Abstract
Background: With extensive use of posterior
stabilized total knee arthroplasty implants, it is increasingly
important to assess the mechanical performance of this design alternative.
The purpose of this study was to examine the wear patterns at the
femoral cam-tibial post interface in a series of posterior stabilized prostheses
retrieved at revision arthroplasty.
Methods: Qualitative and quantitative wear analysis
was performed over the surface of the stabilizing posts from twenty-three
retrieved total knee components that had been implanted for a mean
of 35.6 months (range, 2.3 to 107.2 months). The implants were designs
from four different manufacturers. Digital images of the anterior,
posterior, medial, and lateral surfaces of the tibial post were
made for quantitative analysis and determination of a post wear score.
Wear was characterized with a grading system that isolates adhesive,
abrasive, and fatigue wear, inferring a weighted score from an estimation of
generated polyethylene debris.
Results: Evidence of wear or damage was observed
on all twenty-three of the stabilizing posts, including those revised
because of infection. On the average, 39.9% (range, 18.5% to
60%) of the post surface demonstrated some form of deformation,
with adhesive wear, or burnishing, being the predominant wear mechanism.
Seven posts (30%) exhibited severe damage with gross loss
of polyethylene. The wear caused premature failure and early revision
of two components: one of these failures was related to isolated
post wear and the other, to severe post wear and subsequent fracture.
Overall, wear was primarily posterior, but wear over the anterior, medial,
and lateral surfaces was also notable.
Conclusions: The cam-post articulation in posterior
stabilized implants can be an additional source of polyethylene
wear debris. The variability in wear patterns observed among designs
may be due to differences in cam-post mechanics, post location,
and post geometry. The surgeon should be aware that the cam-post
interface is not an innocuous articulation, and manufacturers should
be motivated to produce implants that maintain the function of the
post while limiting wear and surface damage.
Cruciate-substituting procedures with use of a posterior
stabilized implant are a popular treatment for patients requiring
primary total knee arthroplasty. Improvements in implant design,
a technically easier procedure in the face of deformity, and restoration
of knee kinematics may all be reasons for the increased use of this
design.
Once the posterior cruciate ligament is cut, options to prevent
anteroposterior instability are limited to the use of (1) a posterior
stabilized design with a tibial post and femoral cam, or (2) a dished
polyethylene insert with a raised anterior lip. There is considerable
controversy with regard to the indications for the use of these
two design options. The posterior stabilized implant is certainly
a proven design, with a higher than 95% survival rate reported
in long-term (greater than ten-year) follow-up studies involving
series of 165 to 2300 cases1-3.
Complications such as instability and so-called patellar clunk have
been identified but are apparently less likely to occur with use
of newer implants. Opponents of this design have voiced concern
that the stabilizing post acts like a bumper in preventing tibial
subluxation and may subsequently lead to substantial problems such
as excessive wear or component fracture, especially in younger,
more active patients.
Wear of the polyethylene insert is accepted by many as the primary
limiting factor in the longevity of current total joint prostheses,
and a substantial amount of information about the mechanical performance
of this component has been reported4-13.
Unfortunately, to our knowledge, specific wear analysis of retrieved
posterior stabilized implants, particularly of the cam-post articulation,
is currently lacking in the literature. With substantial heterogeneity
among designs, including variable post location and geometry, it
is increasingly important to appraise the performance of this articulation.
The current study is a qualitative and quantitative wear analysis
of the stabilizing posts from retrieved posterior stabilized total
knee components. It was our hypothesis that the tibial post may
be an additional source of polyethylene damage and wear debris.
Twenty-three posterior stabilizing tibial inserts from twenty-three
patients were retrieved at the time of revision over a two-year
period (1996 through 1998) by four arthroplasty surgeons at our center.
Components were collected, decontaminated, and catalogued with use
of a procedure developed by our implant retrieval laboratory. Implants
had been inserted over a period of seventeen years (1981 through
1998), with an average duration of implantation prior to retrieval
of 35.6 months (range, 2.3 to 107.2 months). Seven inserts had been
implanted as primary components, and sixteen had been inserted during
revision arthroplasty. The tibial inserts were modular cemented designs
from four different manufacturers: there were four Genesis I and
three Genesis II inserts (Smith and Nephew Richards, Memphis, Tennessee);
one Kinematic I, one Kinematic II, and five Kinemax inserts (Howmedica,
Rutherford, New Jersey); two AMK and five Coordinate inserts (DePuy,
Warsaw, Indiana); and one Insall-Burstein II and one CCK insert
(Zimmer, Warsaw, Indiana). Differences in designs by individual
manufacturers represent the evolution of a similar model type (Kinematic
I, Kinematic II, and Kinemax inserts and Genesis I and II inserts)
or options for increasing constraint (AMK and Coordinate inserts
and Insall-Burstein II and CCK inserts). Eighteen inserts were made
of standard ultra-high molecular weight polyethylene, and five were
manufactured from Hylamer (DePuy). Five of the Kinematic/Kinemax
components were heat-pressed. The polyethylene sterilization procedure
was determined for twenty-one of the twenty-three inserts; fourteen were
gamma-irradiated in air, five were sterilized with ethylene oxide,
and two were sterilized with gas plasma. The components were revised
for numerous reasons, including infection (seven), aseptic loosening
(six), instability (five), pain (two), osteolysis (one), stiffness
(one), and recurrent effusions (one). The primary underlying diagnosis
was osteoarthritis in twenty-two patients and rheumatoid arthritis
in one.
All twenty-three tibial inserts were examined for evidence of
surface wear over the stabilizing post. Whenever surfaces articulate
with each other, wear occurs; this can be seen as damage to one
or both surfaces and generally involves progressive loss of material14. An important concept that must
be kept in mind when one attempts to quantify the wear of polyethylene
is the difficulty in distinguishing wear, in which material is lost,
from plastic deformation (creep or cold flow), in which the polymer
is distorted in shape without loss of material. This can be especially
difficult when an attempt is made to differentiate adhesive wear
seen as burnishing, but it is easier when one is trying to identify
fatigue damage such as delamination and pitting. Although the use of
the term wear to describe all forms of surface
and subsurface deformation may be semantically imprecise, we will
continue to use it in this study, acknowledging that the damage
can occur without the subsequent generation of debris.
With use of gross visual examination and stereomicroscopy at
ten times magnification, a wear grade was assigned to each area
believed to have undergone deformation (Table I). The grading system is used in
an attempt to quantitatively evaluate polyethylene damage and generated
debris on the basis of one of three mechanisms of wear: adhesive,
abrasive, and fatigue wear12,14-16.
It assigns a grade of 1 to 5 on the basis of this mechanism as well
as the severity of the wear, thereby inferring a weighted score
through an estimation of the volume of generated polyethylene debris.
This system does not infer a grade on the basis of the potential
destructiveness of the particles produced. Although we acknowledge
that smaller particles produced from burnishing can be potentially
more destructive, we could not apply this concept in a macroscopic
setting and we chose to use a system based on the overall volume
of debris. Previously defined damage morphologies7,
with the term morphology describing what we see
(such as delamination, pitting, and abrasions), are also encompassed
by this system in order to facilitate comparison with other wear
studies. A digital camera (D-600L; Olympus Optical, Melville, New
York) was used to obtain images of the anterior, posterior, medial,
and lateral surfaces of the post. The worn or damaged surface area
was quantified, with use of image-analysis software (Scion Image;
Scion, Frederick, Maryland), by outlining the affected area with
use of edge-detection and manual tracing. The percentage of the
surface area affected was calculated from a measurement of the entire
area of each surface (Fig. 1). A wear score, calculated as the
sum of the product of the wear grade and the percent of the surface area
affected, was assigned to each of the four post surfaces. The wear
scores from each of the four surface regions were combined to determine
the total post wear score for the individual components. The maximum
score assigned to each region was 5, and the maximum total post
wear score was 20.
Although the purpose of the study was primarily to document wear
of the stabilizing post, analysis of variance was applied to the
wear scores to compare surface locations and prosthetic designs.
The Pearson product-moment correlation coefficient, comparing wear
score and period of implantation, was also determined. When comparisons
were made between prosthetic models, wear scores were normalized
for the period of implantation by dividing the total post wear score
by the period of implantation. Unless otherwise stated, all values
are reported as the mean and the standard deviation.
General examination revealed evidence of wear, in varying degrees,
over some portion of the post surface of all twenty-three retrieved
implants. All three mechanisms of wear occurred, and the morphology of
damage included pitting, delamination, scratching, abrasions, burnishing,
and surface deformation. There was no notable evidence of embedded cement
or metal debris. The mean area of wear, as a percentage of the post’s
total surface area, was 39.9% (range, 18.5% to
60%). The predominant mechanism was adhesive wear, or burnishing, which
occurred on 91% (twenty-one) of the twenty-three posts
and covered a mean of 24% (range, 0% to 48%)
of the entire post surface area. Fatigue wear, mainly delamination,
was evident on 35% (eight) of the twenty-three posts and
covered a mean of 11% (range, 0% to 64%)
of the overall surface area. Abrasive damage, primarily scratching, was
evident on 52% (twelve) of the twenty-three posts and covered
a mean of 4% (range, 0% to 19%) of the
surface area.
Severe wear, primarily in the form of delamination and involving
gross loss of polyethylene, was apparent on 30% (seven)
of the twenty-three posts. Three of these inserts had been heat-pressed,
all were standard ultra-high molecular weight polyethylene, and
all had been gamma-irradiated in air. Although we did not perform
quantitative wear analysis of the condylar articular surfaces of
the tibial insert in this study, gross visualization of these surfaces
did reveal morphologies of damage similar to those of the stabilizing
post. Certainly, delamination of the post surface seemed to correspond
with similar damage over the articular surface of five of the seven
severely worn inserts, two of which had been heat-pressed.
Conversely, one AMK insert (DePuy), which had been revised because
of pain after forty-five months of implantation, exhibited severe
wear over a portion of the anterior and posterior post surfaces
without evidence of similar wear of the articular surface. Intraoperative
records on the removal of this component described an inflammatory
cyst and generalized reactive synovitis with notable evidence of
polyethylene debris. The implant had not been heat-pressed and had
been gamma-irradiated in air.
Severe post wear influenced the structural integrity and the
capacity to prevent posterior subluxation of two of the twenty-three
implants. One of these implants, a revision Genesis I prosthesis
(Smith and Nephew Richards) that had been removed because of instability
at thirty-five months, had a fractured post that had dissociated
from the tibial platform. The polyethylene was split through a section
of the post that had been subjected to appreciable posterior wear
(Fig. 2).
The second insert, a Kinemax insert (Howmedica) that also had been
revised because of instability, had complete disintegration of 4.1
mm of the vertical height, or 33% of the superior apex,
of the post.
The quantitative post wear score was determined for twenty-two
of the twenty-three retrieved inserts (Table II). Wear analysis of one component
was not possible, as iatrogenic damage during removal was too extensive
for accurate assessment. The mean post wear score was 3.56 3.18
(range, 0.7 to 12.4), of a maximum of 20, after a mean period of
implantation of 35.7 months (range, 2.3 to 107.2 months). The Pearson
product-moment correlation coefficient (r) comparing the period
of implantation with the wear score was 0.61, suggesting an increasing wear
score with increasing duration of implantation (Fig. 3). With the numbers
available, no significant differences in normalized post wear scores
were noted among the design types. Similarly, no significant differences
in normalized post wear scores were noted between inserts manufactured
from Hylamer and those made of standard ultra-high molecular weight
polyethylene, or among sterilization procedures. The normalized
post wear scores for the heat-pressed inserts were appreciably higher
than those for the non-heat-pressed inserts, but no significance
could be determined with the small numbers available.
The post wear score averaged 1.54 ± 0.65
for the components revised because of infection, averaged 3.7 ± 1.48 for those revised because of aseptic loosening, was
12.4 for the component revised because of osteolysis, averaged 3.1 ± 1.96 for the components revised because of pain, averaged
4.5 ± 3.98 for those revised because of instability,
and was 7.3 for the component revised because of recurrent effusions.
It was noted that the wear scores and the wear patterns of the implants
revised because of infection were similar to those of the rest of
the group after normalization for the period of implantation. Fatigue
wear was not observed over the post of any of the implants revised
because of infection, but the mean duration of implantation was
relatively short (16.5 months; range, 2.3 to 31.3 months).
Wear occurred over all four surfaces of the post, and the prevalence
varied among implant designs (Fig. 4). The most predominant location
of wear was the posterior surface, which was worn on all twenty-two
of the inserts. The posterior surface also had the highest wear
scores, with a mean of 1.86 ± 1.53. The prevalence
of anterior wear was also substantial, occurring on 36% (eight)
of the twenty-two implants, with a mean wear score of 0.93 ± 1.56. The overall extent of anterior wear was skewed
by the high prevalence and severity of anterior wear of the Kinematic/Kinemax
(Howmedica) inserts. All six of these inserts had evidence of anterior
wear, and the normalized wear scores for the anterior surface were
significantly higher (p < 0.05) than those for the other
designs.
With the numbers available, no significant difference between
medial and lateral wear of the posts could be identified. The predominant
morphology of wear over the medial and lateral surfaces was burnishing,
with minimal evidence of fatigue. Posts with a relatively wider
medial-lateral dimension had increased evidence of damage over these
surfaces. This was the case with six implants, the five Coordinate
implants (DePuy) and the CCK implant (Zimmer), that had an ultra-constraining
post to resist varus-valgus deformity and to limit tibial rotation
to a few degrees. The normalized wear scores for the medial and
lateral surfaces of these constraining posts were substantially
higher than the normalized wear scores for the same regions of implants
of a similar model type. This was reflected by the relatively high
mean normalized wear scores in the two groups containing these implants
(Fig. 4).
In two of the implants in those groups, considerable wear of the
medial condylar articular surface appeared to have caused a mild
varus-valgus deformity that appeared to have accelerated wear over the
lateral surface of the post.
Wear analyses of the articular and inferior surfaces of tibial
inserts from retrieved total knee implants have been extensively
reported in the literature7,10-12.
Component design, in addition to polyethylene thickness, sterilization,
manufacturing procedure, and counterface roughness, is an important mechanical
factor that influences wear5,9-13.
Cruciate-substituting designs were developed to prevent uncontrolled
sliding of the femoral component without requiring excessive conformity, thereby
reducing polyethylene damage. The presence of an additional interface,
between the femoral cam and the tibial post, does raise concern
regarding the generation of additional polyethylene debris.
It is clear, both from knowledge of implant mechanics and from
examination of retrieved implants, that wear or damage occurs over
some portion of the post surface in almost all posterior stabilized
implants. Unfortunately, it is impossible to determine the influence
of this wear on the overall generation of polyethylene debris or
its contribution to implant failure. It would appear from this sample,
though, that the damage may not always be irrelevant or without
consequence. Given that, on the average, approximately 40% of
the post surface exhibited some form of wear, the post surface must be
considered a notable contact surface and a potential source of additional
debris. Although the majority of the wear appeared to be adhesive
in nature, other forms of damage can occur, releasing larger volumes
of potentially destructive particles4,6,15.
In this study, severe tibial post wear appeared to be associated
with negative outcomes known to be related to the accumulation of
polyethylene debris, such as osteolysis and recurrent effusions.
Of course, a direct association cannot be established without somehow
distinguishing the debris from this interface from that from the
tibial articular and inferior surfaces. In at least one patient,
however, isolated post wear with nominal wear of the articular surface
and the undersurface was believed to be related to the initiation
of a reactive synovitis that resulted in revision secondary to pain.
In addition, remarkable posterior wear and subsequent fracture of
the post resulted in gross instability and early revision of another
implant.
It was not unexpected that severe wear of the post coincided
with a similar severity of wear over the condylar articular surface
of the tibial insert in many of the implants. This might indicate
a predisposition of the polyethylene of these implants to fatigue
as a result of intrinsic factors such as oxidation, crystallinity,
or manufacturing method. It may also be secondary to operative factors
such as malalignment or instability of the components. In these cases,
the presence or absence of a stabilizing post may not affect the
inevitable or ultimate failure of the implant, and the observed
wear patterns would not be representative of a normally
functioning implant. Another important factor may have been that
70% (sixteen) of the twenty-three implants were inserted
during revision arthroplasty, which may mean that the knees with
those implants had less ligamentous stability or were less so-called flexion-extension-gap
balanced than the knees in which the implants were inserted during
primary arthroplasty.
Although we only analyzed implants that failed, we might extrapolate
similar wear patterns to those that do not fail, by isolating the
components that were retrieved for nonmechanical failure. In our
study, comparing implants retrieved because of infection with those
that failed because of mechanical reasons revealed similar wear
patterns and overall wear scores after normalization for the period
of implantation.
This study did not demonstrate appreciable differences in overall
wear scores among the prosthetic designs, but it did highlight variable
wear patterns among them. Because we know that the stabilizing bar
acts as a contact guide in limiting tibial subluxation and ensuring
appropriate femoral rollback, it is not surprising that polyethylene
wear occurred over the posterior surface of the post in all of the stabilized
implants in our series. The prevalence of anterior wear in the Kinematic/Kinemax
implants (Howmedica) is somewhat distinct and is most likely a factor
of the cam-post design. This implant has a more anteriorly positioned
post and was designed to resist hyperextension. The impingement
of the anterior surface of the femoral cam on the anterior surface
of the post at the end point of knee extension predisposes this
area to damage. Similarly, the distinct patterns of wear in other implants
may be due to the location and geometry of the cam-post complex.
The apparent increase in medial and lateral wear of the more constrained posts
is most likely secondary to reduced clearance between the medial
and lateral surfaces of the cam. If components are implanted in
a malaligned position or if asymmetrical wear of the condylar articular
surface occurs, the reduction in clearance might further predispose
the tibial post to accelerated wear over these surfaces.
The observation of the Genesis I implant (Smith and Nephew Richards)
with the fractured post also demonstrates the importance of post
design. It is our hypothesis that the relatively tall and thin geometry
of the post may have predisposed this implant to failure. The considerable
wear of the inferior aspect of the post combined with the forces of
a long lever arm during cam-post engagement probably contributed
to the fracture. Interestingly, this was an older design, and the
manufacturer has since revised its geometry by shortening the height (reducing
the levering potential) and increasing the thickness.
An obvious limitation of this study is the limited number of
retrievals and hence the small number of specific designs. This
study was not meant to reveal problems with specific implants but
rather was observational of the variability in the geometry and architecture
of the stabilizing post and, as such, provokes the question as to
whether these differences may somehow affect wear.
In conclusion, posterior stabilized implants may contribute to
the production of additional wear debris and hence may influence
the prevalence of negative outcomes such as osteolysis, aseptic
loosening, and reactive synovitis. The surgeon should be aware that
the cam-post interface is not a completely innocuous articulation,
and manufacturers should be motivated to produce implants that maintain
the function of the post while limiting wear or damage. Further
analysis of the entire articular surface of these implants, and
a comparison of these wear patterns with those of retrieved current-generation
cruciate-retaining designs, would be invaluable. In addition, evaluation
of ultra-high-conforming designs as an alternative to posterior stabilized
implants should be included in future wear analyses.
Colizza WA; Insall JN; and Scuderi GR: The posterior stabilized total knee prosthesis. Assessment
of polyethylene damage and osteolysis after a ten-year-minimum follow-up. J Bone Joint Surg Am,1995.77: 1713-20, 771713
1995
[PubMed]
Font-Rodriguez DE; Scuderi GR; and Insall JN: Survivorship of cemented total knee arthroplasty. Clin Orthop,1997.345: 79-86, 34579
1997
[PubMed]
Ritter MA; Campbell E; Faris PM; and Keating EM: Long-term survivor analysis of the posterior cruciate
condylar total knee arthroplasty. A 10-year evaluation. J Arthroplasty,1989.4: 293-6, 4293
1989
[PubMed]
Amstutz HC; Campbell P; Kossovsky N; and Clarke IC: Mechanism and clinical significance of wear debris-induced
osteolysis. Clin Orthop.,1992.276: 7-18, 2767
1992
[PubMed]
Blunn GW; Joshi AB; Minns RJ; Lidgren L; Lilley P; Ryd L; Engelbrecht E; and Walker PS: Wear in retrieved condylar knee arthroplasties. A comparison
of wear in different designs of 280 retrieved condylar knee prostheses. J Arthroplasty,1997.12: 281-90, 12281
1997
[PubMed]
Harris WH: The problem is osteolysis. Clin Orthop,1995.311: 46-53, 31146
1995
[PubMed]
Hood RW; Wright TM; and Burstein AH: Retrieval analysis of total knee prostheses: a method
and its application to 48 total condylar prostheses. J Biomed Mater Res,1983.17: 829-42, 17829
1983
[PubMed]
Landy MM, and Walker PS: Wear of ultra-high-molecular-weight polyethylene components
of 90 retrieved knee prostheses. J Arthroplasty,1988.3(Suppl): 73-85, 3(Suppl)73
1988
Lewis P; Rorabeck CH; Bourne RB; and Devane P: Posteromedial tibial polyethylene failure in total knee
replacements. Clin Orthop,1994.299: 11-7, 29911
1994
[PubMed]
Wasielewski RC; Galante JO; Leighty RM; Natarajan RN; and Rosenberg AG: Wear patterns on retrieved polyethylene tibial inserts
and their relationship to technical considerations during total
knee arthroplasty. Clin Orthop,1994.299: 31-43, 29931
1994
[PubMed]
Wasielewski RC; Parks N; Williams I; Surprenant H; Collier JP; and Engh G: Tibial insert undersurface as a contributing source of
polyethylene wear debris. Clin Orthop,1997.345: 53-9, 34553
1997
[PubMed]
Wright TM, and Bartel DL: The problem of surface damage in polyethylene total knee
components. Clin Orthop,1986.205: 67-74, 20567
1986
[PubMed]
Wright TM; Rimnac CM; Stulberg SD; Mintz L; Tsao AK; and Klein RW
McCrae C: Wear of polyethylene in total joint replacements. Observations
from retrieved PCA knee implants. Clin Orthop,1992.276: 126-34, 276126
1992
[PubMed]
Hutching IM. Tribology:
friction and wear of engineering materials. London: CRC
Press; 1992
Briscoe BJ: Wear of polymers: an essay on fundamental aspects. Tribology Internat.,1981.14: 231-43, 14231
1981
Schmalzried TP, and Callaghan JJ: Wear in total hip and knee replacements. J Bone Joint Surg Am,1999.81: 115-36, 81115
1999
[PubMed]