 Wear-through of the polyethylene at the
periphery leading to metal-on-metal apposition, dissociation
of the polyethylene from the metal backing, and fracture of
the polyethylene at the edge of the metal skeleton have been
reported. The problems of metal-backed patellae are yet to be
solved. High forces are generated at the periphery where the
polyethylene is thin and subject to compression and shear. In
our series the percentage of patella replacements with tilting
or slight subluxation is high, approaching 40%. Each of these
tilted patellae have increased peripheral stresses. The
combination of tilt and subluxation of the patella, even with
successful, painless arthroplasties, does not bode well for
the durability of currently used metal-backed patellae.
* *
*
The last 20 years have provided valuable information as to
what is and is not successful in the design of knee
replacement. The superior results of early total knee designs
have led us to conclude that a number of features, namely
congruent articular geometry, thick tibial plastic, and
all-plastic patellae represent the basis for superior designs
available today. For cemented fixation a symmetric
all-polyethylene tibia gives less expensive and equally good,
if not superior, results. For cementless fixation a more
flexible bicondylar design with into-bone fixation may
represent the basic principles for future research to provide
better bone ingrowth and reduce loosening. Finally,
all-polyethylene and cemented patella is recommended.
Return to Dr. Cohen's Main Page
Return to Top |
