STAIR ASCENT KINEMATICS INCREASE UHMWPE DAMAGE IN TOTAL KNEE REPLACEMENTS
+*Babalola, O. M.; *Furman, B D; *Wright, T M
+*Hospital for Special Surgery, Department of Biomedical Mechanics & Materials, New York, NY
Polyethylene damage in total knee implants is partly controlled by
kinematics. In vitro tests provide a fairly good picture of causes and
modes of wear when compared to retrievals, but do not give a complete
clinical representation of actual in vivo kinematics. Usually restricted to
simulating level gait, these tests do not incorporate activities of daily
living, such as stair ascent and decent, kneeling, and rising from a chair,
which can contribute to increased wear and altered wear patterns. Though
current studies have attempted to incorporate more rigorous testing1-4, in
vivo data have generally not been used as input for tibial insert testing. To
begin to address this limitation, we performed tests aimed at
characterizing wear patterns of tibial inserts produced by more
physiologic in vitro testing, comprised of normal gait and clinically
acquired stair climbing kinematics, with the premise that this would
produce wear patterns more representative of in vivo kinematics and
those observed in retrieved implant components. We used two types of
knee implants, which differed in their articular surface designs, to
determine the influence of contact geometry on the wear results.
MATERIALS AND METHODS
Four NexGen Cruiciate Retaining Augmentable (CR) and four NexGen
Legacy Posterior Stabilized (LPS) tibial inserts (Zimmer, Warsaw, IN),
all of size E and 9mm thickness, were used. The CR inserts were
machined from 1050 compression molded sheet UHMWPE and the LPS
inserts were directly molded from 1900 resin. All were gamma irradiated
in nitrogen to 25 kGy. Six additional inserts, 3 per design, were kept
unloaded in a 37ºC deionized water bath as soak controls. All inserts were
pre-soaked for 30 days prior to testing. Corresponding NexGen CR and
LPS femoral components and metallic tibial trays were used.
Knee simulation was conducted on a 4-station Instron/Stanmore KCI
knee simulator with AP springs pre-compressed (no gap, spring constant
14.48N/mm). Two test conditions were run. The first used the ISO
Standard for normal gait, providing a baseline for the study. Axial loads
were applied from 0 to 40º of flexion, and the max load was 2279N run at
a frequency of 1.4Hz. The second incorporated two protocols: the ISO
Standard and stair climbing. The stair climbing protocol was acquired
from a previous gait study of 22 patients, who had received the Genesis II
total knee system (Smith & Nephew, Memphis, TN) and were asked to
walk, ascend and descend stairs, and rise from a chair while twisting.5
While performing these tasks, patients were monitored with a 3-D motion
analysis system in conjunction with a force plate and fluoroscopy. Nine
of these test subjects (69 6 yrs, 5 males/4 females, ht = 174 8 cm, wt
= 85 10 kg) had stair climbing data available for computations of the
inputs (Fig. 1).
Quadriceps forces were calculated using a model described by Smidt,
et al.6 The stair climbing input was incorporated at a ratio of 70:1 to
ensure a more clinically relevant test of the inserts based on previous
reports.1,7 The compressive axial force, flexion angle, AP shear force and
IE torque curves were determined from the gait analysis and were used as
inputs for the knee simulator. The flexion was from 20 to 100º, and the
max load was 2828N run at a frequency of 1Hz. Because the simulator’s
maximum flexion range was only 80°, the femoral components were pre-
mounted with 20º of flexion.
Figure 1. ISO gait kinematics (black); stair ascent kinematics (gray)
100
Fle x ion Angle (de g)
80
Ax ia l (N)
60
40
20
0 0
-20
0 100 200
ste p #
400
300
AP Torque (Nm )
AP Force (N)
200
100
0 0
-100 0 100 200
-200
-300
ste p #
All inserts were assessed for evidence of wear, the presence of damage
modes, and the percent contact surface area that showed wear damage.
Gravimetric wear measurements were acquired, corrected for moisture
absorption, and plotted versus total cycles. Wear rate (mg/Mc) was
determined from linear regression, and the wear rates from the two
conditions (normal gait) were compared with those from the second
(normal gait with incorporated stair climbing cycles).
RESULTS
In the first million cycles (Mc), the wear rate of the CR inserts for the
first test condition was 4.3 1.3 mg/Mc with a damage area of 22% of the
total available area on the insert. The LPS inserts had a wear rate of
11.0 2.7 mg/Mc, with a damage area of 17%. The damage modes of both
insert types were burnishing, pitting, and scratching.
The wear rates from test condition 2 were 3.9 0.4 mg/Mc and
10.4 2.9 mg/Mc for the CR and LPS inserts, respectively, after 1 Mc of
testing. The damage area was 28% for the CR insert and 24% for the
LPS. The posts of the LPS inserts showed only slight posterior damage
and no medial or lateral damage.
Stair climbing resulted in similar AP displacements and greater IE
rotations than normal gait: CR: 5.0mm, 9°, LPS: 6.0mm, 8.25° for gait;
CR: 4.5mm, 16.25°, LPS: 6.5mm, 19° for stairs.
Figure 2. A) Damage areas of inserts; B) Damage maps of stair and gait
at 1Mc (gait damage map in orange, stair in gray and scratches in blue)
A. Damage Area (mm2)
Tibial insert ~0.5Mc 1 Mc
Test 1 CR 635 624
PS 533 536
Test 2 CR 696 891
PS 626 752
B. Damage Maps
CR PS
DISCUSSION
An understanding of wear patterns and their causes can be gained by in
vitro testing, but only if that testing includes loads and kinematics that
simulate in vivo condition. Our results demonstrate that the addition of
bouts of stair climbing kinematics into a simulation of normal gait
markedly increased the wear damage area but did not affect the wear rate
as compared to when no stair climbing was included in the tests. Stair
climbing resulted in a 35% increase in damage area for the CR inserts and
28% increase for the LPS inserts compared to normal gait. The dominant
damage mode was burnishing. By 1Mc, the damage area of CR inserts
increased 72% and PS inserts 53%. Wear and damage modes were not
affected, but longer duration tests could show otherwise. In comparing
designs, the CR femoral component had flatter condyles and resulted in
larger contact areas in the ML direction. The LPS component was
slightly more conforming in the AP direction and resulted in larger AP
contact areas and AP displacements.
3000
The addition of more vigorous tests, better reflecting in vivo
2500
conditions, is essential to preclinical testing of implants. The damage
2000
1500 patterns observed in this study will be compared to those observed in
1000 retrieval implants, and we speculate that the addition of other loading
500
protocols based on clinical activities may explain more clearly the wear
0 100 200 patterns observed. Future plans include running additional tests to 4Mc.
ste p #
References: [1]McLeod et al J. Biomechanics, 8:369, 1975 [2]D’Lima et al.
15 Clin Orthop 392:124, 2001 [3]Muratoglu et al. Clin Orthop 404: 89, 2002 [4]
Wimmer et al. ORS 27:159, 2002 [5]Banks et al. ORS 25:431, 2000 [6]
10
5 Smidt, Gary Biomechanics 6: 79, 1973 [7]Benson et al. ORS 27:1006, 2002.
Acknowledgements: The authors acknowledge our colleagues in the Motion
-5
0 100 200 Analysis Lab and Jeremy Rawlinson for their contributions, and NIH Grant
IR01AR049793 and Zimmer Inc. for their support of this research.
-10
ste p #
50th Annual Meeting of the Orthopaedic Research Society
Paper No: 0296