(iopynght I 982 by The J(’urn1I of Bone ,,,I Joini Surgery.

I, to, jorate1

The Influence of Total Knee-Replacement Design

on Walking and Stair-Climbing
BY THOMAS P. ANDRIA((HI, PH.D.t, JORGE 0. GALANTE, M.D.t, AND REX W. FERMIER, B.S.1’, CHICAGO, ILLINOIS

Frin (lit !)ejirt!iieFit 0/ Vrthopcdi Surerv

#{149} . Rush-Presbvte,ian-St . Luke s Medical Center. CIzua,t,o

ABSTRACT: The relationship between gait and gait. An improved understanding of gait and the variables
prosthetic design was studied during level walking and associated with total knee designs is essential in address-
stair-climbing for twenty-six asymptomatic patients ing this controversy. Quantitative studies of gait during
after total knee replacement. An age-matched group of activities of daily living are needed to generate this infor-
fourteen control subjects was also studied. Five designs mation, and will be useful for the evaluation of total knee-
of total knee replacement - Geomedic, Gunston, total replacement devices and for providing understanding of
condylar, duopatellar, and Cloutier - were used. the loading patterns that may occur during normal activity.
Differences in gait could be identified on the basis Several studies have evaluated gait in patients with
of prosthetic design. The more stressful stair-climbing knee disease. These investigations included kinematic
test produced more clearly differentiated function analyses I .6.7.9.Ii.2I , time-distance measurements, and
among the different designs. Patients who were treated force-plate measurements. There have also been several
with the least-constrained cruciate-retaining (Cloutier) kinetic and force-analysis studies of function in normal
design of prosthesis were the only group that had a subjects and in patients after treatment for knee dis-
normal range of motion during climbing up and down abilitiesaIa14l7ao. The common finding of these studies
stairs. Two groups of patients with semiconstrained was that patients who appear to be clinically asymptomatic
(total condylar and Geomedic) designs had a lower after joint replacement have abnormal gait patterns. Cur-
than normal range of knee fiexion while descending rently, little is known about the nature of the gait abnor-
stairs. Patients with the other designs of prosthesis had mality in patients after total knee replacement or its rela-
a normal range of knee motion on stair-climbing. tionship to total knee-replacement design.
Kinematic and anatomical differences among the five The purpose of this study was to evaluate the relation-
designs did not have as great an influence on function ship between gait and total knee-replacement design. The
during level walking as they did during stair-climbing. prosthetic knees that were selected for this study were
The results of this study indicate that after total knee considered to be representative of cruciate-sacrificing and
replacement even asymptomatic patients with excellent sparing designs with varying amounts of constraint. The
clinical results have an abnormality of gait. The fea- parameters ofgait that we observed included time-distance
tures of the abnormality were common to most of the patterns and motion and moments of the knee joint. The
patients in the series, and consisted of a shorter than gait of patients who had received one of five different de-
normal stride length, reduced mid-stance knee flexion, signs of total knee replacement was evaluated and corn-
and abnormal patterns of external flexion-extension pared with that of control subjects.
moment of the knee. Although an explanation of these
Materials and Methods
abnormalities of gait is not completely possible at this
Twenty-six patients. in five experimental groups. werestudied during level
time, they appear to be related to the interaction of the
walking and stair-climbing. Patients were grouped according which of five total
to

kinematics of the knee and surrounding soft tissues. knee designs they had received. The five implants selected for this study were the
Geomedic, Gunston, total condylar. duopatellar, and Cloutier designs.
CLINICAL RELEVANCE: It appears that patients with
The five designs of prosthesis were selected to represent varying shapes of the
less constrained cruciate-retaining designs of total knee articular surfaces and the retention of one, both, or neither cruciate ligament. The
Geomedic prosthesis has fairly congruous articular surfaces. requires removal of
replacement have a more normal gait during stair-
the anterior cruciate ligament, permits retention of the ps)sterior cruciate ligament.
climbing than patients with more constrained and does not include a patellar flange or resurfacing. The Gunston prosthesis con-
sists of two separate semicircular runners that articulate with two independent tibial
cruciate-sacrificing designs. During level walking, pa-
components, permits retention of both cruciate ligaments. and does not include
tients with five quite different designs all had abnor- patellar resurfacing or a patellar flange. The total condylar design requires the sac-
rifice of both cruciate ligaments. with anterior-posterior stability provided by the
malities of gait in spite of a successful clinical result.
conformity of the tibial articulating surfaces; all patients with this design had patel-
lar resurfacing. The duopatellar prosthesis permits retention of the posterior
cruciate ligament. includes a patellar flange. and allows patellar resurfacing, which
There is currently a great deal of controversy regard- was performed in all of the patients whom we examined. The Cloutier prosthesis
ing which type of total knee prosthesis provides better allows retention of both cruciate ligaments and the femoral condyles are asymmet-
neal, diverge. and have varying radii of curvature. The tibial component of the
Cloutier device consists of flat articular surfaces supported on a metal retainer, and
the design has a patellar flange. but patellar resurfacing was not performed in our
Funded in part by National
Institutes of Health Grants patients.
KO4AMO()493 and ROIAM2O7O2 and
by the Arthritis Foundation. The patients selected for this study were matched according to postoperative
1 Department of Orthopedic Surgery . Rush-Presbyterian-St. Luke’s pain. function, passive range of motion, and joint stability. A point system based
N’lcdical Center. 753 Wcst Congress Parkway, Chicago. Illinois 60612. on The Hospital for Special Surgery knee.rating system was used to quantitate

132tt THE JOURNAL OF BONE AND JOINT SURGERY

 INFLUENCE OF TOTAL KNEE-REPLACEMENT DESIGN ON WALKING AND STAIR-CLIMBING 1329
TABLE I

COMPARISON OF FIVE GROUPS OF PATIENTS ACCORDING TO PROSTHETIC DESIGN

No. of No. of Diagnosis ( No. of Knees) Average Passive Average

Type of No. of Average No. of No. of Bilateral Unilateral Rheumat. Time Range Knee
Prosthesis Knees Age Men Women Replacements Replacements Arth. Osteoarth. Postop.5 of Motiona Ratingt
( Yrs.) ( Mos.) (Degrees) (Points)

Gunston 7 59.2 1 6 3 4 2 5 50.0 100 92.1

( 19.7) (9.7) (8) (6.0)
Geomedic 7 58.7 3 4 5 2 5 2 37.9 105 93.3
(18.8) (10.3) (12) (3.4)
Duopatellar 6 62.2 2 4 0 6 3 3 21.8 115 96.2
(5.2) (11.3) (16) (4.3)
Total condylar 8 59.5 2 6 6 2 7 1 16. 1 103 95.2
(9.9) (9.3) (9) (1.9)
Cloutier 8 63.8 2 6 7 1 6 2 12.0 112 90.9
(10.5) (1.8) (11) (8.4)
Total 36 60.7 10 26 21 15 23 13 27.0 107 93.6
(12.8) (17.2) (12) (5.4)

* Standard deviations are in parentheses.

t Maximum score = 100 points.

postoperative status. All patients were evaluated at least one year postoperatively, anatomical axes. Angular velocity and acceleration about the longitudinal axis of
and to qualify for this study the result had to score 85 points or more on the basis of the limb segment were assumed to be negligible. The inertial properties of the
this system. All patients had an excellent clinical result, were able to walk without limb segments were approximated from data derived from the literature’. It was
aids, had little or no pain, and were able to climb stairs in a reciprocal manner. assumed that the fiexion-extension axis remained perpendicular to the plane of
No attempt was made to match the patients in terms ofdiagnosis, involvement progression, that the abduction-adduction and internal-external rotation axes at the
of other joints. sex distribution. age. or presence of bilateral total knee replace- hip joint moved with the thigh segment, that the axes of the knee joint moved with
ment. However, patients with moderate or severe involvement of other joints that the shank segment, and that the ankle joint moved with the foot. The components
was associated with pain were not included in the study. since the result would not of the moment vector were resolved into directions such that moments producing
have achieved the clinical rating that was required as a prerequisite for inclusion. flexion-extension, abduction-adduction, and internal-external rotation at each joint
The resulting population for the experimental study consisted of thirty-six could be identified. In addition, time-distance measurements and parameters of
knees in twentysix patients (Table I). In eighteen patients with twenty-three in- limb motion were derived for these measurements. For the purpose of this study,
volved knees rheumatoid arthritis was diagnosed, while the remaining eight pa- only data about the knee joint will be described.
tients with thirteen involved knees had osteoarthritis. Six patients had had a previ. The moments were analyzed for both patterns and magnitudes . The patterns
ous operation: four on the fool (three Gunston and one Geomedic prosthesis), one were identified by timing the occurrence of the relative maximum and sign rever-
total hip replacement (duopatellar prosthesis), and one osteotomy of the hip sal The sensitivity
. of the moment patterns and magnitudes to shifts in the position
(Geomedic prosthesis). No patient had a passive knee-flexion contracture of more of the fixed joint center (for normal gait) were studied by theoretically moving the
than 5 degrees. The posterior cruciate ligament was retained in the patients with the joint center a total of thirty millimeters in five-millimeter increments (fifteen mil-
appropriate designs of prosthesis. Four of the eight knees with the Cloutier pros- limeters on either side of the nominal position) in three orthogonal directions.
thesis still had the anterior cruciate ligament. while no documentation was avail- Differences between the averages for the experimental and control groups
able on the presence of the anterior cruciate ligament in patients with the Gunston were tested for significance using the Student t test with a level of p < 0.05. A
design. least-square curve was used to fit first. second, and third-order polynomials to the
The control population consisted of fourteen healthy adults (seven men and relationships of the various measurements of gait and walking speed. An analysis
seven women) with an average age of 62.4 years (standard deviation, 6.3). Both of the multiple correlation coefficient was used to determine which order of
the control group and the study group were tested using the same protocol. The gait polynomial was used. Differences in regression coefficients were tested indepen-
of each subject was measured during twelve stride cycles, each of which occurred dently among patients and control subjects, assuming a Student t distribution.
midway during a separate walking trial on a ten-meter walkway. Each subject was
instructed to walk at three nominal speeds: slow, normal, and fast. Patients were Results
never requested to walk faster or slower than would normally be comfortable. All
measurements of gait were analyzed in relation to walking speed for each sub- Level Walking
ject’2. Each subject was also observed while ascending and descending a three.step
staircase . The range of walking speeds of the patients with total
The experimental observations were based on an idealization that considered
knee replacement was 0.5 to I .92 meters per second,
the lower extremities as a three-dimensional linkage with movable joints at the hip.
knee, and ankle. The joints were assumed to have fixed axes of motion. The ap- while the range of walking speeds for the normal mdi-
proach was similar to that reported by others5#{176}20. Motion of the lower limbs was
viduals was 0.61 to 2.05 meters per second. The average
monitored by observing the spatial position of six light-emitting diodes that were
placed at the anterior superior iliac spine, at the center of the greater trochanter, self-selected walking speeds were placed in three catego-
over the mid-point of the lateral joint line of the knee, on the lateral aspect of the
ries (slow, normal, and fast) on the basis of the instruc-
malleolus at the ankle. at the base of the calcaneus, and at the base of the fifth
metatarsal. The positions of the joint centers at the hip. knee, and ankle on the tions given at the time of testing. The patients’ average
sagittal plane were located relative to the position ofthe light-emitting diodes at the
greater trochanter, lateral joint line of the knee, and lateral malleolus. respectively.
slow and normal speeds were both 79 per cent of those of
The position of the center of the knee joint on the frontal plane was located by the normal group (Fig. 1). The range of self-selected
identifying the mid-point of a line between the peripheral margins of the medial
speeds tended to overlap those of the three nominal-speed
and lateral plateaus at the level of the joint surfaces. The hip joint was located 1 .5
centimeters distal to the mid-point of a line from the anterior superior iliac spine to categories, producing measurements of gait that were dis-
the pubic symphysis. The ankle joint was estimated to be at the mid-point of a line
tributed over a range of walking speeds. Each of the other
from the tip of the lateral malleolus to the tip of the medial malleolus. The joint
centers were located and marked on each subject prior to the observation of gait. measurements of gait was considered with walking speed
The three-dimensional position of each light-emitting diode was sampled
as an independent variable in order to separate differences
seventy-five times per second using an optoelectronic system consisting of two op.
tical digitizers and electronic signal-conditioning2. Ground-reacflon force mea- in gait that were due to factors other than changes in
surements were acquired simultaneously with the measurements of limb position
speed”.
using a piezoelectric force platform. The force platform provided the three compo-
nents of ground.reaction force. vertical twisting moments. and location of resultant The stride length of the patients with total knee re-
forces at the foot.
placement was shorter than that of the control group over
To calculate the moments, each segment of the limb (thigh. shank, and foot)
was idealized as a rigid body with a coordinate system chosen to coincide with the range of walking speeds (Fig. 2). The relationship be-

VOL. 64-A, NO. 9. DECEMBER 1982

1330 T. P. ANDRIACCHI, J. 0. GALANTE, AND R. W. FERMIER

20 Normals
TKR Patients
0
0)
U)
5
E

-
0)
0)
a. 10
Ct)

0 05

0
“Slow” “Normal “ “Fast”
FIG. I
A comparison of the average self-selected walking speeds of the patients and the control subjects when they were asked to walk at nominal slow,
normal. and fast speeds. There were no significant differences between the average walking speeds of the patients according to the design of the
prosthesis.

tween stride length and walking speed for the control sub- tern of the fiexion-extension moment during gait was not
jects and the patients was best fit by a first-order least- altered by displacing the position of the joint center by as
square polynomial. The intercept of the curve for the much as fifteen millimeters on either side of the nominal
groups ofpatients was found to be less (p < 0.05) than that position of the joint. These results are similar to those re-
for the control group, while the slopes of the two curves ported by other investigators’9 There was no significant
(24 per cent of height per meter per second) were not dif- difference between the regression coefficients in the rela-
ferent. tionship of stride length to walking speed among the pa-
The majority (75 per cent) of the patients with knee tients in the five groups.
replacement had abnormal patterns of flexion-extension The change in the angle of knee fiexion during stance,
moments during stance phase. In response to an assumed defined as the difference between the angle of knee fiexion
error of fifteen millimeters in the location of all three axes at heel-strike and the relative maximum flexion occurring
of rotation, the largest change in the magnitude of the during mid-stance (Fig. 3), was also less in the patients
moment ( 16.5 per cent) was for the moment about the than in the control subjects over the range of walking
Ilexion-extension axis. However, the normal biphasic pat- speeds. The relationship between changes in knee fiexion

20

( 0

100
a)
I
2 90

a,
C
a)
-J Normal
Sublects
a)

70
(I)

60

TKR
50 Patients

‘1

0 05 0 5 20
Walking Speed (m/sec)

FIG. 2
An illustration of the difference in the relationship between stride length and walking speed in patients and control subjects. The curve for the
control subjects (n = ninety-seven and r = 0.87) and for the patients (n - 108 and r = 0.85) had significantly different intercepts.

ThE JOURNAL OF BONE AND JOINT SURGERY

 INFLUENCE OF TOTAL KNEE-REPLACEMENT DESIGN ON WALKING AND STAIR-CLIMBING 1331

Mid - stance

C
0

a)
20

a) Normal
a)
C
Subjects

a)
0 l0
C
a
U,

TKR
Patents

0 0.5 l.0 5 20
Walking Speed (m/sec)

FIG. 3
An illustration of the difference in the change in the relationship between knee fiexion during stance and walking speed in control subjects (n =

ninety-seven and r = 0.65) and in patients n = 108 and r = 0.57).

during stance and walking speed was found to be similar to maximums tending to extend the joint (Fig. 4). The times
the relationship between stride length and walking speed when the relative maximums occurred as well as the
(Figs. 2 and 3). An increase in walking speed was as- number of occurrences of crossing the zero axis were used
sociated with an increased change in knee flexion during to classify three distinct patterns. The pattern charac-
stance. A first-order polynomial was found to best fit the terized as normal was found to occur in 80 per cent of the
change in the relationship of knee flexion to walking speed control subjects (Table II). The two abnormal moment pat-
during stance for both the control subjects and the patients. terns (Fig. 4) tended to maintain an extrinsic moment pre-
The intercept of the curve in the patients with a total knee dominantly tending to either flex the joint (flexional mo-
replacement was found to be less (p < 0.05) than that for ment pattern) or extend the joint (extensional moment pat-
control subjects, while the slopes were not different. tern) throughout stance phase. The abnormality in the flex-
Again, there were no differences between the patients in ional and extensional moment patterns occurred during the
terms of the different designs of prosthesis. A patient who middle portion of stance phase. The distribution of the pat-
was walking at the same speed as a control subject tended terns for each of the groups of patients and for the control
to change the angle of knee flexion substantially less dur- subjects is shown in Table II.
ing stance than did the control subject. The reduced stride Differences in the maximum magnitude of the
length in the patients appeared to be coupled with a re- moments tending to flex and extend the knee joint were
duced change in the amount of knee flexion during the consistent with the distribution (Tables III and IV) of the
middle portion of the stance phase. moment patterns. The two groups of patients (those with
A third characteristic of abnormal gait in the patients total condylar and those with Geomedic prostheses) with a
with knee replacement was the pattern of external fiexion- majority of extensional moment patterns had a sig-
extension moment about the nominal fixed center of the nificantly lower than normal moment tending to flex the
knee joint. The moments were compared by selecting the knee. The only other moment amplitude that was different
test conducted at a walking speed that was closest to one
meter per second for each subject (Fig. 1). The average TABLE II
trial speed that was used when comparing the moments of FLF.XION-EXTENSION MOMENT PATTERNS OF THE KNEES
the controls ( I .02 ± 0. 14 meter per second) and of the pa-
tients (0.97 ± 0.09 meter per second) was not different. Type of Moment Pattern ( No. of Kn ees)
Prosthesis Normal Flexional Extensional
The variations in the patterns were too large to be ex-
plained by systematic errors in the location of the nominal Total condylar 0 3 5
Geomedic I I 5
joint center. In all of the patients the moment patterns in
Duopatellar I 3 2
the di rections of abduction- adduction and internal-external Gunston 3 3 1
rotation were normal. Cloutier 3 2 3
Total 8 II 16
The normal curve of flexion-extension moment was
Control group I 1 3 0
characterized by a biphasic pattern, with two relative
maximums tending to flex the joint and two relative s Walking speed of one meter per second.

VOL. 64-A, NO. 9. DECEMBER 1982

1332 T. P. ANDRIACCHI, J. 0. GALANTE, AND R. W. FERMIER

TABLE III

MAXIMUM M OMENTS IN THE K NEE DURING LEVEL WALKING ( PER CLVT Boor WLIGHT X HLIGHT)*t

Type of
Prosthesis Extension Flexion Adduction Abduction External Internal

Totalcondylar 2.72 l.86t 3.05 0.41 0.45 0.39

( 1.42) (0.96) (0.93) (0.39) (0. 14) (0. 15)
Geomedic 2.00 l.801 4.00 0.40 0.60 0.40
( 1 .00) (0.80) (0.70) (0.02) (0.30) (0.20)
Duopatellar 2.07 2.78 3.78 0.22 0.61 0.44
( 1.26) ( 1.70) ( 1.57) (0. 1 1) (0.31) (0. 15)
Gunston 2.40 2.30 4.20 0.701 0.80 0.40
(1.99) (1.71) (0.24) (0.08) (0.60) (0.30)
Cloutier 2.49 2.67 3.70 0.41 0.56 0.50
(1.07) (1.87) (1.43) (0.07) (0.24) (0.23)
Control group 2.29 2.74 3.57 0.25 0.54 0.41
( 1 . 10) (0.98) (0.90) (0.25) (0. 14) (0. 19)

Walking speed of one meter per second.

t Standard deviations are in parentheses.
Significant difference from the control value.

from normal occurred in the patients with the Gunston the control subjects. In addition, when the patients were
prosthesis, who had an abnormally higher moment tending grouped according to moment patterns, the patients with
to abduct the knee joint. All of the other measurements of the flexional moment pattern had a significantly larger
moments showed no difference between the patients and angle of knee flexion at heel-strike ( 10.6 ± 5.7 degrees)
than did patients with the extensional moment pattern
(4.24 ± 3. 1 degrees). Both groups tended to have the
:;: C
smaller amount of change in knee flexion during stance, as
: 2 described earlier.

Stair-Climbing
: All patients were able to climb up and down stairs in a
reciprocal manner. The speeds of ascending and descend-
ing stairs were not different for the patients and the control
l000/o subjects (0.32 ± 0.09 meter per second). For the patients
% Stance Phase
30
T KR
C Fle.ional Moment

I 0 Total
0 Condylor .- S -I

a LA
0
-C
-0 0 \%,I (±91) ±3 I)
Geomedic , S -I

Ea) 20
0’-
Duo
30 I- #{149}
50 l#{244}0#{176}/ Patellor
0/0 Stance Phase
30

C TKR Gunston t S
20
0 Estensionol Moment
I 0)
I0
U-
S ‘I 80% Pattern
r274l% ti8% 2 5)
0 Cloutier ‘ . 1

I
C 0
0
C
a)
E 20
Normal
0

LU 30
U 50 lO0#{176}/
60 70 80 90 100
Stance Phase
Knee Flexion (Degrees)
FIG. 4

An illustration of the characteristics of the three patterns of flexion- FIG. 5

extension moments about a fixed point at the knee joint. The average The differences in total range of the knee while climbing
of motion up
times t,-t, were characteristic of the so-called normal pattern, and the stairs among the five groups of patients and the control subjects. The pa-
average times t-t. were characteristic of the patients with total knee re- tients with the Cloutier prosthesis were the only group whose range of
placement. motion did not differ from that of the control group.

THE JOURNAL OF BONE AND JOINT SURGERY

 INFLUENCE OF TOTAL KNEE-REPLACEMENT DESIGN ON WALKING AND STAIR-CLIMBING 1333

TABLE IV

MAXIMUM MOMENTS IN THE KNEE DURING STAIR-CLIMBING (PAR CLNT BODY W.u;HT X H!,GHT)5

Extension Flexion Adduction Abduction Internal External

Ascending stairs
Patients 1.08 3.76 2.47 0.52 0.52 0.18
(0.87) (1.70) (1.31) (0.39) (0.11) (0.09)
Control subjects 0.84 4. 1 1 2.56 0.33 0.51 0.13
(0.91) (1.72) (1.05) (0.28) (0.15) (0.06)

Descending stairs
Patients 3.28 8.67t 2.10 1.41 1.18 0.46
(2.10) (3.10) (1.1) (1.10) (0.31) (0.17)
Control subjects 2.89 1 1.12 2.57 1.23 1.20 0.40
(1.04) (2.47) (1.51) (1.05) (0.28) (0.20)

Standard deviations are in parentheses.

t Significant difference from the control value. There were no significant differences between the five groups of patienLs.

in all five groups. with the exception ofthose with the total (Fig. 7). Pattern 1 was characterized by a waveform that
condylar prosthesis, the speed while descending stairs did reached a maximum in a flexion direction during mid-
not differ from that of the controls (0.49 ± 0.07 meter per support phase, while Pattern 2 was characterized by a sign
second). The patients with the total condylar prosthesis reversal from flexion to extension at 44 per cent of stance
descended stairs at a speed of 0.3 1 ± 0. 10 meter per sec- phase. The variations of these patterns could not be cx-
ond. plained by systematic differences in the location of the
The patients with the Cloutier prosthesis had the same
range of motion of the knee while climbing up stairs as the 5.0
control subjects (Fig. 5). The patients in the other groups,
40
however, used a significantly smaller range of knee motion C

than the control group while climbing up stairs. Patients

I; Pattern I
with the duopatellar, Gunston, and Cloutier prostheses
: 2.0
were not different than the control subjects during de-
scending stairs, while patients with the total condylar and l.0
Geomedic
knee motion
prostheses
(Fig. 6).
had significantly smaller ranges of
i 0.0
Two distinctive patterns of flexion-extension mo- E .0
ft
ments of the knee while ascending stairs were identified 2.0
.
Pattern 2 \ I
C 0
Total
3.0-
Condylar
U4.0

Geomedic 1- #{149} -4

500/0 l00#{176}/
0
Duo
Patellar
0/0 Stance Phase
FIG. 7

The characteristics of the two types of llexion-extension moment pat-

Gunston 4- . 4 terns found during stair-climbing. The sign reversal at t = 44 per cent
was the characteristic used to separate these patterns.

Cloutier 4- . nominal center of the knee joint of as much as one and

one-half centimeters of displacement from the fixed cen-
ter. All control subjects had Pattern- I flexion-extension
Normal p #{149}
moment of the knee during stair-climbing. Five patients
.I& ,
with the total condylar, two with the Geomedic, and one
60 70 80 90 100 with the duopatellar prosthesis had Pattern-2 moment
Knee Flexion (Degrees) characteristics at the knee, while all of the patients with a

FIG. 6
Gunston or Cloutier prosthesis had Pattern- I moments.
The maximum magnitude of the moment components dur-
The differences in the total range of motion of the knee while descend-
ing stairs among the five groups of patients and the control group. The ing stair-climbing did not differ among the five groups of
range of motion of the patients with a total condylar prosthesis and those
with a Geomedic prosthesis was significantly smaller than that of the
patients. All patients had a flexion moment that was less
control subjects. than that of the control subjects when descending stairs

VOL. 64-A, NO. 9. DECEMBER 982

1334 T. P. ANDRIACCHI, J. 0. GALANTE, AND R. W. FERMIER

(Table IV). In the patients who climbed up stairs with the there may be factors other than clinical differences or dif-
Pattern-2 flexion-extension moment, the moment tending ferences in prosthetic design that influence the way that
to extend the knee ( I .24 ± 6.3 per cent body weight x people walk after total knee replacement.
height) was higher than that for the control subjects. One explanation for these abnormal characteristics of
gait could be that after total knee replacement the patients
Discussion
continued to walk with a pattern that they had learned prior
The results of this study indicate that differences in to treatment. Since we did not obtain preoperative mea-
gait can be identified on the basis of prosthetic design. In surements for these patients, this explanation could not be
addition to measurements of gait during level walking, evaluated in our study. However, there is some indication
more stressful tests such as stair-climbing are useful in de- from other studies that the gait characteristics of patients
termining differences. The patients who were treated with with untreated arthritis are different from those of normal
the least-constrained cruciate-retaining (Cloutier) design subjects .,
of prosthesis were the only ones who had a normal range Another possible cause for these abnormalities in
of motion during ascending and descending stairs. With walking is abnormal muscle function. Several characteris-
the Cloutier design both cruciate ligaments are retained, tics of the gait in these patients seem consistent with this
and the prosthesis has flat tibial surfaces that minimally explanation. The abnormality in gait occurred during the
constrain anterior-posterior and rotatory displacement. In middle portion of the stance phase, when normally the
assessing the reasons for the more normal gait of these pa- quadriceps muscle is eccentrically contracting to balance
tients during stair-climbing, some aspects of the kinematic gravitational and inertial forces at the knee. In the process,
behavior of the prosthesis appear to be factors. However, the quadriceps allows the knee to flex to approximately 20
the range of motion of the knee alone does not provide the degrees during mid-stance in a controlled manner. Patients
entire explanation, since all of the groups of patients had with total knee replacement do not flex the knee in this
the same passive range of knee flexion. Because differ- manner. The abnormal patterns of flexion-extension
ences were observed during activity, the possibility exists moments seen in this study are also consistent with the
that the interaction of the kinematics of the knee joint with explanation that the muscles are functioning abnormally,
the surrounding ligaments and muscles produces func- since the flexor and extensor muscle groups of the knee are
tional differences between designs. As the knee flexes the primary internal structures that are capable of equili-
from zero to 90 degrees, the contact area between the brating the extrinsic flexion-extension moments.
femur and tibia will move posteriorly io.i:t increasing the Abnormal muscle function could be caused by a par-
mechanical advantage of the quadriceps muscles sig- tial loss of proprioceptive control or by a reduction or im-
nificantly by increasing their moment arm about the condy- balance in muscle capacity after joint replacement. Both
lar contact point. This explanation is consistent with the proprioceptive control and muscle capacity can be
finding that the myoelectric activity of the quadriceps influenced by the interaction of the joint kinematics with
muscles under constant moment is reduced by nearly a fac- the surrounding soft tissues and muscles. It is possible that
tor of two as the knee moves from full extension to 40 de- different prosthetic designs impose different strains on the
grees of flexiont. The mechanical advantage of this poste- soft tissues (where afferent receptors are located) and thus
nor movement may not be possible to the same degree in a alter the feedback on joint position, which depends on a
more constrained design, and thus gait during stair- combination of skin, muscle, and joint receptors i6#{149}Simi-
climbing is compromised. larly, a loss of quadriceps-muscle capacity or mechanical
The presence and function of the posterior cruciate advantage could inhibit its use during gait and produce the
ligament may also be a factor influencing the patient’s abnormality of an extensional moment pattern.
ability to descend stairs. The patients with the total condy- As indicated earlier, the different kinematics as-
lar (posterior-cruciate-sacrificing) prosthesis had a reduced sociated with the various prosthetic designs may sig-
descent velocity as well as a reduced range of knee mo- nificantly alter the mechanical advantage of the quadriceps
(ion. As the knee flexes while the patient descends stairs, mechanism. The total explanation of the abnormality in
the posterior cruciate ligament is in a position to resist the gait may include loss of both proprioception and muscle
forward thrust of the femur on the tibia. In the absence of capacity.
the ligament, the joint must be stabilized primarily by In assessing the results of this study, it is important to
constraint built into the prosthesis, since neither the mus- consider the limitations of the methods and assumptions
cular nor the secondary ligamentous restraints are in a used in acquiring the data. The moments reported in this
mechanically advantageous position to provide an efficient study were calculated about a fixed point at the geometric
substitute. center of the knee joint. Thus, for every subject, external
The differences in design among the five total knee moments were calculated about the same relative point at
replacements did not influence gait during level walking as the knee joint. The sensitivity of both the magnitude and
much as during stair-climbing. The fact that common patterns of moments in the position of the fixed nominal
characteristics of gait were present in a rather in- joint center has been tested. Both in this study and in an-
homogeneous mixture of subjects tends to indicate that other study 19 the normal patterns of the flexion-extension

THE JOURNAL OF BONE AND JOINT SURGERY

 INFLUENCE OF TOTAL KNEE-REPLACEMENT DESIGN ON WALKING AND STAIR-CLIMBING I 335

moment of the knee were not altered by relatively large for the latter subject due to the fact that he or she was walk-
changes ( ± 1 .5 centimeter) from the nominal position of ing at a slower speed. However, without collecting mea-
the joint center. The abnormality of gait in these patients surements over the range of walking speeds, we would not
was sufficiently large so that the characteristics of the pat- know if a subject was walking with a normal relationship
terns that we have identified would not be altered by errors between stride length and walking speed or with the ab-
in the position of the joint center. normal relationship seen in the knee-replacement patients.
In comparing the characteristics of gait among the The normalization of moments by the product of the sub-
groups of patients and the control subjects, walking speed ject’s height and weight was assumed to reduce variability,
was used as an independent variable. Many aspects of gait but no statistical assessment was performed.
change with changes in walking speed. Therefore, in order It is not completely possible at this time to explain the
to differentiate between the effects of differing walking abnormalities in gait observed in this study. However,
speeds and actual abnormalities ofgait, gait measurements their existence raises some new questions regarding the
were compared in terms of their relationship to walking designs of total knee-replacement prostheses and the re-
speed. In spite of the fact that the patients’ so-called nor- habilitation methods used after operation. We still do not
mal walking speed was slower than that of the control know the effects of these abnormalities of gait on energy
group, we were able to identify differences between the consumption and joint forces, nor do we know the effects
mechanisms of walking in the two groups. Thus, this ap- of these forces on bone, implant, and cement interfaces.
proach allowed interpretation of results when comparing At this stage in the evaluation of total knee-replacement
data between two groups walking at different speeds. For design, functional testing of patients treated with these
example, if two control subjects were observed, one walk- devices constitutes an important source of new informa-
ing at the average speed for the control group and the sec- tion.
ond walking at the average speed for the knee-replacement
NoTi: The authors gratefully acknowledge the substantial contritrulions of E)r. J. M.
patients, the stride length would be expected to be shorter (butter and Ms. Susan Massarello.

References
1. ANDRIAHI, T. P.; OGLE, J. A.: andGALANTE, J. 0.: Walking Speed as a Basis for Normal and AbnormalGait Measurements. J. Biomech..
10: 261-268. 1977.
2. ANDRIM’HI. T. P.; HAMPTON, S. J.; SCHULTZ, A. B.; and GALANTE, J. 0.: Three Dimensional Coordinate Data Processing in Human Motion
Analysis. J. Biomech. Eng. . 101: 279-283, Nov. 1979.
3. ANDRIACHI, T. P.; ANDERSSON, G. B. J.; BJ#{246}RK, R.; and ORTENGREN, R.: A Study of the Relationship between External Moments and
Myoelectric Activity about the Knee. Trans. Orthop. Res. Soc. , 7: 255, 1982.
4. ANDRIAC-HI. T. P.; ANDERSSON. G. B. J.; FERMIER, R. W.; STERN, D.; and GALANTE, J. 0.: A Study of Lower-Limb Mechanics during
Stair-Climbing. J. Bone and Joint Surg. , 62-A: 749-757, July 1980.
5. CAPPOZZO, AUREI.IO; LEO, TOMMAs0; and PEDOTTI. ANTONIO: A General Computing Method for the Analysis of Human Locomotion. J.
Biomech. , 8: 307-320, 1975.
6. CHAO, E. Y.: Functional Evaluation of Total Knee Replacement Patients Through Gait Analysis. American Society of Mechanical Engineers,
Publication 75-APMB-5, April 1975.
7. Cl-lAO, E. Y.; LAUGHMAN. R. K.; and STAUFFER. R. N.: Biomechanical Gait Evaluation of Pre and Postoperative Total Knee Replacement
Patients. Arch. Orthop. Traumat. Surg., 97: 309-317, 1980.
8. DEMPSTER. W. T.. and GAUGHRAN. G. R. L.: Properties of Body Segments Based on Size and Weight. Am. J. Anat.. 120: 33-54, 1967.
9. GOLDFLIES. M. L.; ANDRIACCHI. T. P.; andGALANTE, J. 0.: The Relationship between Varus Deformity and Moments at the Knee During Gait
and the Changes at the Knee after High Tibial Osteotomy. Trans. Orthop. Res. Soc. , 6: 54, 1981.
10. GOODFELLOW, JOHN, and O’CONNOR, JoHN: The Mechanics of the Knee and Prosthesis Design. J. Bone and Joint Surg. . 60-B(3): 358-369,
1978.
II. GRIEVE, D. W., and GEAR. R. J.: The Relationships between Length of Stride. Step Frequency, Time of Swing and Speed of Walking for
Children and Adults. Ergonomics. 9: 379-399. 1966.
12. HARRINGTON. I. J.: A Bioengineering Analysis of Force Actions at the Knee in Normal and Pathological Gait. Biomed. Eng., 11: 167-172,
I976.
13. IsEKI. F., and TOMATSU, T.: The Biomechanics of the Knee Joint with Special Reference to the Contact Area. Keio J. Med., 25: 37-44, 1976.
14. JoHNsoN. F., and WAUGH, W.: Methods for Routine Clinical Assessment of Knee-Joint Forces. Med. Biol. Eng. Comput., 17: 145-154, 1979.
15. KETTELKAMP. D. B., and LEAVERTON. P. E.: Gait Characteristics of the Rheumatoid Knee. Arch. Surg., 104: 30-34, 1972.
16. MCCLOSKY, D. I.: Kinesthetic Sensibility. Physiol. Rev., 58: 763-820, 1978.
17. M-GROUTHER, D. A.: Evaluation ofa Total Hip Replacement. J. Biomed. Mater. Res. Symp., No. 5, Part 2, pp. 271-283 (in J. Biomed. Mater.
Res.. vol. 8), 1974.
18. MORRISON. J. B.: The Mechanics of the Knee Joint in Relation to Normal Walking. J. Biomech., 3: 51-61, 1970.
19. NISSAN. M.: Review of Some Basic Assumptions in Knee Biomechanics. J. Biomech., 13: 375-381, 1980.
20. PAUL, J.: Force Actions Transmitted in the Knee of Normal Subjects and by Prosthetic Joint Replacement. In Total Knee Replacement. pp.
126-131. The Institute of Mechanical Engineers. 1974.
21. RITTMEN. NANCY; KETTELKAMP, D. B.; PRYOR, PHILIP; SCHWARTZKOPF, G. L.; and HILLBERRY, BEN: Analysis of Patterns of Knee Motion
Walking for Four Types of Total Knee Implants. Clin. Orthop. , 155: 1 1 1- 1 17, 198 I.
22. STAUFFER, R. N.: LAUGHMAN. R. K.; CHAO, E. Y.; and ILSTRUP. D. M.: Biomechanical Evaluation of Knee Function, before and after Total
Knee Replacement. Trans. Orthop. Res. Soc., 7: 55, 1981.
23. WINTER, D. A.: Overall Principle of Lower Limb Support during Stance Phase of Gait. J. Biomech., 13: 923-927, 1980.

VOL. 64-A, NO. 9. DECEMBER 1982