The Journal of Arthroplasty Vol. 24 No. 6 Suppl.

1 2009

Analysis of Wear in Retrieved

Mobile and Fixed Bearing Knee Inserts
Gerard A. Engh, MD, Rebecca L. Zimmerman, MS,
Nancy L. Parks, MS, and C. Anderson Engh, MD

Abstract: One purported advantage for mobile bearing (MB) knee implants is reduced
polyethylene wear. Twenty-three retrieved Low Contact Stress (LCS) MB and 31 Anatomic
Modular Knee (AMK) fixed bearing tibial polyethylene inserts, both Enduron (Depuy, Warsaw,
IN) and sterilized by gas plasma, were graded for top and backside wear using the Hood grading
system. For articular surface wear, there was no statistical difference between pitting and
scratching, but burnishing was twice as much for MB inserts (P = .003). For backside scores, there
was a minimal amount of pitting for both, but the scratching score was twice (P = .000) and the
burnishing score was 3 times greater for MB inserts (P = .000). At more than 5 years in situ, the
linear wear measurements were similar for the 2 groups (mean penetration was 0.329 mm for MB
and 0.320 mm for fixed bearing). Keywords: mobile bearing, fixed bearing, knee implants.
© 2009 Published by Elsevier Inc.

Polyethylene wear is a major limitation of the long-term
success of total knee arthroplasty (TKA) [1,2]. Although
new sterilization techniques adopted in the mid 1990s
have improved the performance of ultrahigh molecular
weight polyethylene, long-term wear is still a concern.
Even with tibial bearings that were sterilized with
conventional current methods (methods other than
gamma-air), polyethylene wear still accounts for most
revisions at the Anderson Clinic that were revised after 2
years in situ.
Mobile bearing (MB) technology was introduced more
than 25 years ago as a means to reduce knee implant
wear. By decoupling the joint and adding mobility
between the tibial insert and the tibial tray, the
articulating tibial geometry can more highly conform
to the femoral condylar geometry. This increase in
conformity leads to a decrease in contact stress and
theoretically a reduction in polyethylene wear. By
allowing this rotation, it also reduces the torsional and
shear forces seen at the bone-implant interface. Clinical
From the Anderson Orthopaedic Research Institute, Alexandria, Virginia.
Submitted January 14, 2009; accepted March 15, 2009.
Benefits or funds were received in partial or total support of the
research material described in this article. These benefits or support
were received from the following sources: Royalties for the AMK Knee
System from Depuy Orthopaedics, INOVA Health System, and a
cooperative agreement that was awarded and administered by the U.
S. Army Medical Research and Materiel Command (USAMRMC), and
the Telemedicine and Advanced Technology Research Center (TATRC),
Under Contract Number W81XWH-05-2-0079.
Reprint requests: Rebecca L Zimmerman, MS, Anderson Orthopae-
dic Institute, 2501 Parkers Lane Suite 200, Alexandria, VA 22306.
© 2009 Published by Elsevier Inc.
0883-5403/09/2406-0007$36.00/0
doi:10.1016/j.arth.2009.03.010
results support this hypothesis with revision rates for
aseptic loosening between 0% and 0.6% [3-5]. Lastly,
the mobility of the insert offers the advantage of
postoperative self-alignment. The ability of the bearing
to align itself allows the contact surface area between
the femoral component and tibial bearing to be
maximized [6].
In contrast, traditional fixed bearing (FB) knees have
less conforming geometry. The decreased level of con-
straint between the femoral component articulating
geometry and tibial bearing allows the soft tissue to direct
movement, which will theoretically allow the knee to
move anatomically. This reduction in constraint will also
lower the torsional and shear forces on the bone-implant
interface. This design, like the MB technology, allows
more postoperative self-alignment than more congruent
FB implants and reduces the likelihood of mechanical
loosening. Therefore, the major differences we see
between the 2 designs are the level of femoral compo-
nent/tibial bearing conformity and the introduction of a
second articulating surface with MB implants.
Retrieval studies are useful in providing the efficacy of
designs in vivo. Previous studies have investigated the
wear patterns on both MB tibial inserts and FB tibial
inserts; however, a direct comparison in the literature has
not been made between FB and MB tibial inserts
machined from the same polyethylene [7-10]. Many of
these studies were also performed on tibial bearings
sterilized by gamma radiation, and exposed to air during
storage [7,10]. This method of sterilization has been
proven to result in oxidation, which prematurely fatigues
the polyethylene and causes a loss in mechanical
properties and greater wear.

28
 Knee Mobile and Fixed Bearing Knee Inserts  Engh et al
In this study, Low Contact Stress (LCS) rotating
platform inserts (Depuy, Warsaw, IN) and Anatomic
Modular Knee (AMK) FB inserts (Depuy) retrieved at
revision TKA were analyzed to evaluate whether the
increased conformity and addition of an articulating
surface affected the in vivo wear of the polyethylene
inserts. Both sets of inserts were manufactured from
Enduron polyethylene (Depuy) and sterilized with gas
plasma. We hypothesized that retrieved MB tibial inserts
would have less surface wear and less linear penetration
than their less conforming FB counterparts.
Materials and Methods
Twenty-three MB and 31 FB tibial polyethylene
inserts all retrieved at the time of revision were
analyzed. The MB retrieved specimens were obtained
from 8 men and 15 women. The average age at revision
surgery was 65 ± 13 years, and the average length of
implantation was 29 ± 31 months (range, 1-102 months).
The average patient weight was 186 ± 36 lb. The reasons
for revision included instability/loosening (61%), infec-
tion (17%), osteolysis (4%), and other (17%).
The FB retrieved specimens were obtained from 17
men and 14 women. The average age at revision surgery
was 67 ± 11 years, and the average length of implantation
was 32 ± 33 months (range, 1-141 months). The average
patient weight was 218 ± 38 lb. The reasons for revision
included instability/loosening (42%), infection (32%),
osteolysis (6%), other (19%).
All tibial inserts were inspected visually for the type and
severity of wear using a technique developed by Hood et
al [11]. All inserts were inspected by 2 independent
observers for evidence of surface deformation, pitting,
embedded cement debris, scratching, burnishing, abra-
sion, delamination, and metal debris. Wear was classified
from 0 to 3, in which 0 indicates no wear, and 3 indicates
severe wear. Wear on both the articulating surface and
backside surface was assessed with the same technique.
The thicknesses of 6 MB and 6 FB tibial polyethylene
inserts in situ 55 to 148 months were measured using
digital calipers. The MB inserts were retrieved from 5
women and 1 man. The average age at revision surgery
was 65 ± 11 years, and the average length of implantation
was 76.6 ± 17.2 months (range, 55-102 months). The
average patient weight was 186 ± 30 lb. Four inserts were
10 mm, and 2 were 15 mm thick. Preoperative diagnosis
included a loose cemented tibia in 2 patients, infection in
2 patients, a loose cemented femur in 1 patient, and
osteolysis/polywear in 1 patient. All MB inserts retrieved
and analyzed were primary TKAs.
The FB inserts were retrieved from 2 women and 4
men. The average age at revision surgery was 67 ± 11
years, and the average length of implantation was 87.7 ±
30.6 months (range, 66-141 months). The average
patient weight was 218 ± 54 lb. One insert was 12 mm,
4 were 14 mm, and 1 was 16 mm. Preoperative diagnosis
included infection in 2 patients, osteolysis/polywear in 1
29
patient, a loose cemented tibia in 1 patient, a femoral
fracture in 1 patient, and a failed patella in 1 patient.
Three FB inserts retrieved and analyzed were from
revision TKAs, and 3 were from primary TKAs.
Linear wear was calculated by measuring the most
deformed or thinned area on the medial and lateral
articulating surface on each tibial insert and subtracting
this value from the original thickness of the insert. The
original thickness was obtained by subtracting the
thickness of the tibial tray from the thickness labeled on
the insert. The deviation from the original thickness
represented the final linear penetration value. Manufac-
turing tolerances of 0.254 mm were neglected when
making our measurements.
Results
The most common form of articular surface wear
identified was burnishing, followed by scratching and
pitting. Surface deformation and abrasion were very
minimal on the articulating surface of both FB and MB
tibial inserts. The average articulating surface deforma-
tion score was 0.9 ± 1.2 for the MB inserts and 0.8 ± 1.3
for the FB inserts. The average articulating surface
abrasion score was 0.6 ± 0.9 for the MB inserts and 0.5
± 1.2 for the FB inserts. There was no delamination,
embedded cement debris, or metal debris observed on the
articulating surface for any of the tibial inserts, FB or MB.
For the 3 modes of wear that were the most prevalent,
there was no statistical difference between MB and FB
pitting and scratching, but there was twice as much
burnishing for the MB inserts (P = .003) (Fig. 1).
Table 1 contains the total wear scores for the different
zones that were scored on the articulating surface of the
insert (medial anterior, medial posterior, lateral anterior,
and lateral posterior). Severe asymmetric wear was not
observed in either the MB or FB inserts.
There was a minimal amount of pitting on the backside
for both MB and FB inserts, but the average scratching
score was twice (P = .000), and the average burnishing
score was 3 times greater for MB inserts (P = .000) (Fig.
2). Surface deformation, embedded cement debris, and
Fig. 1. The MB vs FB articular surface scores and P values.
30 The Journal of Arthroplasty Vol. 24 No. 6 Suppl. 1 September 2009

Table 1. Total Articular Surface Wear Scores for Individual Table 2. Average Linear Wear Rate for the Medial and Lateral
Zones Compartments of MB and FB Tibial Inserts
Medial Medial Lateral Lateral Linear Wear
Anterior Posterior Anterior Posterior Linear Wear (mm) Rate (mm/y)
AMK Medial compartment
Average 2.8 2.5 2.3 2.1 MB 0.407 ± 0.394 (range, 0.263 to 1.207) 0.061 ± 0.041
Standard 1.6 1.6 1.4 1.6 FB 0.381 ± 0.554 (range, −0.157 to 1.357) 0.054 ± 0.083
deviation P value for MB vs FB medial .699
LCS compartment measurements
Average 3.8 3.6 3.7 3.4 Lateral compartment
Standard 1.8 1.7 1.6 1.7 MB 0.252 ± 0.359 (range, 0.057 to 0.177) 0.035 ± 0.040
deviation FB .259 ± .285 (range, −0.183 to 0.730) 0.029 ± 0.029
P value for MB vs FB lateral .937
compartment measurements
abrasion were very minimal on the backside of both FB
and MB tibial inserts. The average backside surface
deformation score was 0.5 ± 0.7 for the MB inserts and pared to the FB knee. The LCS knee allows a very large
0.4 ± 1.9 for the FB inserts. The average backside area of contact between both the medial and lateral
embedded cement debris score for the FB inserts was femoral condyles and the tibial polyethylene insert
0.1 ± 0.3, and the average backside abrasion score was throughout a wide range of motion. The overall increase
0.1 ± 0.3. There was no delamination or metal debris in conformity has been documented to decrease contact
observed for either group of tibial inserts. stress; however, as evidenced in this study, it also
Table 2 contains the average linear penetration values increases the contact between the femoral component
(mm) and linear wear rates (mm/y) for both the medial and tibial component resulting in a larger percentage of
and lateral compartments of the FB and MB knees. There burnishing wear [6,12,13]. The increased contact area
were no statistically significant differences between either results in a dominant wear mechanism that is similar to
group. total hip arthroplasty.
All MB tibial inserts were retrieved from the patient's Wear on the backside of FB tibial inserts has been
primary TKA. The FB knees were retrieved from 3 knees recognized as a potential source of polyethylene debris,
that were primary TKAs and 3 knees that were revision even with the tibial insert firmly locked to the tray [14,15].
TKAs. The linear measurements were further grouped for Therefore, the addition of a second articulating surface for
the FB TKAs into primary and revision measurements. MB knees raises concerns. This retrieval study did confirm
The linear measurements for the medial compartment in that on the backside of MB inserts, pitting, scratching, and
the primary FB TKAs was 0.340 ± 0.306 mm, and for the burnishing were significantly greater than their FB
lateral compartment was 0.404 ± 0.313 mm. The linear counterparts. The pitting and scratching observed were
measurements for the medial compartment in the most likely a result of third body debris. The scratches
revision FB TKAs was 0.421 ± 0.818 mm, and for the observed on the backside of the MB inserts were
lateral compartment 0.114 ± .346 mm. concentric and were aligned with rotational movement.
Prior studies have attributed these concentric rings on the
Discussion backside of rotating platform tibial inserts to third body
The MB knee was designed to increase conformity and debris trapped between the tibial tray and tibial insert
consequently decrease polyethylene wear when com- [10]. When compared to FB tibial inserts with improved
locking mechanisms, it appears that the introduction of
the second articulating surface creates more visual wear.
Ho et al [7] performed a similar study comparing FB and
MB knees; however, the tibial inserts in their study were
sterilized by gamma-air methods. They grouped the wear
modes into low-grade and high-grade wear, where low-
grade wear consisted of burnishing, abrasion, and surface
deformation, and high-grade wear consisted of pitting,
scratching, and delamination. In accordance with their
study, we found that MB knees increased low-grade wear;
however, unlike their study, we did not find a significant
increase in high-grade wear for FB knees. They reported
that 22.2% of the FB tibial inserts they evaluated had
severe wear through from delamination, whereas no
delamination was present in our study for either FB or MB
inserts. We also observed during this study that the gas
Fig. 2. The MB vs FB backside surface scores and P values. plasma–sterilized inserts we analyzed have very low wear
 Knee Mobile and Fixed Bearing Knee Inserts  Engh et al
rates. Prior studies have reported linear wear rates of
0.127 and 0.350 mm/y [16,17]. We observed wear rates of
approximately 0.05 to 0.06 mm/y for the medial
compartment and approximately 0.03 mm/y for the
lateral compartment of both fixed and MB knees. These
wear rates are in agreement with Kops and Swarts [9],
who reported a penetration rate of 0.054 mm/y for the
superior surface of LCS knees. They also analyzed
polyethylene tibial inserts sterilized with current meth-
ods. This is encouraging information that regardless of
design, improvements in sterilization techniques have
made a large contribution to reducing polyethylene wear.
Although visually the LCS knee appears to have more
articular surface wear than the AMK knee, this may be
deceptive. A visual grading scale, such as the Hood scale,
limits the user to a 2-dimensional assessment of the
surface. The extent of damage from burnishing wear, in
particular, may be underestimated using this 2-d grading
system. Burnishing wear occurs from the articulation of
the metal femoral component polishing the surface of the
polyethylene tibial component. In the Hood grading scale,
burnishing is scored by the percentage of the surface that
is burnished, ranging between 1 (mild) and 3 (severe).
The important factor that is not accounted for using this
2-d system is that polishing of the surface coupled with
cold flow of the polyethylene will eventually deform the
surface, meaning more material will be lost, but the score
will remain the same. For instance, the LCS knee is more
conforming than the AMK knee; therefore, there is more
contact between the femoral component and the tibial
insert. Hypothetically, this contact may result in a large
percentage of surface being polished or burnished, which
using a visual grading scale would warrant a score of 3.
This polishing may be very shallow and may not result in
a large amount of material removal, but it is termed
“severe.” On the other hand, the AMK knee may have a
smaller area that has been burnished. This area of
burnishing may be very deep, but because it covers less
than 50% of the surface, it is considered “moderate.” In
terms of actual material lost in our hypothetical example,
the AMK may have more material removed, but with our
visual grading scale, the LCS would be given the severe
score. Because of this phenomenon, the factor that is
more important when assessing wear damage is volu-
metric loss of material. Further investigation into the
volumetric loss of material is currently being done at
our laboratory.
Preliminary results from our volumetric measurements
have shown that surface deformation caused by burnish-
ing produces a larger volume of material loss than wear
modes, which are visually more unappealing, such as
pitting and scratching. This is a concern because
burnishing wear generates smaller particles, which have
been shown to produce a greater cellular response than
larger particles [18]. Previous studies have been con-
ducted on the in vivo generation of particulate debris for
the MB and FB prostheses and subsequently their
31
revision rates for osteolysis. Garcia et al [8] found in
their study of retrieved rotating platform knees that
greater levels of damage were present on both the inferior
and superior surface of the tibial inserts in knees with
greater radiographic radiolucency scores and mechani-
cally loose components. Minoda et al [19] collected the
synovial fluid 1 year after implantation from 22 TKA
patients with well-functioning prostheses. These were
conventionally sterilized MB and FB knees. They did not
find a statistically significant difference between the
number of particles, particle size, or particle shape. In
contrast, Huang et al [20] reported their observations for
the revision rate of MB and FB knees that were revised
for osteolysis. Osteolysis was identified intraoperatively in
16 (47%) of 34 knees in the MB group and in 6 (13%) of
46 knees in the FB group. In this study, there was a
similar revision rate for osteolysis between the 2 groups.
There are several imitations to this study. First is the
subset of inserts. The inserts analyzed were retrieved in
conjunction with some form of failure; therefore, they
may not represent the larger population of well-function-
ing knees. In addition, only a specific implant from a
single manufacturer was analyzed to represent FB and
MB knees. The findings may not be applicable to other
inserts of different designs or by different manufacturers.
Lastly, the study does not further expound on the reason
for failure and how it affects the wear pattern. The
different types of failure observed in this study are
markedly different and could affect the type of wear that
was observed.
The information presented in this study warrants further
investigation into the volume of material lost from the 2
implant designs. Visually, the benefits of reduced wear
with MB tibial inserts was not realized in our study, and
linear penetration was similar for the 2 groups.
References
1. Lonner JH, Siliski JM, Scott RD. Prodromes of failure in
total knee arthroplasty. J Arthroplasty 1999;14:488.
2. Sharkey PF, Hozack WJ, Rothman RH, et al. Insall Award
Paper: why are total knee arthroplasties failing today? Clin
Orthop Relat Res 2002;404:7.
3. Callaghan JJ, O'Rourke M, Iossi M, et al. Cemented
rotating-platform total knee replacement: a concise fol-
low-up, at a minimum of fifteen years, of a previous report.
JBJS Am 2005;87:1995.
4. Buechel F. Long-term followup after mobile-bearing total
knee replacement. CORR 2002;404:40.
5. Buechel F. Mobile-bearing knee arthroplasty: rotation is
our salvation! JOA 2004;19(4 Suppl 1):27.
6. Stukenborg-Colsman C, Ostermeier S, Hurschler C, et al.
Tibiofemoral contact stress after total knee arthroplasty:
comparison of fixed and mobile-bearing inlay designs. Acta
Orthop Scand 2002;73:638.
7. Ho F, Ma H, Liau J, et al. Mobile-bearing knees reduce
rotational asymmetric wear. CORR 2007;462:143.
8. Garcia R, Kraay M, Messerschmitt P, et al. Analysis of
retrieved ultra-high-molecular-weight polyethylene tibial
32 The Journal of Arthroplasty Vol. 24 No. 6 Suppl. 1 September 2009
components from rotating-platform total knee arthroplasty.
JOA 2009;24:131.
9. Kops A, Swarts E. Quantification of polyethylene degradation
in mobile bearing knees: a retrieval analysis of the anterior-
posterior-glide (APG) and rotating platform (RP) low contact
stress (LCS) knee. Acta Orthopaedica 2007;78:364.
10. Atwood S, Currier J, Mayor M, et al. Clinical wear
measurement on low contact stress rotating platform knee
bearings. JOA 2008;23:431.
11. Hood R, Wright T, Burstein A. Retrieval analysis of total knee
prostheses: a method and its application to 48 total condylar
prostheses. J of Biomedical Materials Research 1983;17:829.
12. McNamara JL. A comparison of contact pressures in tibial
and patellar total knee components before and after service
in vivo. CORR 1994;299:104.
13. D' Lima DD. Polyethylene contact stresses, articular con-
gruity, and knee alignment. CORR 2001;392:232.
14. Conditt M, Thompson M, Usrey M, et al. Backside wear of
polyethylene tibial inserts: mechanism and magnitude of
material loss. JBJS 2005;87-A:326.
15. Conditt M, Ismaily S, Alexander J, et al. Backside wear of
modular ultra-high molecular weight polyethylene tibial
inserts. JBJS 2004;86-A:1031.
16. Lavernia C, Sierra R, Hungerford D, et al. Activity level and
wear in total knee arthroplasty: a study of autopsy retrieved
specimens. JOA 2001;16:446.
17. Benjamin J, Szivek J, Dersam G, et al. Linear and
volumetric wear of tibial inserts in posterior cruciate-
retaining knee arthroplasties. CORR 2001;392:131.
18. Fisher J, McEwen H, Tipper J, et al. Wear, debris, and
biologic activity of cross-linked polyethylene in the knee.
Benefits and potential concerns. CORR 2004;428:114.
19. Minoda Y, Kobayashi A, Iwaki H, et al. Characteristics of
polyethylene wear particles isolated from synovial fluid after
mobile-bearing and posterior-stabilized total knee arthro-
plasties. J Biomed Mater Res B Appl Biomater 2004;71:1.
20. Huang C, Ma H, Liau JJ, et al. Osteolysis in failed total
knee arthroplasty: a comparison of mobile-bearing and
fixed-bearing knees. J Bone Joint Surg Am 2002;84-A:
2224.