CLINICAL ORTHOPAEDICS AND RELATED RESEARCH
Number 331, pp 107-117
In Vivo Knee Kinematics Derived Using
an Inverse Perspective Technique
Douglas A. Dennis, MD*?**;Richard D. Komistek, PhD****;
William A. H o ~PhD**;
, and Stefan M. Gabriel, PhDt
Sixty-four subjects having implanted and non-
implanted knees were studied using fluoro-
scopic videos. Each subject, flexing in the
sagittal plane, performed successive deep knee
bends under fluoroscopic surveillance. Fem-
orotibial contact in the sagittal plane was then
determined using image matching and discrete
digitization. At full extension, the mean con-
tact point of the normal and posterior stabi-
lized implanted femurs was anterior to the tib-
ial midpoint in the sagittal plane. The average
position was 6.49 mm (+3 - +13 mm) for the
normal knees and 0.30 mm (0 - +4 mm) for the
posterior stabilized knees. The implanted pos-
terior cruciate retaining and anterior cruciate
ligament deficient knees differed from the
other knee types. Their average initial contact
was posterior. The average contact at full ex-
tension for the posterior cruciate retaining and
anterior cruciate ligament deficient knees was
-5.13 mm (-2 - -8 mm) and -5.45 mm (-2 -
-14 mm), respectively. The femur of the nor-
mal knee contacts the tibia anterior to the mid-
point in the sagittal plane in full extension and
translates posteriorly during flexion. The fe-
mur of the posterior stabilized knee contacts
the tibia anteriorly, slightly less than the nor-
mal knee, and rolls back posteriorly during
flexion similar to normal knees. The femurs of
From the *Rose Musculoskeletal Research Laboratory,
Denver, CO; **Colorado School of Mines, Golden,
CO; and ?Johnson & Johnson Professional, Inc., Rayn-
ham, MA.
Reprint requests to Douglas A. Dennis, MD, 1601 East
19th Avenue, Suite 5000, Denver, CO 80218.
107
0 1996 Lippincott-Raven Publishers
the posterior cruciate retaining and anterior
cruciate ligament deficient knees contact the
tibia posterior in extension, but translate ante-
riorly during midflexion in a substantial num-
ber of cases, which is kinematically opposite of
the normal knees. The abnormal anterior
femoral translation observed in the posterior
cruciate retaining knees may be a factor in the
premature polyethylene wear seen in retrieval
studies.
Failure in the early years of total knee arthro-
plasty was most commonly due to aseptic
loosening, often associated with component
malalignment, soft tissue imbalance, or use
of constrained prosthesis.*.14.15.'8~*7~38Isolated
polyethylene wear failure in this era was un-
common. With improved instrumentation and
soft tissue balancing techniques, failure sec-
ondary to mechanical loosening has been
minima1.9.35.40f"' More recently, failures sec-
ondary to catastrophic polyethylene wear
have been observed, attributed to less con-
forming articular geometries,4.6.'3,25,*8.44 thin
or poor quality p0lyethylene,4,6,*5.~~ polyeth-
ylene sterilization methods,".39 or disturbed
knee kinematics.41 A better understanding of
knee joint kinematics is important to explain
the premature polyethylene wear failures ob-
served, and serves as the purpose of this in-
vestigation.
To date, most experimental studies of
knee kinematics have involved cadaveric, in

108 Dennis et al
vitro analyses, or have not tested the knee in
a weightbearing mode. l6.17.19.20.23.26.29.3 1.33.34
Others have used exoskeletal linkages and
skin markers that permit error due to unde-
sired motion between markers and the under-
lying bone.24 T h e present study investigates
knee kinematics using an in vivo, weight-
bearing methodology.
MATERIALS AND METHODS
Femorotibial contact position was determined for
64 implanted and nonimplanted knees (Institu-
tional Research Review Board approval). Sixteen
knees were normal, 10 were anterior cruciate lig-
ament deficient, 13 were implanted with posterior
cruciate retaining total knee arthroplasty, and 25
were implanted with posterior cruciate substitut-
ing total knee arthroplasty (Press-Fit Condylar
Designs, Johnson and Johnson Professional In-
corporated, Raynham, MA). Five subjects with
posterior stabilized knees and 3 subjects with
posterior cruciate retaining knees were tested bi-
laterally. Each subject was asked to perform 3
successive deep knee bends to maximum flexion
under fluoroscopic surveillance in the sagittal
plane. Each subject placed the foot of their leg to
be studied on a designated marker. The subjects
were initially fluoroscoped at full extension.
Patients were examined using a Siemens Sire-
mobil2000 Digital Xray image intensifier system
(Iseline, NJ). The fluoroscope images were stored
on videotape for subsequent redigitization using a
frame grabber. Clinical data for the subjects ana-
lyzed are shown in Table l .
The contact position between the femur and
the tibia was determined using 2 methods: ( 1 )
TABLE 1. Clinical Data
Group Male Female Total
Normal 7 9
Anterior cruciate 5 5
ligament deficient
Posterior stabilized 17 8 25
Posterior cruciate 6 7
retaining
Clinical Orthopaedics
and Related Research
3-dimensional image matching and (2) 2-dimen-
sional discrete digitization. The procedure using
image matching techniques was used initially and
discrete digitization was used as a verification.
Three-Dimensional Image Matching
Image matching involved the creation of a library
of 3-dimensional images of the tibial and the
femoral components. The initial 3-dimensional
computer assisted design models of the tibial and
femoral components were entered into a com-
puter algorithm and then reoriented at I" incre-
ments for the 2 out of plane rotations: (1) internal
and external rotation and ( 2 ) abduction and ad-
duction rotation. Internal and external rotations
were created for k20" and abduction and adduc-
tion rotations were created for ?lo". Each com-
ponent library, therefore, contained 861 images
corresponding to each combination of internal
and external and abduction and adduction rota-
tion. These computer assisted design model im-
ages were transformed into silhouettes to mimic
the appearance of the real tibial and femoral com-
ponents in the xray fluoroscopic images. Each sil-
houette was then centered in the middle of its
image, scaled to a constant predetermined size,
and rotated so that its principle axis was oriented
parallel to the horizontal axis of the image. The
result is a set of images of the prosthetic compo-
nents oriented at different out of plane rotations,
each normalized to a canonic size and in plane
orientation.
The images were then interactively analyzed
on a computer workstation using the Khoros im-
age analysis and display software (Albuquerque,
NM). The contours (silhouettes) of the tibial and
femoral prosthetic components were extracted in
each of the fluoroscopic images corresponding to
Hospital
Average for Special
Age Surgery Score
16 36.6 Not applicable
10 44.4 Not applicable
65.9 91.1
13 60.8 93.2
Number 331
October, 1996
C femoral and tibial components (C).
the desired increments of flexion. Each silhouette
was normalized for size and in plane rotation, and
matched to 1 of the images in the component li-
brary. The resulting match determines the 2 out of
plane rotation angles. The third angle is given by
the in plane rotation angle. The set of 3 angles
completely determines the 3-dimensional orienta-
tion of the prosthetic component in the Newton-
ian (world) reference frame.
The translation (position) of the component is
determined by an additional calculation. The dis-
tance between the component and the xray source
is determined by the apparent size of the compo-
nent in the 2-dimensional xray image (the dis-
tance is inversely proportional to the size). The
horizontal and vertical position of the component
in the Newtonian reference frame is determined
by its horizontal and vertical position in the im-
age, scaled by the distance between the compo-
nent and the xray source. This set of translational
displacements (horizontal, vertical, and in depth)
In Vivo Knee Kinematics 109
Fig 1A-C. The image matching process be-
gins with a 2-dimensional fluoroscopy image
(A). The femoral and tibial components were
matched with a 3-dimensional computer as-
sisted design image (B). The remaining 2-di-
mensional images from the fluoroscopic
videoframe, such as the bones, are removed
from the image allowing for a true 3-dimen-
sional image of the relative position between
completely determines the 3-dimensional posi-
tion of the prosthetic component in the Newton-
ian reference frame.
The best fit image according to the computer
algorithm was then overlaid onto the fluoroscopic
image (Fig 1). At each increment of flexion, the 2
out of plane rotations (internal and external rota-
tion and abduction and adduction) were deter-
mined via the computer algorithm from the
transformation matrix. With the tibial and
femoral out of plane rotations known from the
chosen images relative to the Newtonian refer-
ence frame, the femoral component rotations rel-
ative to the tibial component in the tibial
reference frame were then mathematically deter-
mined. The in plane contact position in the sagit-
tal plane was determined for each knee using a
relative transformation of the femoral coordinates
into the tibial reference frame. Each position of
the femoral component was transferred into the
tibial reference frame. The computer algorithm

110 Dennis et al
then located the medial and lateral femoral
condyle positions closest to the tibial plateau in
the tibial reference frame along the longitudinal
axis. The 3-dimensional coordinates of these
points were determined and designated as the
femorotibial contact position. The contact posi-
tion was measured from the tibial plateau mid-
point, designated as the origin position. Positions
anterior to the midpoint were denoted as positive
and positions posterior to the midpoint were de-
noted as negative.
Two-Dimensional Discrete Digitization
Discrete digitization was used to analyze the nor-
mal and anterior cruciate ligament deficient knees
and to verify the results obtained using the imag-
ing matching procedure used for the prostheti-
cally implanted knees. Designated points were
digitized on the femoral, tibial, and patellar com-
Fig 2. A schematic detailing the discrete points
(letters) that were digitized using a workstation
computer. The relative positions of those points
were used in a computer algorithm to determine
the femorotibial contact positions.
Clinical Orthopaedics
and Related Research
ponents (Fig 2). Femoral and tibial transforma-
tions relative to the Newtonian reference frame
were established. The coordinates of these points,
within their specified reference frames, were used
as input to a mathematic model that determined
relative angular rotations and translations. The
relative rotations between the femoral and tibial
reference frames were determined using transfor-
mation matrices of each component about the
Newtonian reference frame. The femorotibial
contact position was determined in the tibial ref-
erence frame. The contact position was measured
from the midpoint of the tibial plateau in the
sagittal plane. Although this method cannot deter-
mine the out of plane rotations and positions, it is
a good test to verify the in plane relative position
of the femur on the tibia.
RESULTS
Average femorotibial contact positions of the
4 knee types in the anteroposterior (AP) di-
rection throughout the range of flexion are
presented in Figure 3. Each of the normal
knees followed a very similar pattern. At full
extension, the femurs for this knee type con-
tacted the tibia anterior to the midline in the
sagittal plane. This average initial contact po-
sition was 6.49 mm anterior to the midline in
the sagittal plane. As this knee type flexed,
posterior femoral rollback occurred as the
contact position moved posteriorly an aver-
age of 14.17 mm throughout the range of
flexion (OO-90O).
The anterior cruciate ligament deficient
knees did not replicate normal knee kinemat-
ics. A high degree of variability was found
when comparing kinematic profiles of the
group of anterior cruciate ligament deficient
knees. They generally were posterior at full
extension. This group of knees produced 3
types of femorotibial contact patterns. The
first pattern involved an anterior translation
during midflexion. The second pattern pro-
duced by this knee type produced minimal
AP contact motion, with the femurs remain-
ing posterior on the tibial plateau during the
full flexion cycle. The third group produced
a combination of the first 2 Datterns. This
Number 331
October, 1996 In Vivo Knee Kinematics 111

- - - Anterior Cruciate Deficient

4
- -Posterior Cruciate Retaining

-E
A

C
o
Posterior Cruciate Substitutin

.-
.=0 4
8
n
.-b
t i
c - 8
8n
g
c
-12
a
-16
0 30 60 90 120 Fig 3. Average AP contact position
Knee Flexion (Degrees) of each knee type.

group would begin posterior, sliding anteri-
orly at midflexion, but again translating pos-
teriorly at higher flexion angles. The anterior
translation of femorotibial contact that oc-
curred for the anterior cruciate ligament de-
ficient knees generally occurred between 30"
and 60" or between 60" and 90" flexion. The
average change in contact position for this
knee type was only 1.6 mm in the posterior
direction during a deep knee bend to 90"
flexion. Although the average anterior trans-
lation was 0.5 mm, some knees did shift ex-
cessively anteriorly throughout the flexion
cycle as far as 13.7 mm.
The posterior cruciate retaining knees pro-
duced similar results to those obtained for the
anterior cruciate ligament deficient knees. At
full extension the femur contacted the tibia
posteriorly to the origin position with the av-
erage initial contact position being 5.13 mm
posterior. Erratic contact patterns seen with
the anterior cruciate ligament deficient knees
were also observed with the posterior cruci-
ate retaining knees. The same 3 contact pat-
terns were observed. Although the average
contact position moved posteriorly 3.79 mm
during flexion, some posterior cruciate re-
taining knees moved anteriorly as much as 7
mm. The anterior translation of femorotibial
contact position occurred between the ranges
of 30" to 60°, or 60" to 90" flexion.
The posterior stabilized knees best repli-
cated normal knee kinematics in the sagittal
plane. Although they did not begin the flex-
ion cycle as far anteriorly as the normal knee
group at full extension, they were the only
other group to begin the flexion cycle ante-
rior to the tibia1 midline in the sagittal plane.
The average initial contact position at full
extension was 0.3 mm anterior. At full exten-
sion, the initial contact position was as far
anterior as 4 mm. The majority of the poste-
rior stabilized knees began the flexion cycle
at the midline of the sagittal plane. As the
knee flexed, the posterior stabilized femoral
components did roll back posteriorly to a
similar position as the normal knee. The av-
erage final contact position at 90" flexion
was 7.62 mm posterior. The average amount
of posterior femoral rollback seen with the
posterior stabilized knee was 7.71 mm al-
though some showed posterior femoral roll-
back as much as 12.3 mm.
The AP femorotibial contact positions
for 5 randomly selected knees in each group
are shown in Figures 4 through 7. The nor-
 Clinical Orthopaedics
112 Dennis et al and Related Research

Fig 4. Sagittal femorotibial contact

0 30 60 90 positions of 5 randomly selected nor-
Knee Flexion (Degrees) mal knees.

ma1 knee and the posterior stabilized groups
had relatively reproducible results (Figs 4,
7), whereas the anterior cruciate ligament
deficient and posterior cruciate retaining
knee groups had highly variable contact pat-
terns among knees within the same group
(Figs 5 , 6 ) .
Numeric data corresponding to Figures 4
through 8 are shown in Tables 2 and 3. Ta-
bles 2 and 3 only display data for the range
of 0" to 90" flexion. Several graphs display
plotted data from 0" to 120" flexion. The ma-
jority of the anterior cruciate ligament defi-
cient and posterior cruciate retaining knees
did not flex to 120" during weightbearing.
Therefore, to make accurate comparisons be-
tween these 2 knee groups and the normal
and posterior stabilized groups, the data be-
tween 0" to 90" are shown.
DISCUSSION
Numerous knee kinematic analyses have doc-
umented the presence of posterior femoral
rollback associated with progressive knee
flexion. Draganich et al," in an anatomic spec-

- -Anterior Cruciate Ligament Deficient 1

Anterior Cruciate Ligament Deficient 2
Anterior Cruciate Ligament Deficient 3
-E
E

--- Anterior Cruciate Ligament Deficient4

Anterior Cruciate Ligament Deficient 5

-16 J I I ' Fig 5. Sagittal femorotibial contact po-

0 30 60 go
sitions of 5 randomlv selected anterior
Knee Flexion (Degrees) cruciate ligament dsicient knees.
 8

- 4

-
E
E

-a
s
.-
.- o
a 4
.-
-
b
2
UY
g
-8

e
t
4
-12
Fig 6. Sagittal femorotibial contact
-16 position of 5 randomly selected pos-
0 30 60 so terior cruciate retaining total knee
Knee Flexion (Degrees) replacements.

imen evaluation of the normal knee, found
13.5 mm of posterior femoral rollback oc-
curred with knee flexion, with 10 mm occur-
ring within the first 30" knee flexion. They ob-
served posterior translation was greater for the
lateral than medial femoral condyle, which
was attributed to the larger radius of curvature
of the lateral femoral condyle. Kurosawa et
a123 performed another anatomic specimen
knee study to assess posterior femoral rollback
using serial lateral radiographs. They also
found posterior femoral translation occurred at
the lateral femoral condyle, measuring 17 mm
on average, whereas the medial femur actually
translated anteriorly 2.2 mm throughout a
range of 0" to 120". El Nahass et all2evaluated
femorotibial translation of normal knees and
those implanted with total knee arthroplasty
during various activities in an in vivo study.
An externally applied orthotic fixture creating
a 3-dimensional magnetic field was applied to
the lower extremity for analysis. The average
posterior femoral rollback was activity depen-
dent, but similar for the implanted (9.2-14.1
mm) and normal knee (10.6-13.9 mm) groups.
Additional studies of posterior femoral
rollback have noted it occurs via a combi-
nation of rolling and sliding in the normal

8
I -Posterior Cruciate Substituting 1 I
- -Posterior Cruciate Substituting 2

- 4 Posterior Cruciate Substituting 3

-
E
E
r o
Posterior Cruciate Substituting 4
Posterior Cruciate Subs
.-*
a
n - 4
z
'C
8
g
n - 8
g
8
-12
Fig 7. Sagittal femorotibial contact
-16 positions of 5 randomly selected pos-
0 30 60 SO 120 terior stabilized total knee replace-
Knee Flexion (Degrees) ments.

114 Dennis et al
Fig 8. Tibia1 component retrieval showing ante-
rior wear pattern. Fluoroscopic analysis showed
anterior femoral slide of greater than 10 mm
with flexion.
knee.21.32.37 Rolling predominates in the
first 20" flexion with sliding predominating
at greater increments of flexion. The post-
erior translation is complex involving the
cruciate ligaments and related to the crossed
4-bar linkage.
The anterior shift of the femur on the tibia
observed in some posterior cruciate retaining
total knee arthroplasty and anterior cruciate
ligament deficient knees in the present inves-
tigation has numerous potential negative
consequences. First, anterior femoral trans-
lation results in a more anterior axis of flex-
ion, lessening maximum knee flexion.10
Second, the quadriceps moment arm is de-
creased, resulting in reduced quadriceps effi-
ciency. Third, anterior sliding of the femoral
component on the tibial polyethylene surface
risks accelerated polyethylene wear (Fig 8).
TABLE 2. Anteroposterior Position (rnrn)
Knee Type 0"
Normal 6.49
Anterior cruciate ligament -5.45
deficient
Posterior stabilized 0.30
Posterior cruciate retaining -5.13
Clinical Orthopaedics
and Related Research
Blunn et al,5 in a sophisticated laboratory
evaluation of polyethylene wear, found dra-
matically increased polyethylene wear with
cyclic sliding as compared with compression
or rolling.
Andriacchil reported the predominant
shear force during gait and stair climbing is
directed posteriorly on the tibia which is
normally resisted by the posterior cruciate
ligament, preventing anteriorization of the
femorotibial contact point. Numerous au-
thors2.7.42 have shown that AP femorotibial
translation is related to the mechanics of the
extensor mechanism, particularly the direc-
tion of pull of the patellar tendon and the de-
gree of knee flexion. At lesser degrees of
flexion the direction of the patellar ligament
pull is anterior, creating an anterior pull on
the tibia. This anteriorly directed shear force
on the tibia is normally resisted by the ante-
rior cruciate ligament. However, at greater
degrees of flexion (> 45"-60"), the direction
of patellar ligament pull on the tibial changes
to posterior. This creates a posterior shear
force on the tibia which is normally resisted
by the posterior cruciate ligament.
The present study has found that the
femorotibial contact position of anterior cru-
ciate ligament deficient knees and those im-
planted with posterior cruciate retaining total
knee replacements is posterior in full exten-
sion in contrast to normal knees in which
femorotibial contact is anterior to the mid-
line in full extension. This can be attributed,
at least in part, to the absence of the anterior
cruciate ligament, allowing unopposed ante-
rior pull of the patellar tendon on the tibia
~
30" 60" 90"
-0.86 -5.55 -7.68
-6.97 -7.55 -7.05
-4.36 -6.77 -7.62
-8.29 -9.06 -8.92
Number 331
October, 1996
~~
TABLE 3. Anteroposterior Displacement (mm)
Knee Type 0 - > 30'
Normal -7.35
Anterior cruciate
ligament deficient -1.52
Posterior stabilized -4.72
Posterior cruciate retaining -3.16
and resulting in the posterior shift of the
femorotibial contact observed in these 2
groups of knees within this flexion range.
Additionally, the .femorotibial contact posi-
tion of these same 2 knee groups shifted an-
teriorly, on average, between 60" and 90"
flexion. It is theorized this may be due to in-
sufficient tension in the posterior cruciate
ligament which poorly resists the posteriorly
directed shear force on the tibia from the ex-
tensor mechanism within this flexion range.
This subsequently allows the anterior shift of
femorotibial contact to occur. Draganich et
all1 found a similar relationship between fem-
orotibial contact position and cruciate liga-
ment integrity. In a cadaveric analysis they
also observed a posterior shift in femorotibial
contact after sectioning of the anterior cruci-
ate ligament. After the addition of posterior
cruciate ligament sectioning, an anterior shift
in femorotibial contact occurred as knee
flexion progressed, as was observed in the
anterior cruciate ligament deficient and pos-
terior cruciate retaining total knee arthro-
plasty groups in the current study.
Other investigators have similarly ob-
served a paradoxical anterior femoral transla-
tion with knee flexion after posterior cruciate
retaining total knee arthroplasty. Stiehl et
al,41 in a similar in vivo weightbearing fluo-
roscopic study of 47 patients implanted with
posterior cruciate retaining total knee re-
placements, observed the average femorotib-
ial contact in full extension was 10 mm pos-
terior to the midsagittal line of the tibia,
translating anteriorly to a point 5 mm anterior
to the midsagittal line with flexion (15 mm
In Vivo Knee Kinematics 115
30 - > 60° 60 - > 90" Total
-4.68 -2.14 -14.17
-0.58 +0.50 -1.60
-2.34 -0.65 -7.71
-0.78 +0.15 -3.79
total anterior translation). Banks and Hodge3
performed a similar 3-dimensional kinematic
analysis of posterior cruciate retaining total
knee arthroplasty using video fluoroscopy
and found maximum posterior femoral roll-
back occurred at less than 60" flexion with an
anterior shift in femorotibial contact occur-
ring at flexion ranges greater than 60". Using
sequential lateral radiographs, Kim et a122
studied femorotibial translation of 52 poste-
rior cruciate retaining total knee arthroplas-
ties. They observed that no posterior femoral
rollback occurred on average. Instead, the fe-
mur shifted anteriorly 0.6 mm. A high vari-
ability in femorotibial translation was found,
as noted in the present study, with contact
shifting anteriorly as much as 14.5 mm with
flexion to 90",whereas others shifted posteri-
orly as much as 12.7 mm.
An advantage of the present experimental
model is that it allows analysis using an in
vivo weightbearing method. The 3-dimen-
sional kinematic method reduces potential
error due to motion between external skin
markers and the underlying bone.24 The im-
portance of weightbearing in kinematic eval-
uation of the knee is supported by the work
of Hsieh and Walker,20 who determined that
in an unloaded knee joint laxity is primarily
determined by soft tissue constraints,
whereas in the loaded knee joint the geomet-
ric conformity of the joint surfaces plays the
major role in controlling joint laxity. Meth-
ods that allow assessment during dynamic
muscle contraction provide a superior esti-
mation of true knee kinematics. Markolf et
al,30 in an in vivo study of knee stability of

116 Dennis et al
49 normal patients, found laxity measure-
ments were reduced by as much as 50% with
the addition of muscle contraction.
SUMMARY
The present video fluoroscopic evaluation
showed similar sagittal kinematic patterns
between anterior cruciate ligament deficient
and posterior cruciate retaining total knee
arthroplasty, with the average femorotibial
contact position posterior in full extension
and shifting anteriorly between 60" to 90"
flexion. The posterior translation in femo-
rotibial contact can be attributed to the ab-
sence of the anterior cruciate ligament,
whereas the anterior translation suggests in-
adequate tension in the posterior cruciate lig-
ament. Although posterior stabilized total
knee arthroplasty more closely duplicated
normal knee kinematics, neither total knee
arthroplasty design duplicated the posterior
femoral rollback of the normal knee. The
paradoxical anterior femoral translation ob-
served may play a role in the premature poly-
ethylene wear reported in retrievals of some
posterior cruciate retaining total knee re-
placements.6.25 Those designs with relatively
flat, nonconforming condylar geometries
may be at greater risk of polyethylene wear if
the posterior cruciate ligament is not properly
tensioned at the time of operation and main-
tained throughout the lifespan of the total
knee implant. The high variability in contact
patterns of posterior cruciate retaining total
knee arthroplasty suggests this procedure is
highly technique sensitive, and strict atten-
tion to proper tensioning of the posterior cru-
ciate ligament is mandatory for the long term
success of this operative procedure.
References
1. Andriacchi TP: Functional analysis of pre- and post-
knee surgery. Total knee arthroplasty and ACL re-
construction. J Biomech Eng 115:575-581, 1993.
2. Andriacchi TP, Stanwyck TS, Galante JO: Knee bio-
mechanics and total knee replacement. J Arthro-
plasty 1:211-219, 1986.
3. Banks SA, Hodge WA: Accurate 3-D measurement
of dynamic knee replacement motions using single-
Clinical Orthopaedics
and Related Research
plane fluoroscopy. Validation and in vivo applica-
tion, Proceedings of the 19th Annual Meeting of the
American Society of Biomechanics, Stanford Uni-
versity, Palo Alto, CA 163-164, 1995.
4. Bartel DL, Bickness VL, Wright TM: The effect of
conformity, thickness, and material stresses in
UHMWPE components for total joint replacement. J
Bone Joint Surg 68A:1041-1051, 1989.
5. Blunn GW, Walker PS, Joshi A, Hardinge K: The
dominance ofcyclic sliding in producing wear in total
knee replacements. Clin Orthop 273:253-260, 1991.
6 . Collier JP. Mayor MB, McNamara JL, Surprenant
VA, Jensen RE: Analysis of the failure of 122 poly-
ethylene inserts from uncemented tibia1 knee com-
ponents. Clin Orthop 273:232-242, 1991.
7. Daniel DM, Stone ML, Barnett P, Sachs R: Use of the
quadriceps active test to diagnose posterior cruciate-
ligament disruption and measure posterior laxity of
the knee. J Bone Joint Surg 70A:386-391, 1988.
8. Deburge A: GUEPAR: GUEPAR hinge prosthesis.
Complications and results with two years follow-up.
Clin Orthop 120:47-53, 1976.
9. Dennis DA, Clayton ML, O'Donnell S, Mack RP,
Stringer EA: Posterior cruciate condylar total knee
arthroplasty: Average 1 I-year follow-up evaluation.
Clin Orthop 281:168-176, 1992.
10. Dennis DA, Komistek RK, Hoff WA, Gabriel SM,
Kettler MM: Passive versus weightbearing range of
motion of the knee determined using fluoroscopy.
The Interim Meeting of the Knee Society, Boston.
MA, 18, 1995.
I I . Draganich LF, Andriacchi T, Andersson GBJ: Interac-
tion between intrinsic knee mechanics and the knee
extensor mechanism. J Orthop Res 5539-547, 1987.
12. El Nahass B, Madson MM, Walker PS: Motion of
the knee after condylar resurfacing. An in vivo
study. J Biomech24:1107-1117, 1991.
13. Feng EL. Stulberg SD, Wixson RL: Progressive sub-
luxation and polyethylene wear in total knee re-
placements with flat articular surfaces. Clin Orthop
299:667 I , 1994.
14 Freeman MAR, Todd RC, Bamert P, Day WH:
ICLH arthroplasty of the knee, 1968-1977. J Bone
Joint Surg 60B:339-344, 1978.
15 Freeman PA: Walldius arthroplasty. A review of 80
cases. Clin Orthop 94:85-91, 1973.
16 Fukubayashi T, Torzilli PA, Sherman MF, Warren
RF: An in vitro biomechanical evaluation of ante-
rior-posterior motion of the knee. Tibia1 displace-
ment, rotation, and torque. J Bone Joint Surg
64Az258-264, 1982.
17 Garg A, Walker PS: Prediction of total knee motion
using a three-dimensional computer-graphics
model. J Biomech 23:45-58, 1990.
I8 Goldberg VM, Henderson B T The Freeman-Swan-
son ICLH total knee arthroplasty. J Bone Joint Surg
62A: 1338-1 344, 1979.
19 Grood ES, Suntay WJ: A joint coordinate system for
the clinical description of three-dimensional mo-
tions: Application to the knee. ASME 105:136-1 44,
1983.
20. Hsieh HH, Walker PS: Stabilizing mechanisms of
the loaded and unloaded knee joint. J Bone Joint
Surg 58A:87-93, 1976.
Number 331
October, 1996
21. Kapendji 1A: The Physiology of the Joints. New
York, Churchill Livingstone 114-123, 1970.
22. Kim H, Pelker RR, Lynch JK, Gibson DH, Irving
JF: Radiographic analysis of the rollback phenome-
non in posterior cruciate-retaining total knee arthro-
plasty: No rollback observed. The Sixty-first Annual
Meeting of the American Academy of Orthopaedic
Surgeons, New Orleans, LA, 424, 1994.
23. Kurosawa H, Walker PS, Garg A, Hunter T Geome-
try and motion of the knee for implant and orthotic
design. J Biomech lk4878-499, 1985.
24. Lafortune MA, Cavanagh PR. Sommer 111 HJ, Kale-
nak A: Three-dimensional kinematics of the human
knee during walking. J Biomech 25:347-357, 1992.
25. Landy MM, Walker PS: Wear of ultra-high-molecular-
weight polyethylene components of 90 retrieved knee
prostheses. J Arthroplasty (suppl):S73-85, 1988.
26. Lew WD, Lewis JL: The effect of knee-prosthesis
geometry on cruciate ligament mechanics during
flexion. J Bone Joint Surg 64A:734-739, 1982.
27. Lewallen DG, Bryan RS, Peterson LF: Polycentric
total knee arthroplasty. A ten-year follow-up study. J
Bone JointSurg66A:1211-1218, 1984.
28. Lewis P, Rorabeck CH, Bourne RB, Devane P: Pos-
teromedial tibia1 polyethylene failure in total knee
replacements. Clin Orthop 299: 11-1 7, 1994.
29. Markolf KL, Finerman GM, Amstutz HC: In vitro
measurements of knee stability after bicondylar re-
placement. J Bone Joint Surg 61A:547-557, 1979.
30. Markolf KL, Graff-Radford A, Amstutz HC: In vivo
knee stability. A quantitative assessment using an in-
strumented clinical testing apparatus. J Bone Joint
Surg 60Af364-674, 1978.
3 1. Markolf KL, Mensch JS, Amstutz HC: Stiffness and
laxity of the knee. The contributions of the supporting
structures. J Bone Joint Surg 58A:583-593, 1976.
32. Muller W: The Knee. Form, Function and Ligament
Reconstruction. New York, Springer 8-17, 1983.
33. Nilsson KG, Karrholm J, Ekelund L: Knee motion in
total knee arthroplasty. A roentgen stereophotogram-
metric analysis of the kinematics of the Tricon-M
knee prosthesis. Clin Orthop 256: 147-161, 1990.
34. Nilsson KG, Karrholm J, Gadegaard P: Abnormal
kinematics of the artificial knee. Roentgen stereo-
In Vivo Knee Kinematics 117
photogrammetric analysis of 10 Miller-Galante and
five New Jersey LCS knees. Actd Orthop Scand 62:
440-446, 1991.
35. Ranawat CS, Boachie-Adjie 0: Survivorship analy-
sis and results of total condylar knee arthroplasty.
Eight-to eleven-year follow-up period. Clin Orthop
226:&13,1988.
36. Ries MD, Greer J, Rose RM, Weaver K, Sauer WL:
Sterilization induced effects on ultra high molecu-
lar weight polyethylene (UHMWPE) performance
properties The Sixty-second Annual Meeting of the
American Academy of Orthopaedic Surgeons, Or-
lando. FL, 397, 1995.
37. Segal P, Jacob M: The Knee. St. Louis, Mosby
3 2 4 2 , 1983.
38. Skolnick MD, Coventry MB, Ilstrup DM: Geo-
metric total knee arthroplasty. A two-year follow-
up study. J Bone Joint Surg 58A:749-753, 1976.
39. Sperling DK, Collier JP, Sutula LC, Saum KA, Cur-
rier JH: Role of sterilization damage in the fatigue
failure of retrieved polyethylene knee components.
The Sixty-second Annual Meeting of the American
Academy of Orthopaedic Surgeons, Orlando, FL,
217. 1995.
40. Stem SH, Insall JN: Posterior stabilized prosthesis.
Results after 9-12 years follow-up. J Bone Joint
Surg 74A:980-986, 1992.
41. Stiehl JB, Komistek RD, Dennis DA, Paxson RD,
Hoff, WA: Fluoroscopic analysis of kinematics after
posterior-cruciate-retaining total knee arthroplasty
using fluoroscopy. J Bone Joint Surg 77B:884-889,
1995.
42. Van EijdenTMGJ, DeBoer W, Weijs WA: The orien-
tation of the distal part of the quadriceps femoris
muscle as a function of the knee flexion-extension
angle. J Biomech 18:803-809, 1985.
43. Vince KG, Insall JN, Kelly MA: The total condylar
prosthesis. Ten- to 12-year results of a cemented
knee replacement. J Bone Joint Surg 71B:793-797,
1989.
44. Wright TM, Rimnac CM, Stulberg SD, et al: Wear of
polyethylene in total joint replacements. Observa-
tions from retrieved PCA knee implants. Clin Or-
thop 276:126-134, 1992.