Comparative Study On The Biomechanics of Knees

Katholieke Universiteit Leuven
Group Biomedical Sciences

Faculty of Medicine
Department of Orthopaedics
A COMPARATIVE STUDY ON THE BIOMECHANICS
OF THE NATIVE HUMAN KNEE JOINT AND
TOTAL KNEE ARTHROPLASTY
Jan Victor, MD
Promotor: J. Bellemans, MD, PhD
Copromotor: J. Vander Sloten Ir, PhD, J. Somville MD, Ir, PhD
Chair: P. Herijgers, MD, PhD
Secretary: F. Luyten, MD, PhD
Jurymembers: F. Catani, MD, PhD, University of Bologna, Italy
C. Delloye, MD, PhD, Universit Catholique de Louvain
R. Verdonk, MD, PhD, Universiteit Gent
P. Broos, MD, PhD, P. Herijgers, MD, PhD, F. Luyten, MD, PhD, KUL
Leuven, 11/09/09
Doctoral thesis in Medical Sciences
CONTENTS
Chapter 1: Introduction ......................................................................... 7
I. Justication ........................................................................................... 7
II. Kinematics of the native knee .............................................................. 9
A. Historical insights ...................................................................................... 9
B. MRI studies.............................................................................................. 11
C. RSA studies.............................................................................................. 12
D. Fluoroscopy studies ................................................................................. 13
E. Conclusion ............................................................................................... 13
III. Kinematics of the replaced knee ........................................................ 15
A. Relation between kinematics and implant design ................................... 15
B. Relation between kinematics and alignment ........................................... 23
Chapter 2: Objectives and methodology ........................................ 33
I. Aims of the thesis ............................................................................... 33
II. Hypotheses ......................................................................................... 34
III. Materials and methods of the ex vivo experiment ............................. 35
A. Methodological sequence ....................................................................... 35
B. Description of data processing ................................................................ 42
C. Knee kinematic simulator ....................................................................... 45
D. Accuracy and sensitivity analysis ............................................................ 48
i. Reference frame stability
ii. Repeatability
iii. Effect of ankle load
iv. Single versus double squats
v. Optical tracking system accuracy
Chapter 3: Results of the ex vivo experiment ............................... 60
I. Intra- and inter-observer variability ................................................... 60
II. Horizontal plane geometry of the distal femur ................................... 73
III. Horizontal plane geometry of the proximal tibia ............................... 85
IV. Component position ........................................................................... 86
A. Coronal plane ........................................................................................... 86
B. Sagittal plane ........................................................................................... 88
C. Horizontal plane ...................................................................................... 88
V. Tibiofemoral kinematics of the native and the replaced knee ............ 90
A. The inuence of muscle load on tibiofemoral kinematics .......................90
B. The inuence of muscle load on replaced knee kinematics .................. 106
C. Effect of quadriceps load on the knee near full extension ..................... 108
D. Changes induced by joint replacement .................................................. 112
E. Case study .............................................................................................. 117
VI. Ligament isometry in the native and the replaced knee ..... 122
A. The native knee ...................................................................................... 122
B The replaced knee .................................................................................. 135
Chapter 4: The in vivo experiment ................................................ 138
I. Kinematics of the knee, implanted with a bicruciate substituting
prosthesis ............................................................................................... 138
Chapter 5: General discussion and conclusion ........................... 151
Acknowledgements .................................................................................... 160
Appendix ............................................................................................ 162
1. List of abbreviations .............................................................................. 163
2. Anatomic denitions ............................................................................. 165
3. Professional career................................................................................. 166
ORTHOPAEDICS IS ALL ANATOMY,
PLUS A LITTLE BIT OF COMMON SENSE
Jack Hughston
7
CHAPTER 1: INTRODUCTION
I. Justication
Total knee arthroplasty (TKA) has evolved towards a reliable and long lasting surgical proce-
dure, offering pain relief and improved function to many patients who suffer from degenera-
tive arthritis of the knee. Improvements in surgical technique and prosthesis design continue
to enhance long-term outcomes
1
. The ability to reproducibly restore walking without pain
and limping in more than 90% of cases has led surgeons to offer this operation to younger
patients.
2
However, the functional demands and activity levels of these younger patients ex-
ceed our current capacity to restore the damaged knee joint. Instability and limited range of
motion are listed as important limiting factors to normal function, even in patients with a well
functioning TKA.
3
When looking at failed knee arthroplasties, instability and stiffness rank
in second and third place amongst the causes of early failure, only preceded by infection.
4,5
Range of motion and stability are two key factors in the success of total knee arthroplasty.
Both features relate intimately to the kinematics of the knee joint.
6-8
Kinematics can be altered
by trauma, degenerative surface deformation, and attrition of ligaments and capsule. Surgical
treatment should aim at restoration of the original kinematics in an attempt to restore original
function.
Several obstacles can prevent achievement of this goal. As total knee arthroplasty is a surface
replacement within an existing soft tissue envelope, alteration of form or size will induce
either stiffness or instability. Ligaments have a given strength and modus of elasticity, so they
should be loaded within physiologic boundaries. The surgical exposure routinely involves
cutting of the anterior cruciate ligament and damage to other ligaments, potentially leading
to instability. For the above-mentioned reasons, the implant that will resurface the damaged
knee must match as close as possible the original anatomy and preserve the ligaments or
substitute for lost ligament function.
Study of the kinematics of the native human knee and evaluation of ligament isometry are
crucial in the quest for improvement of knee surgery and total knee arthroplasty. This work
aims at a close comparison between the behaviour of the native and the replaced knee joint
with special emphasis on the critical factors of alignment, ligament isometry, and kinemat-
ics. Making use of a kinematic simulator, these parameters will be studied and compared
in the experimental ex vivo setting, before and after the implantation of a prosthesis. The
experimental ex vivo methodology will be validated in performing independent analyses and
comparing outcomes to existing knowledge. Additional, in vivo work will compare the kine-
matics of a deep knee bend, with the obtained experimental data. A full description of aims
and hypotheses is given in chapter 2.
8
Chapter 1: Introduction
REFERENCES
1. Robertson O, Knutson K, Lewold S and Lidgren L: The Swedish Arthroplasty Register 1975-1997. An update
with special emphasis on 41.223 knees operated on in 1988-1997. Acta Orthop Scand 2001; 72(5): 503-513.
2. Kurtz S, Mowat F, Ong K et al: Prevalence of primary and revision total hip and knee arthroplasty in the United
States (1990-2002). J Bone Joint Surg 2005; 87(A): 1487-1497.
3. Noble PC, Gordon MJ, Weiss JM, et al: Does total knee replacement restore normal knee function? Clin Orthop
2005; 431:157-165.
4. Lavernia C, Lee DJ, Hernandez VH: The increasing nancial burden of knee revision surgery in the United
States. Clin Orthop 2006; 446:221-226.
5. Sharkey PF, Hozack WJ, Rothman RH et al: Why are TKAs failing today? Clin Orthop. 2002; 404:7-13.
6. Victor J, Banks S and Bellemans J: Kinematics of posterior cruciate ligament-retaining and substituting total
knee arthroplasty. J Bone Joint Surg 2005; 87B: 646-655.
7. Victor J and Bellemans J: Physiologic Kinematics as a Concept for better Flexion in TKA. Clin Orthop 2006;
452: 53-58.
8. Bellemans J, Banks S, Victor J et al: Fluoroscopic analysis of the kinematics of deep exion in total knee arthro-
plasty. Inuence of posterior condylar offset. J Bone Joint Surg. 2002; 84:50-53.
9
II. Kinematics of the Native Knee
A. HISTORICAL INSIGHTS
The description of the relative motion between rigid bodies is called kinematics. The study of
how the femur moves on the tibia has a long history. The rst known description dates back
to 1836 when Weber and Weber
1
described the movement on the medial side to be like a
cradle. Since that rst description, based upon direct visual observation of a cadaveric speci-
men, several methods have been used to examine the kinematics of the human knee.
The rst radiological study was performed by Zuppinger
2
, who stated that the femur rolled
back across the tibia during exion as a result of the so-called rigid four-bar link mechanism
provided by the two cruciate ligaments. In 1971, Frankel
3
introduced the concept of the in-
stant centre of rotation to the orthopedic community. He emphasised as one link (rigid body)
rotates around the other, there is at any given moment in time a point with zero velocity. That
point is called the instant centre of rotation (g 1).
He concluded that due to the shapes of the bones and the restraints on motion imposed by
the ligaments, capsule, and muscles, the instantaneous centres for successive positions of the
links of the knee move.

(A) (B) (C)
Fig 1 The instant centre of rotation determined with the method of Reuleaux.
3
The points on the body move from
position 1 to position 2, and from 2 to 3. The perpendicular bisector for a pair of displacement points is
drawn. The intersection of the bisectors is the point with zero velocity, called the instant centre of rotation.
The contact point is represented by the blue arrow.
A. For a circular body undergoing rotation with gliding, the instant centre of rotation and the contact point
remain xed
B. For a non-circular body undergoing rotation with gliding, the instant centre of rotation remains xed but
the contact point shifts
C. For a circular body undergoing rotation with rolling, the instant centre of rotation coincides with the
contact point and moves accordingly
The initial work by Frankel was carried out by taking true lateral X-rays of the knee in
patients lying on the side, at discrete intervals of 10 to 20 in the range of full extension to
90 of exion. The knee was treated as if it were a planar mechanism. In other words, the
movement of the knee was reduced to a two-dimensional projection of a three-dimensional
reality (g 2). Menschik introduced in 1974 the concept of the four bar linkage, represent-
10
ing the cruciate ligaments in two dimensions as two rigid bars and the lines connecting their
insertion points on femur and tibia as the two other rigid bars
4
. The instant centre of rotation
is located at the point of intersection of the cruciate ligaments. This rigid four bar linkage
was later widely popularized by Mller.
In the years following, the limitations of this methodology became clear, with the major aw
being the inability to ascertain the location of the axes of rotation before performing kine-
matic analyses.
5
In 1983, Grood and Suntay presented a joint coordinate system providing a
geometric description of the three-dimensional rotational and translational motion between
two rigid bodies, applied to the knee joint. The main step forward was the denition of a
oating axis. With this model, the described joint displacements became independent of the
order in which the component rotations and translations occur.
6

Fig 2 The moving instant centre of rotation of the knee, as
described by Frankel
3
in a two-dimensional model
The new mathematical insights led to the concept of the helical axis and opened the door for
a correct scientic description of the kinematics of the knee.
7
However, as the mathematical
accuracy improved, the complexity increased and the model appeared to be impractical and
difcult to apply to the clinical setting: the clinicians failed to understand the engineers.
Hollister, and later Churchill, tried to bridge the gap.
8,9
Hollisters model essentially described
knee motion as pure rotations occurring around two axes: the so called exion-extension
axis and the so called longitudinal rotation axis, with the understanding of the exion-
extension axis not being exactly located in the coronal plane and the longitudinal axis not
being exactly located in the sagittal plane.
8
As a consequence, these mathematical simple ro-
tations meant in reality exion-extension, varus-valgus and internal-external rotation of the
knee joint, once again confusing the clinician trying to apply this knowledge to the practical
setting. Churchill addressed this problem by allowing a mathematical error in the kinematic
description, based on a loaded rig experiment with an ankle load of 100 N and a combined
hamstrings load of 30N.
9
They used an optimization technique to identify the locations of the
so-called optimal exion and longitudinal rotation axes. Knee motion was then described
with the following formula:
K = !
OF
+ !
LR
+R
!
+R
X
+R
Y
+R
Z
, where K= complete three-dimensional motion, !
OF
= rotation
about the optimal exion axis and !
LR
= rotation about the longitudinal rotation axis. He con-
cluded that the optimal exion axis coincided with the transepicondylar axis if one accepts
the following errors and limitations: residual rotation (R
!
)=2,9, residual translations (R
X
+R
Y

+R
Z
)= 3,4 mm and applicable motion range 5 to 90 of exion. Despite those limitations,
the advantage of this approach was the link between the kinematic description and certain
anatomic landmarks, allowing clinicians to apply this knowledge in practice.
11
In recent times, technological progress allowed more advanced tools to be used, including
ex vivo studies with MRI on cadaveric specimens
10,11
, in vivo analyses using 2D uoroscopy
with shape matching techniques, based on CT models
12-15
, roentgen stereo photogrammetric
analysis
16
and open dual coil MRIs
17-19
. These newer methods revealed a more complete,
three-dimensional insight in the morphology and kinematic patterns of the normal knee in
loaded and unloaded conditions.
B. MRI STUDIES
Iwaki and coworkers applied MRI scans to six unloaded cadaver specimens for the descrip-
tion of the surface geometry and relative passive movements of the femur and the tibia
10
. They
visualized an approximate mid-sagittal section of the medial and lateral femoral condyle and
analysed the contours of the bones. On the medial side of the femur, in the mid-sagittal sec-
tion, the arc of two circles was described: the anterior or extension facet and the posterior
or exion facet. In their description, the mid-sagittal section of the medial side of the tibia
consisted of an extension facet, a exion facet and a posterior horizontal surface. The exten-
sion facet, accommodating the anterior horn of the medial meniscus, sloped upwards by 11
relative to the exion facet, dened as the area of the tibia that is in contact with the femur
from approximately 20

to 120 of exion. The posterior horizontal surface accommodated
the posterior horn of the medial meniscus. The mid-sagittal section of the lateral femoral con-
dyle was found to be circular in its posterior part as on the medial side with a radius being 1
mm smaller. No clear extension facet was described on the lateral femoral condyle. The mid-
sagittal section of the lateral tibia plateau essentially revealed a central area of 24 mm depth,
called the tibial articular facet. Anterior and posterior to this area, the tibia accommodated the
anterior and posterior horn of the lateral meniscus.
The main kinematic observation was the centre of the femoral exion facet (medial condyle)
remaining relatively constant along the X-Y axis in the sagittal plane, suggesting almost
pure sliding of the condyle from 0-110 of exion. The initial (between 10 and 30) rela-
tive posterior movement of the medial tibiofemoral contact area was described as a rocking
motion as the contact shifts from the tibial extension facet onto the exion facet. From 110
to 120 of exion, a posterior motion of the centre of the femoral exion facet of 2 mm was
described. On the lateral side, rolling and sliding took place simultaneously, translating the
lateral femoral exion facet centre 4 mm between full extension and 45 and 15 mm between
45 and 120.
In a later paper in 2004, Pinskerova
20
reworked these ndings with the focus on the different
patterns of translation of the tibiofemoral surface contact points and the condylar centers.
Limitations of these studies include the MRI does not yield a homogenous Cartesian coordi-
nate set, the denitions of the exion facet centers in two dimensions only and the discrete
measurements with the absence of data between 45 and 90 of exion. The same group
17

studied MRI data of 13 knees in living volunteers performing unloaded exion. On the medial
side, the AP position of the femoral condyle remained relatively stable, from full extension to
110 of exion, provided that adjustment was made for the 2 mm translation during the rock-
ing motion on the extension facet between -5 to 30 of exion. On the lateral side, 3 mm
12
posterior motion was observed as the knee was released from the screw-home mechanism,
1 mm posterior motion between 0 and 60 and 13 mm between 60 and 110 of exion. It
was concluded that the kinematics of unloaded exion of the knees of the living subjects was
comparable to the kinematics observed in the cadaver specimens. In four subjects only, a
loaded exion movement under form of a squat from 0 to 90 of exion was examined. The
main difference with the unloaded movement was the anterior translation of the medial femo-
ral condyle by 4 mm. Two additional tests were performed: isometric hamstrings contraction
at 90 of exion caused no change medially and an anterior translation of the lateral femoral
condyle of 2 mm. Forced internal rotation of the tibia did not change kinematics between 0
and 90 of exion, whereas forced external rotation of the tibia suppressed the normal rota-
tional kinematics and forced the knee to ex as would a hinge, without axial rotation. Limita-
tions of this study are equal to the previous study with the additional drawback of the small
sample size and limited exion (up to 90 only) of the loaded exion measurement.
Nakagawa and co-workers
18
completed the previously described work in the high exion arc,
studying unloaded exion in 20 Japanese volunteers with an open MRI. Active exion was
measured from 90-133 and passive exion from 90 to 162. They found a mean posterior
translation of the medial condyle and lateral condyle of 2 mm and 13 mm respectively from
90-130 of exion. Pushing the knee further, from 133 to 162 of exion, caused further
posterior translation, medially by 4.5mm and laterally by 15 mm, subluxing the femur behind
the tibia.
Johal and co-workers
19
studied the full range of motion of a loaded squat (wall supported
squat sit) with 10 volunteers, using an interventional MRI. They observed posterior transla-
tion of the lateral femoral condyle of 22 mm from hyperextension to 120 of exion. The
medial condyle moved forward 1.7 mm from hyperextension to 30 of exion and started
translating posterior from 90 exion onwards: 3.6 mm between 90 and 120, and an ad-
ditional 8.4 mm from 120 to full exion. This differential translation on the medial and
lateral side, lead to an external rotation of 20 of the femur relative to the tibia during ex-
ion. Beyond 120 of exion, posterior translation was equal on the medial and lateral side
and no further tibiofemoral rotation was occurring. The subluxing nature of the tibiofemoral
articulation beyond 120 of exion inspired the authors to question the ability of total knee
replacement prostheses of the condylar type to achieve ranges of exion above 120 in a
physiologic manner.
C. RSA STUDIES
Karrholm et al
16
conrmed the data published by Hill in an in vivo RSA experiment where
volunteers performed a step-up activity from 50 to 65 of exion to full extension. They also
found anterior translation of the femoral condyle in early exion (3 mm from 0 to 50 of
exion). Forced external rotation of the foot suppressed physiologic internal rotation of the
tibia with exion. The main weakness of this study is the variability of the motion that was
performed and the fact that they only studied a single motion from moderate exion to full
extension and not a full motion cycle.
13
D. FLUOROSCOPY STUDIES
Fluoroscopy allows following knee motion in real time in vivo, with an obvious drawback of
delivering two-dimensional images. A preliminary CT scan allows making a 3D bone model
that can convert the 2D uoroscopic image with a shape-matching technique
12
. Lu and co-
workers studied active knee extension in open chain at xed angular speed in 8 subjects and
found no difference between the unloaded (ankle free) and loaded (5 kg attached at ankle)
setting. Half of the tibiofemoral rotation occurred between full extension and 30 of exion.
14

Komistek and co-workers used this technology to study gait and deep exion activities in
ve volunteers. They found signicantly less translation and rotation of the femur during
gait than during deep exion.
13
They conrmed previous ndings of less translation on the
medial side than on the lateral side but the most important result of their work seems to be
the impressive inter-individual variability. For deep exion, the medial condyle translation
ranged from +3 mm to -9 mm versus +1,4 to -30 mm for the lateral condyle. The average
tibiofemoral rotation during exion was 13.
Banks group used CT-derived bone models for model registration and added MRI derived
articular surfaces for obtaining higher accuracy of the contact areas. They observed the great-
est femoral external rotation during the squat activity, but reported no posterior subluxation
of either femoral condyle in maximum knee exion.
12
In comparing kneel, squat and stair
climb motions, they found knee kinematics to vary signicantly by activity.
E. CONCLUSION
On the basis of the published data one can conclude the normal kinematic pattern of the
human knee consists of internal rotation of the tibia relative to the femur with increasing
exion, following a greater posterior translation of the lateral femoral condyle than of the
medial femoral condyle. However, different methodologies seem to discover different kine-
matic patterns. Multicenter ex vivo cadaver studies and in vivo studies on normal volunteers
clearly demonstrate a signicant inter-individual and activity dependent variability: the mean
values fail to disclose the full story. It appears that rotational patterns are optional and under
control of the forces imposed on the joint by foot position, body inertia and muscular action.
An important clinical question to be raised is how surgical repair of the damaged knee joint
can respect the specic individual anatomic and geometric characteristics and still allow for
sufcient kinematic freedom to adapt to the variety of external conditions imposed on the
loaded knee joint.
14
REFERENCES
1. Weber WE, Weber E. Mechanics of the human walking apparatus. Translated by Maquet P and Furlong R Berlin
etc: Springer-Verlag, 1992: 75 (original publication: Mechanik der menschelichen Gehwerkzeuge. Gttingen,
1836).
2. Zuppinger H. Die aktive exion im unbelasten Kniegelenk: Zricher Habil Schr. Wiesbaden: Bergmann, 1904:
703-763.
3. Frankel VH, Burstein AH and Brooks DB. Biomechanics as determined by analysis of the instant centers of
motion. J Bone Joint Surg 1971; 53-A:945-977.
4. Menschik A. Mechanik des Kniegelenks. Teil 1 Z Orthop. 1974; 112:481-495.
5. Soudan K, Van Audekercke R, Martens M. Methods, difculties and inaccuracies in the study of human joint ki-
nematics and pathokinematics by the instant axis concept. Example: the knee joint. J Biomech 1979;12:2733
6. Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: ap-
plication to the knee. J Biomech Eng. 1983; 105:136-144.
7. Blankevoort L, Huiskes R, de Lange A. Helical axes of passive knee joint motions. J Biomechanics. 1990;
23:1219-1229.
8. Hollister AM, Jatana S, Singh AK, Sullivan WW, Lupichuk AG. The axes of rotation of the knee. Clin Orthop
1993; 290:259-268.
9. Churchill DL, Incavo SJ, Johnson CC, Beynnon BD. The transepicondylar axis approximates the optimal ex-
ion axis of the knee. Clin Orthop 1998; 356:111-118.
10. Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement 1: the shapes and relative movements of the
femur and the tibia in the unloaded cadaver knee. J Bone Joint Surg. 2000:1189-1195
11. Eckhoff DE, Hogan C, DiMatteo L, et al: Difference between the epicondylar and cylindrical axis of the knee.
Clin Orthop 2007; 461:238-244.
12. Banks SA, Hodge WA: Accurate measurement of three-dimensional knee replacement kinematics using single-
plane uoroscopy. IEEE Trans Biomed Eng 1996; 43:638-649
13. Komistek RD, Dennis DA, Mahfouz M. In vivo uoroscopic analysis of the normal human knee. Clin Orthop.
2003; 410:69-81.
14. Lu TW, Tsai TY, Kuo MY, Hsu HC, Chen HL. In vivo three-dimensional kinematics of the normal knee dur-
ing active extension under unloaded and loaded conditions using single-plane uoroscopy. Med Eng Phys.
2008;30:1004-1012.
15. Moro-oka T, Hamai S, Miura H, Shimoto T, Higaki H, Fregly BJ, Iwamoto Y, Banks SA. Dynamic acitivity
dependence of in vivo normal knee kinematics. J Orthop Res. 2008; 26:428-34.
16. Karrholm J, Brandsson S, Freeman MAR: Tibiofemoral movement 4: changes of axial rotation caused by forced
rotation at the weight bearing knee studied by RSA. J Bone Joint Surg 2000; 82-B:1201-1203.
17. Hill PF, Vedi V, Williams A, Iwaki H, Pinskerova V, Freeman MAR: Tibiofemoral movement 2: The loaded and
unloaded living knee studied by MRI. J Bone Joint Surg 2000; 82-B:1196-1198.
18. Nakagawa S, Kadoya Y, Todo S, Kobayashi A, Sakamoto H, Freeman MAR, Yamano Y. Tibiofemoral move-
ment 3: full exion in the living knee studied by MRI. J Bone Joint Surg. 2000; 82-B: 1199-1200.
19. Johal P, Williams A, Wragg P, Hunt D, Gedroyc W. Tibio-femoral movement in the living knee. A study of
weight bearing and non-weight bearing knee kinematics using interventional MRI. J Biomech. 2005; 38:269-
276.
20. Pinskerova V, Johal P, Nakagawa S, et al: Does the femur roll-back with exion? J Bone Joint Surg 2004; 86-B:
925-931.
15
III. Kinematics of the Replaced Knee
A. RELATION BETWEEN KINEMATICS AND IMPLANT DESIGN
J. Victor and J. Bellemans: Physiologic kinematics as a concept for better
exion in TKA. Clin Orthop. 2006; 452: 53-58.
Abstract
Functional outcome after total knee arthroplasty is determined by strength, stability and
range of motion. Flexion in the replaced knee is suboptimal for many patients and kinemat-
ics after total knee arthroplasty is abnormal. The relation between kinematics of the replaced
knee and postoperative exion is analysed and compared to normal knee kinematics. Specic
characteristics that relate to better exion are dened: posterior condylar offset, femoral roll-
back and external femoral rotation during exion relate to a better post-operative range of
motion. A rationale for a guided motion knee arthroplasty is developed and positioned within
the current state of the art knowledge on total knee arthroplasty.
Introduction
Total knee arthroplasty (TKA) has evolved to a successful operation with acceptable longev-
ity helping patients with incapacitating arthritis of the knee joint. Over the years, the success
of this procedure has been measured mainly by means of survivorship analysis.
1
Survivor-
ship measures durability but does not reect the functional results of the procedure. Despite
the introduction of functional parameters in the knee society rating system
2
, we believe the
information on the functional outcome of TKA is inadequate. The function score as part of
the knee society rating system has several limitations. It only covers gait and the ability to
go up and down stairs, ignoring a number of daily activities of importance to the patient.
2
In
addition, it uses a scoring system that has a ceiling effect because of the linear relationship
between walking capacity and the number of points allocated. Consequently, all patients who
are able to walk 2 km and go up and down stairs are positioned within the good or excellent
group withour further stratication. Lastly, there is insufcient information as to what mat-
ters for the patient in terms of functional capacity.
3
Another reason for the apparent lack of information on functional outcomes is validated
functional outcome scores often are not used in traditional follow-up studies. A structured
literature search from 1995 to April 2003 of English-language databases even concludes that,
TKA is a generally effective procedure, but the current English-literature does not support
specic recommendations about which patients are most likely to benet from it.
4
However, a recent study by Noble et al
5
looked specically at functional activities that play
an important role in daily life and compared a group of 243 TKA patients versus 257 in-
dividuals (matched for age and gender) without previous knee disorders. They concluded
activities that placed few demands on the knee in terms of either muscle control or ROM
were equally important to both groups. More demanding activities like squatting, dancing,
carrying heavy objects, kneeling, turning and cutting, were avoided by TKA patients.
5
Based
16
on the experience in the outpatient clinic and the description of the activities, we assume four
main factors play a role in functional decits: range of motion, intrinsic stability, strength
and proprioception.
3
Although the original arthritic disease accounts partially for the decit
in the form of irreversible sequelae, the kinematic abnormality of the replaced knee accounts
for the other part.
Of all functional parameters, range of motion (ROM) is most often quoted by patients and
orthopaedic surgeons. Reliable information on range of motion can be found in uoroscopic
studies that measure motion in an accurate and reproducible way.
6-14
Fluoroscopy in multi-
center studies suggest a substantial limitation of the weightbearing ROM of the replaced knee
with typical values between 100 and 110.
7,9,11,15
Because these studies include many different surgeons and implant designs, it may be as-
sumed that this is a limitation of our current technology. There is an ongoing discussion as to
whether the surgical trauma or the prosthesis design plays a role in this functional decit.
We explore evidence for the hypothesis that more normal kinematics will lead to better
exion in the replaced knee. Results of studies relating dynamic in vivo kinematic features
against a control group of posterior stabilized TKAs were analysed. The characteristics al-
lowing exion in the physiologic knee joint were compared to kinematic patterns in the
replaced knee. We analysed kinematic abnormalities and related these to current limitations
of total knee arthroplasty.
Flexion of the normal knee
Flexion of the human knee occurs along the six degrees of freedom in space and includes
rotation along the horizontal axis (exion), translation along the sagittal axis (roll-back of the
femur) and rotation over the coronal axis (femoral external rotation). This concept of coupled
motion is a sine qua non for a physiologic exion arc.
16
The motion is guided mainly by the
cruciate ligaments and the surface geometry of the tibial plateau with its concave and stable
medial side versus a convex and sloped lateral side (g 1). The roll-back is greater on the
lateral side that on the medial side.
10,17
On the lateral side, this clears the back of the knee in
deep exion. On the medial side, the greater condylar offset allows the posterior femur to
accommodate the posteromedial tibia without impingement.
8
The rotational behavior in the
horizontal plane also affects patellofemoral tracking. The internal rotation of the tibia relative
to the femur reduces the Q-angle with increasing exion, thereby stabilizing the patella.
Fig 1 Differential anatomy of the tibial
surface of the human knee with a concave
medial and a convex lateral plateau is
shown.
17
Flexion of the replaced knee
As in the normal, physiologic knee joint, the motion is guided by the surface geometry. The
cruciate ligament system is absent because the anterior cruciate ligament (ACL) is almost al-
ways sacriced. If the posterior cruciate ligament is retained, it can offer a constraint against
sagging of the tibia, but in the absence of its anterior counterpart it cannot guide the motion as
in the normal knee.
12
Other systems substitute the posterior cruciate ligament under the form
of an interaction between the polyethylene insert and the femoral component: the cam-post
mechanism. None of the current systems substitute for the ACL: therefore, the guidance of
the motion is limited to a forced roll-back in exion in those so called posterior stabilized
systems. In addition, these systems are traditionally designed for symmetric roll-back, giv-
ing equal posterior translations on the medial and lateral sides. If the knee follows its normal
rotational pattern in the horizontal plane, with asymmetric femoral roll-back, this can cause
a cam-post conict and post wear.
18
Kinematic patterns of the replaced knee have been well
documented in uoroscopic in vivo studies.
6-14,19
Some patterns can be related directly to
an important functional parameter that can be measured as a hard endpoint: exion of the
replaced knee.
Posterior condylar offset
The posterior condylar offset has been described as a determinant for exion.
8
Based on the
analysis of 150 consecutive posterior cruciate ligament retaining TKA patients, a correlation
between the operative restoration of the posterior condylar offset and the maximum exion
of the knee was shown. A mean reduction in exion of 12 was found with every 2 mm de-
crease in offset. This phenomenon can be explained by inadequate clearance in the back of
the knee, with resulting impingement of the posterior polyethylene rim against the back of
the femur. The posterior soft tissues can be caught between, leading to posterior knee pain.
The restoration of this offset relates directly to surgical technique and prosthesis design. An-
terior referencing with downsizing of the femoral component is a common cause and should
be avoided. The authors recommend posterior referencing with accurate restoration of the
posterior condylar offset. This surgical technique has a trade-off: in order to accommodate
the front of the femur without running the risk of notching or overhang, many femoral sizes
of the implant are needed with intervals of 2 to 3 mm in the sagittal plane.
Femoral roll-back
The posterior translation of the femur relative to the tibia is another determinant of exion
in the replaced knee. In synergy with the restoration of the posterior condylar offset, femo-
ral roll-back helps to clear the posterior aspect of the knee joint, especially on the lateral
side.
8,10,11,14
In addition, it has been shown femoral roll-back increases the lever arm of the
quadriceps muscles at a critical phase of the stair-climbing cycle and provides increased me-
chanical advantage for the quadriceps muscle to extend the knee.
20
In a prospective random-
ized trial comparing cruciate retaining versus cruciate substituting TKA, a difference in ki-
nematic behavior was observed.
14
Forward sliding of the femur during exion, the so-called
paradoxical motion, occurred mainly in cruciate retaining knees and more on the medial than
18
on the lateral side. Similar ndings were reported in a matched pairs study comparing cruci-
ate retaining and posterior stabilized TKAs
14
and in a multi-center summation analysis.
10
Forward sliding of the femur during exion can be attributed to several mechanisms. In ex-
tension, the starting position of the femur may be overly posterior in relation to the tibia (g
2). Near full extension, the quadriceps pull has a forward-directed vector component that is
countered in the normal knee by the ACL. Absence of the ACL allows the tibia to slide for-
ward, putting the femur in a too posterior starting position. During the exion arc, forward
sliding of the femur will occur as a compensatory mechanism. This phenomenon is observed
in ACL decient knees as well as in all TKAs that do sacrice the ACL. Second, in case of
insufcient posterior cruciate ligament function of the cruciate retaining knee, the tibia is
pulled backwards by the co-contraction of the hamstrings, allowing the femur to continue its
forward slide. This phenomenon is only observed in cruciate retaining and not in posterior
stabilized knees as the cam-post mechanism prevents this.
14
A correlation between the antero-
posterior positions of the femur relative to the tibia was observed. The replaced knee lost
1.3 of exion with every mm of forward sliding of the femur. This observation is found in
several other studies with similar magnitudes in different designs.
7,8
It is a conrmation of the
relationship between kinematic pattern and the outcome of a functional parameter.
A. B. C.
Fig 2 Tibiofemoral relationships near full extension in the sagittal plane. (A) Physiologic relationship with the
posterior condyle lined up with the posterior border of the tibia. (B) Abnormal relationship after TKA due to
anterior instability and nonanatomical surface geometry of tibial and femoral implant. (C) Correct relation-
ship after TKA with a stable implant and physiologic sagittal geometry.
Femoral external rotation
Abnormal rotational behavior has been described in several papers dealing with uoroscopic
analysis of the replaced knee.
6-14
The abnormal pattern is more outspoken in cruciate retain-
ing than in posterior stabilized knees. The center of rotation in the horizontal plane is on the
medial side in the normal knee, whereas it is highly variable in the replaced knee.
14
To our
19
knowledge, no study is available relating functional outcome with axial rotation between
femur and tibia during exion. A theoretical advantage for the patellofemoral mechanism
is apparent. Femoral external rotation during exion reduces the Q-angle and patellar shear
force and patellofemoral joint reaction force decreases. All modern posterior stabilized knee
designs induce femoral roll-back equally on the medial and lateral sides. This mechanism
should deliver a centre of rotation in the horizontal plane in the middle of the knee. That ki-
nematic studies on posterior stabilized knees show a more medial center of rotation
7,10,11,13,14,19
can be explained by the isometric nature of the medial collateral ligament that will not allow
too much roll-back on the medial side and forces the replaced knee more in the direction of
the original kinematic pattern. A second disadvantage of symmetric roll-back is a potential
conict in the front of the knee. Impingement of the patellar tendon with the front of the poly-
ethylene insert and contact between the patella and the post have been described.
21
Discussion
As total knee arthroplasty is a surface replacement within the existing soft tissue sleeve, it
functions within normal anatomic and physiologic boundaries. Impaired functionality after
TKA is attributed to sequelae of the arthritic disease, the surgical trauma and the design of the
prosthesis. Recent information on the outcome of minimally invasive procedures suggests
the reduction of the surgical trauma offers early improvement and faster rehabilitation.
22-24

This effect levels off after 3 months to a result similar to that in patients who had a standard
exposure.
22-23
This means factors other than exposure and extensor mechanism violation are
involved in the reduced functionality after TKA.
Aberrant kinematics including abnormal tibiofemoral rotation in the horizontal plane and
forward sliding of the femur on the tibia, the so-called paradoxical motion, have been dem-
onstrated in TKA patients.
6-14,19
Forward sliding of the femur on the tibia negatively affects strength because of the shorter
moment arm between the tibiofemoral joint reaction force and the force exerted by the patel-
lar tendon on the tibial tuberosity. The relationship between kinematic behavior and knee
exion is the second link to postoperative function. A direct relationship between a kinematic
parameter and the functional outcome is hard to conrm since few papers have related kine-
matic features to clinical outcomes. A second limitation of current kinematic studies using
in vivo uoroscopic analysis is they typically study small, selected groups of patients and
no control group. Also, the typical activities studied (gait, deep knee bend, step-up) do not
necessarily represent the spectrum of kinematic conditions patients routinely perform. In one
randomized controlled trial comparing between CR and PS knees, the uoroscopic analysis
was directed to a subgroup of patients.
14
Other studies have reported matched pairs
7,11,13,19
in
comparing CR and PS knees. All studies
7,11,13,14
but one
13
related the more natural kinematics
of the PS group to better exion. In one paper, the number of investigated patients was high
enough (40) to show statistical signicance.
11
The highest number of patients was enrolled in
a multicenter summation analysis.
10
The authors refer to the smaller magnitudes of femoral
rollback during deep exion as a main reason why knee exion is reduced after TKA. One
group compared xed and mobile bearing PS knees and found normal axial patterns of rota-
tion but decreased femoral rollback as compared to the normal knee.
22
Maximum weightbear-
ing exion was not mentioned. The comparison of kinematic data on the replaced knee with
20
the characteristics of the normal knee supports the rationale for a knee implant that envisions
a better kinematic performance: physiologic rotation in the horizontal plane, anatomic pos-
terior offset and physiologic roll back. Differential medial and lateral surface geometry (g
3A, B) and guidance by the cruciate ligaments or an intrinsic asymmetric cam-post mecha-
nism (g 3C) can provide this (Ries MD, Victor J, Bellemans J, Otto J, McKinnon B, Parikh
A, Sprague J, Salehi A. Effect of guided knee motion and high exion TKA on kinematics,
implant stresses and wear. Proceedings of the AAOS 2006 Annual Meeting, Chicago Illinois,
2006).
25
Preserving the anterior and posterior cruciate ligament has been attempted and re-
sulted in more normal kinematics and better exion.
6,13,15
Surgical feasibility limits the poten-
tial of this solution. The other solution is to provide kinematic guidance and stabilization by
the prosthetic design through surface geometry and the cam-post mechanism.
25

A. B. C.
Fig 3 Internal rotation of the tibia relative to the femur during exion can be induced by (A, B)
asymmetric surface geometry and (C) an asymmetric cam-post mechanism. This reduces the Q
angle as the knee moves into exion. Also note the similarities in surface geometry design as
compared to the normal anatomy in g 1.
Some authors fear intrinsic motion guidance and a more anatomic tibial geometry could lead
to accelerated wear.
26
These authors generally prefer the original symmetric toroidal design
with its inherent stability, surgical reproducibility, and wear characteristics. They suggest as
long as designs do not try to reproduce normality in kinematics, wear will not be an issue.
26

As polyethylene wear should be a primary concern with every new development in knee
arthroplasty, it is worth taking a closer look at the multifactorial process leading to failure.
First, osteolysis in TKA occurred with the advent of modularity in total knees.
27
Backside
wear was recognized as a main source of particles in the replaced knee. Improvements have
since been made and the solution of a monobloc tibial component is not incompatible with a
more kinematic knee design. Second, it has been recognized multidirectional sliding is det-
rimental for polyethylene.
28
However, a guided kinematic pattern cannot be compared to the
uncontrolled multidirectional sliding pattern of the unstable total knee; the sliding velocity
is lower than in the unstable knee and the pattern of motion is unidirectional. Third, the poor
polyethylene performance in knee arthroplasty may be related to inadequate quality control
in manufacturing, gamma-irradiation-in-air and shelf aging.
29
Excellent long-term wear per-
formance in a total knee design that offers little conformity but incorporates compression
21
molded vacuum sterilized polyethylene supports the idea a more kinematic knee is possible
from a material fatigue and wear standpoint.
30
This is conrmed by recent work on implant
stress and wear in a kinematic knee design, using dynamic nite element analysis and a vir-
tual Oxford Rig (Ries MD, Victor J, Bellemans J, Otto J, McKinnon B, Parikh A, Sprague J,
Salehi A. Effect of guided knee motion and high exion TKA on kinematics, implant stresses
and wear. Proceedings of the AAOS 2006 Annual Meeting, Chicago Illinois, 2006).
Conclusion
Numerous in vivo uoroscopic studies in patients who had TKA have shown abnormal ki-
nematic patterns as compared to the normal human knee.
6-14,19
Some factors have been cor-
related with inferior function after TKA. There is emerging evidence achievement of better
kinematic patterns will help the patients in their functional performance. As the role of the
kinematics is recognized, it can be expected new surgical techniques and implant designs
will be developed to better serve the functional needs of the patient. The main caveats are
surgical feasibility of performing the operation in a reproducible way and with durable mate-
rials. Close clinical followup and better functional monitoring will be needed to support the
hypothesis of better functionality through kinematic normality.
REFERENCES
1. Murray DW, Carr AJ, Bulstrode C. Survival analysis of joint replacements. J Bone Joint Surg Br. 1993;75:697
704.
2. Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat
Res. 1989;248: 1314.
3. Witvrouw E, Victor J, Bellemans J. Rock B, Van Lummel R, Van Der Slikke R, Verdonk R. A correlation
study of objective functionality and WOMAC in total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc.
2002;10:347351.
4. Kane RL, Saleh KJ, Wilt TJ, Bershadsky B. The functional outcomes of total knee arthroplasty. J Bone Joint
Surg Am. 2005;87: 17191724.
5. Noble PC, Gordon MJ, Weiss JM, Reddix RN, Conditt MA, Mathis KB. Does total knee replacement restore
normal knee function? Clin Orthop Relat Res. 2005;431:157165.
6. Banks SA, Fregly BJ, Boniforti F, Reinschmidt C, Romagnoli S. Comparing in vivo kinematics of unicondylar
and bi-unicondylar knee replacements. Knee Surg Sports Traumatol Arthrosc. 2005;13: 551556.
7. Banks SA, Markovich GD, Hodge WA. In vivo kinematics of cruciate-retaining and -substituting knee replace-
ments. J Arthroplasty. 1997;3:297304.
8. Bellemans J, Banks S, Victor J, Vandenneucker H, Moermans A. Fluoroscopic analysis of the kinematics of
deep exion in total knee arthroplasty: inuence of posterior condylar offset. J Bone Joint Surg Br. 2002;84:50
53.
9. Dennis DA, Komistek RD, Mahfouz MR. In vivo uoroscopic analysis of xed-bearing total knee replace-
ments. Clin Orthop Relat Res. 2003;410:114130.
10. Dennis DA, Komistek RD, Mahfouz MR, Haas BD, Stiehl JB. Multicenter determination of in vivo kinematics
after total knee arthroplasty. Clin Orthop Relat Res. 2003;416:3757.
11. Dennis DA, Komistek RD, Stiehl JB, Walker SA, Dennis KN. Range of motion after total knee arthroplasty: the
effect of implant design and weight bearing conditions. J Arthroplasty. 1998;13: 748752.
12. Stiehl JB, Komistek RD, Cloutier JM, Dennis DA. The cruciate ligaments in total knee arthroplasty: a kine-
matic analysis of 2 total knee arthroplasties. J Arthroplasty. 2000;15:545550.
13. Udomkiat P, Meng B, Dorr LD, Wan Z. Functional comparison of posterior cruciate retention and substitution
knee replacement. Clin Orthop Relat Res. 2000;378:192201.
14. Victor J, Banks S, Bellemans J. Kinematics of posterior cruciate ligament-retaining and -substituting total knee
22
arthroplasty: a prospective randomised outcome study. J Bone Joint Surg Br. 2005; 87:646655.
15. Cloutier JM, Sabouret P, Deghrar A. Total knee arthroplasty with retention of both cruciate ligaments: a nine to
eleven-year follow-up study. J Bone Joint Surg Am. 1999;81:697702.
16. Daniel DD, Akeson WH, OConnor JJ, eds. Knee Ligaments: Structure, Function, Injury and Repair. New York:
Raven Press; 1990.
17. Hill PF, Williams A, Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement 2: the loaded and un-
loaded living knee studied by MRI. J Bone Joint Surg Br. 2000;82:11961200.
18. Puloski SK, McCalden RW, MacDonald SJ, Rorabeck CH, Bourne RB. Tibial post wear in posterior stabilized
total knee arthroplasty: an unrecognized source of polyethylene debris. J Bone Joint Surg Am. 2001;83:390
397.
19. Ranawat CS, Komistek RD, Rodriguez JA, Dennis DA, Anderle M. In vivo kinematics for xed and mobile-
bearing posterior stabilized knee prostheses. Clin Orthop Relat Res. 2004;418:184190.
20. Andriacchi TP, Galante JO, Fermier RW. The inuence of total knee-replacement design on walking and stair-
climbing. J Bone Joint Surg Am. 1982;64:13281335.
21. Verborgt O, Victor J. Post impingement in posterior stabilised total knee arthroplasty. Acta Orthop Belg.
2004;70:4650.
22. Hart R, Janecek M, Cizmar I. Stipcak V, Kucera B, Filan P. Minimally invasive and navigated implantation for
total knee arthroplasty: x-ray analysis and early clinical results. Orthopade. 2006; 35:552557.
23. Laskin RS. Minimally invasive total knee arthroplasty: the results justify its use. Clin Orthop Relat Res.
2005;440:5459.
24. McGrory B, Callaghan J, Kraay M, Jacobs J, Robb W, Wasielewski R, Brand RA. Editorial: minimally in-
vasive and small-incision joint replacement surgery: what surgeons should consider. Clin Orthop Relat Res.
2005;440:251254.
25. Ries MD, Bellemans J, Victor J. The high-performance knee. In: Bellemans J, Ries MD, Victor J. Total Knee
Arthroplasty: A Guide to Get Better Performance. Heidelberg: Springer; 2005:303310.
26. Dorr LD. Contrary view: wear is not an issue. Clin Orthop Relat Res. 2002;404:9699.
27. Wasielewski RC, Parks N, Williams I, Surprenant H, Collier JP, Engh GA. Tibial insert undersurface as a con-
tributing source of polyethylene wear debris. Clin Orthop Relat Res. 1997;345: 5359.
28. Blunn GW, Walker PS, Joshi A, Hardinge K. The dominance of cyclic sliding in producing wear in total knee
replacements. Clin Orthop Relat Res. 1991;273:253260.
29. Bohl JR, Bohl WR, Postak PD, Greenwald AS. The Coventry Award: the effects of shelf life on clinical outcome
for gamma sterilized polyethylene tibial components. Clin Orthop Relat Res. 1999;367:2838.
30. Ritter MA, Berend ME, Meding JB, Keating EM, Faris PM, Crites BM. Long-term follow up of anatomic gradu-
ated components posterior cruciate-retaining total knee replacement. Clin Orthop Relat Res. 2001;388:5157.
23
B. RELATION BETWEEN KINEMATICS AND TIBIOFEMORAL ALIGNMENT
J. Victor. Rotational alignment of the distal femur: A literature review. Orthop
Traumatol Surg Res. 2009; in press doi: 10.1016/j.otsr.2009.04.011
Abstract
Malalignment still accounts for an unacceptable number of failures in total knee arthroplasty.
This section presents a literature review of previously published papers describing rotational
alignment of the distal femur and discussing different techniques in obtaining correct rota-
tional alignment of the femoral component in TKA. Based on the published values, the fol-
lowing mean angular relationships between the rotation axes of the distal femur in the axial
plane can be calculated: the posterior condylar line is on average 3 internally rotated relative
to the surgical trans-epicondylar axis (TEA), 5 relative to the anatomical TEA and 4 rela-
tive to the perpendicular to the trochlear anteroposterior axis. The greatest inter-individual
variability is described for the trochlear AP axis. The worst track record regarding inter- and
intra-observer variability is for the TEA.
Introduction
A condylar knee prosthesis functions as a surface replacement in the soft tissue envelope
that surrounds the knee. Consequently, positioning and sizing of the components will largely
affect the post-operative result. Any misplacement or wrong sizing will affect loads on the
interface and tension in the ligaments. This will lead to aberrant kinematic behaviour induc-
ing stiffness, instability and early loosening. Several clinical studies have demonstrated this
relationship. Skolnick and Coventry were the rst authors to report the relation between
clinical outcome and proper alignment in the coronal plane.
1
Lotke and Ecker were able to
demonstrate a signicant correlation between a good clinical result and a well-positioned
prosthesis, using a roentgenographic index describing alignment of the tibial and femoral
component in the three planes.
2
They concluded that more effort should be expended on the
development of better instruments to position the devices components, so that the operation
will become a precise, well engineered procedure and will not depend on gross, subjec-
tive perceptions of alignment and mechanical relationships. Since that time, multiple papers
have conrmed the relation between malalignment in the coronal plane and early failure of
TKA.
3-14
It was only later when more emphasis was directed towards rotational alignment of
the tibial and femoral component. This topic continues to challenge the surgeon in reducing
the number of outliers in horizontal plane position of the tibial and the femoral component.
This is reected in multiple recent studies, specically dealing with this subject.
Rotational alignment of the femoral component will affect exion stability, tibiofemoral and
patellofemoral kinematics, and alignment in exion. The relation between rotational align-
ment of the tibial and femoral component, and patellofemoral stability and function was
recognized in the early days of knee arthroplasty. Mochizuki and Schurman were the rst
to emphasize the detrimental effect of inverse rotational alignment on the patella and post-
operative function.
15
This was later conrmed in numerous papers.
16-22
Berger was the rst to
use CAT scans to evaluate the rotational alignment of the components.
23
He underlined the
24
clinical relationship between the femoral component in internal rotation and patellofemoral
complications. He compared thirty patients with isolated patellofemoral complications after
TKA to a group of twenty patients with well functioning TKA. The degree of patellofemoral
malfunctioning was directly related to the amount of component internal rotation: mild com-
bined internal rotation (1-4) caused lateral patellar tracking and patellar tilting, moderate
combined internal rotation (5-8) caused patellar subluxation and severe combined internal
rotation (7-17) caused patellar dislocation or component failure. He used the anatomic
transepicondylar axis and the tibial tuberosity as rotational landmarks for the femoral and
tibial component respectively. Similar outcomes were reported by Matsuda, using a different
prosthesis.
19
He reported a statistically signicant correlation of internal rotation of the tibial
and femoral component with the patellar tilt angle and clinical symptoms. Akagi compared
two consecutive groups, one with TKA implanted in neutral alignment with respect to the
posterior condylar line, and one with an external rotation of 3-5.
18
The externally rotated
group had less need for release of the lateral retinaculum (6% versus 34%) and postoperative
patellar tracking was signicantly better. Rhoads found in a laboratory study that internal
rotation and medial translation of the femoral component caused patellar maltracking, but
reported no predictable changes by externally rotating the femoral component. Anouchi per-
formed a cadaver experiment in changing femoral component position from neutral (accord-
ing to the posterior condylar line) to 5 internal and 5 external rotation. Patellar tracking
was closest to normal in the group that was externally rotated. Internal rotation caused severe
patellar maltracking to the medial side.
21

Traditionally, most emphasis has been put on the dangers of internally rotating the femoral
component. More recently, the detrimental effects of excessive external rotation of the femo-
ral component have been outlined. Olcott and Scott described symptomatic exion instability
resulting from the oversized exion gap on the medial side of the knee as a consequence of
excessive external rotation of the femoral component.
24
Miller and coworkers were able to
demonstrate increased shear forces on the patella as a result of the induced maltracking.
25

Finally, Hanada related the excessive external rotation of the femoral component to varus
alignment in exion, leading to mechanical overload on the medial side of the joint.
26
Despite everyone being convinced of the clinical importance of correct rotational alignment,
the dilemma of choosing the correct reference remains unsolved. Three obstacles interfere
with surgical consistency: the unsolved issue of the correct axis, the variability of the refer-
ence axes and the inter- and intra-observer variability in the intra-operative determination
of these references. On top of these practical concerns, confusion rules in terminology and
denition of index references.
There is a distinct difference between the correct rotational alignment of the femoral com-
ponent and the normal rotational alignment of the distal femur. This divergence can be
explained by the fact that the natural tibial plateau has an average varus orientation of about
3.
27,28
The perpendicular coronal cut of the tibia will change this angle. Consequently, the
femoral component will not be correctly aligned in exion if it follows the natural anatomy.
A rotational compensation to the same degree as the correction of the tibial cut in the coronal
plane will generally be advocated. Consequently, as the literature refers to correct rota-
tional alignment of the femoral component, the authors always refer to the adapted rota-
tional alignment of the femoral component, which is different from the normal situation. One
25
should keep this in mind when comparing normal knee kinematics with the replaced knee
kinematics.
It is yet unclear what is the best rotational reference to which all other parameters can be
compared. This consensus is needed to provide a common reference frame to work in and
refer to. Several authors have claimed the denition of this rotational reference.
The rotation of the femoral component can be described, relative to surface derived land-
marks on the distal femur or relative to its relation with the tibia. Distal femoral references
include the posterior condylar line (PCL)
27,29
, the anatomical transepicondylar axis (TEA)
30,31
,
the surgical TEA
23,32,33
, and the trochlear anteroposterior (AP)
34,35
axis. References relating to
the relative position of the tibia include the exion gap symmetry
36
and the tibial mechanical
axis alignment in exion.
26
Fig 1 Rotation reference axes in the distal femur.
Denitions:
The Antero-Posterior (AP) axis runs from the deepest point of the trochlea to the centre of the femoral
notch, with the femur viewed along its mechanical axis.
The Anatomical Trans-Epicondylar (TEA) axis runs from the medial to the lateral epicondyle
The Surgical TEA runs from the medial sulcus to the lateral epicondyle
The Posterior Condylar Line is the tangent line to the most posterior part of the femoral condyles with the
femur viewed along its mechanical axis.
The Femoral Transverse Axis connects the centres of the two best-t spheres to the femoral condyles. See
chapter 2, III.
26
Dorr et al
36
addressed the issue of rotational alignment and exion gap symmetry as early as
1986 but it was John Insall
37
who popularised the gap technique for obtaining a rectangular
exion gap that is equal to the extension gap. Freeman
38
introduced the use of a tenser to use
the tibial resected plateau as a basis for rotational alignment of the femoral component. The
importance of the rectangular gap was reiterated by Laskin, who stated that knees, left with
a trapezoidal exion gap had inferior clinical results compared with those with a rectangular
exion gap.
29
The rst scientic study addressing the accuracy of balancing the exion and
extension gaps was by Grifn.
39
Using the surgical TEA he was able to obtain a rectangular
exion gap in 89% of 104 consecutive TKAs. However, equality of the exion and extension
gaps was only obtained in 57% of the patients. The authors found a tendency for the exion
gap to be slightly looser on the lateral side, reecting physiologic laxity of the knee being
greater on the lateral side, as documented earlier by our group.
40

Pairwise comparison between reference axes and techniques
As appears from the above, all claimed rotational reference axes are relative axes compared
to a so-called secondary axis considered inferior by the author. Table I shows an overview
of all published relative relationships between these axes. A discussion of these studies is
given below, ordered in pairs that have been studied in their mutual relationship (g 1).
1. Tensioned gap technique versus PCL
In the tensioned gap technique, the knee is tensed in exion after ligamenteous release in
extension.
37,38,41
The tensing of the gaps can be performed manually or force controlled, with
laminar spreaders or a tensor device. The idea is to establish equal loads in the medial and lat-
eral compartment and to resect the posterior femoral surfaces parallel to the cut tibial surface.
Several authors have compared this technique to different methods of measured resection.
Laskin
29
compared the Posterior Condylar Line (PCL) to the rotational alignment obtained
by tensing the exion gap with laminar spreaders. For neutral and varus knees, he found a
very consistent relationship of 3.20.3 of external rotation, relative to the PCL, by tensing
the gaps. For knees with a coronal plane anatomical axis of more than 10 valgus, the values
were signicantly higher and less consistent: 10.14.2.
Fehring
22
compared the tensioned gap technique to measured resection. In 100 posterior sta-
bilised TKAs he used gap tensing with laminar spreaders and performed a rectangular gap
resection. Based upon the size of the resected posterior femoral bone, a virtual assessment
of the resection relative to the PCL+3 was made. The mean external rotation relative to the
PCL was 2.6 with a large variation (range -7/+8). He concluded that 45% of the patients
would have had a rotational error of at least 3 if the PCL +3 would have been used as a
reference. This is in contradiction with the results published by Laskin. These results have to
be considered in the light of the assumptions made by the author. First, he assumed that every
single knee was correctly balanced and that the tensioned gap technique yielded the correct
rotational alignment. As no post-operative assessment of the rotational alignment was made,
all conclusions depend on this assumption. Second, no compensation was made for existing
bone deciencies, despite the fact that many surgeons who use measured resection would do
this. Third, the series comprised both varus and valgus knees.
27
2. Epicondylar axis versus PCL
Mantas was the rst to measure the relationship between the PCL and the anatomic TEA. He
used 19 paired cadaveric femora and found a mean angle of 4.9, being consistent left and
right.
42
Grifn performed a clinical measurement during surgery on 107 arthritic patients.
He found a mean external rotation of the surgical TEA compared to the PCL of 3.7.
32
This
angle was smaller in patients with varus malalignment (3.3) than in patients with valgus
malalignment (5.4). As the standard deviation was greater than 2 in all groups, the authors
concluded the posterior condyles are potentially unreliable references for femoral component
rotation.
Berger
23
examined 75 embalmed anatomic specimen femurs with a calliper. He described
the medial sulcus as a clearly discernible, reproducible landmark and measured the angle
between the surgical TEA and the PCL. He found a mean value of 3.5 for the male speci-
mens and 0.3 for the female specimens. In comparing the anatomical TEA the angles were
respectively 4.7 and 5.2. It should be noted that the study involved 35 femurs of known
gender and 40 of unknown gender, assumed to have the same sex distribution. Yoshioka et al
examined 32 cadaveric femora and reported a small gender based difference: a condylar twist
angle of 5 in males and 6 in females.
31
No gender difference was reported in other studies.
Arima and Whiteside however were not convinced the epicondylar axis was a reliable ref-
erence.
34
In a cadaver study on 30 specimens they reported the mean angle between the
anatomical TEA and the PCL to be 4.4 but the range was excessive: -4.5 to 15.5 giving a
standard deviation of 2.9. These results improved a little when the landmarks were identied
on radiology instead of clinically but they were still not reliable.
Akagi used a CT scan to compare the pre- and post-operative relationship between the ana-
tomical TEA and the PCL. He found a mean value of 6.8 on the pre-operative scans of 26 pa-
tients undergoing TKA.
18
This angle was reduced after TKA (comparing with the prosthetic
PCL) to 3.21.7 in the group that underwent TKA by referencing off the PCL and adding
3-5 of external rotation to the instrument.
Grifn et al used MRI taken in patients with minor soft tissue pathology to examine the re-
lationship between the TEA and the PCL.
43
They described the medial epicondyle as a bony
ridge surrounding a central sulcus, present in all knees that were imaged. As the authors used
MRI, they drew the PCL according to the posterior cartilage, not to the bony border of the
condyles. The mean value of the posterior condylar angle was 3.111.75. Interestingly,
younger patients seemed to have a smaller angle than older patients. It was hypothesized this
might be due to posterolateral cartilage wear, increasing with age. This nding could not be
conrmed however, based on the data published in the study. No signicant differences be-
tween males and females were found. Matsuda examined the relation between the PCL angle
and varus deformity. As the angle between PCL and anatomic TEA was consistent in normal
and in varus knees (6.03 and 6 resp.), he concluded varus constitution is not associated with
posterior medial condyle dysplasia.
44
28
Table I Overview of the published mutual relationships between the surface derived axes.
Abbreviations:
Instrumented: instr
Anatomic trans-epicondylar axis: anat TEA
Surgical trans-epicondylar axis: surg TEA
Perpendicular to trochlear anteroposterior axis: !Troch AP
Posterior condylar line: PCL
Tensioned Gaps: Gap Tens
Osteoarthritic: OA
Patients: pts
Medial: med
Lateral: lat
29
3. Trochlear AP axis versus PCL
Arima and Whiteside stated the trochlear AP axis, dened as the line connecting the deepest
point of the trochlea to the centre of the notch, was a reliable rotational landmark.
34
The mean
angle with the PCL was 3.8 of external rotation. Still the range was high: -1 to 10 , SD 2.
Radiographic examination improved this to 3.1 with a range of 0.5 to 7. In a clinical paper
on the valgus knee, Whiteside reported less patellar complications if the trochlear AP axis
was used than when the PCL was used as a reference.
35
Nagamine investigated the reliability of the anatomical TEA and the trochlear AP axis versus
the PCL in patients with different types of arthritis and in volunteers with normal knees.
45
The
mean values were as follows: the PCL was 6.0 2.4 internally rotated, relative to the ana-
tomical TEA. The values for normal knees, medial tibiofemoral arthritis and patellofemoral
arthritis were 5.8 2.7, 6.2 1.9 and 6.4 2.4 respectively, being not signicantly dif-
ferent. In contrast, the angle between the line perpendicular to the trochlear AP axis and the
anatomical TEA showed a mean internal rotation of 1.4 3.3. The distinct groups of nor-
mal knees, medial tibiofemoral arthritis and patellofemoral arthritis displayed angles of 2.3
3.1, 0.1 3.3 and 1.3 3.3 respectively, showing a signicant difference between the
normal knees and the knees with medial patellofemoral arthritis. The authors concluded the
PCL is more reliable than the trochlear AP axis in knees with medial tibiofemoral arthritis.
4. Trochlear AP axis versus tensioned gap technique
Hanada et al. used 12 cadaveric knees to compare both techniques. In the rst six specimens,
a TKA was inserted using the perpendicular to the trochlear AP axis as a reference.
26
In the
second group of six specimens, the tensioned gap technique was used, creating an equal load
of 35N medially and laterally. Alignment was measured as the angle between the trochlear
AP axis and the projected extension of the tibial mechanical axis. This angle was 0.6 4
in the normal knees. In group 1, where the AP axis was used for rotational reference, the
postoperative angle was -0.5 0.2. In group 2 with the tensioned gap technique, the angle
was 8.5 3.3. The tibia shifted in varus as the knee went into exion. As a consequence,
peak pressures on the medial side of the knee were greater than on the lateral side upon axial
loading. Also, the patellar groove shifted laterally.
5. Anatomic versus Surgical TEA
Yoshino studied the relationship between the anatomic and the surgical TEA. In 48 patients
with osteoarthritis, a CT scan was performed prior to TKA. The medial sulcus could only
be determined in 30% of the knees. As the arthritis was more severe, the sulcus was more
difcult to locate. In those knees with a discernable sulcus, the angle between anatomic and
surgical TEA was 3.21. The angle between the anatomic TEA and PCL was 6.4 1.6,
between the surgical TEA and PCL it was 3 1.6. No relation between the progression of
disease severity and the condylar twist angle could be demonstrated.
30
Conclusion
Rotational alignment of the femoral component in total knee arthroplasty remains an impor-
tant challenge. The number of outliers in postoperative axial alignment has long been over-
looked, as measurement of axial alignment after TKA is not easy. It requires CT scan, scatter
reduction software and correct understanding of the reference axes. Nevertheless, as outlined
in the introduction, the clinical consequences of rotational malalignment are signicant and
often lead to important functional impairment or revision surgery.
Some conclusions can be drawn from a review of available publications in the literature. The
reported ranges and standard deviations are generally high, indicating an important inter-
individual variability. It is probably not wise to rely systematically on a single reference axis
or technique for every patient. Pitfalls and caveats related to specic situations have been
summarized in table II.
Based on the papers discussed, the following mean angular relationships between the rota-
tion axes of the distal femur in the axial plane can be calculated: the posterior condylar line
is on average 3 internally rotated relative to the surgical TEA, 5 relative to the anatomical
TEA and 4 relative to the trochlear AP Axis. The relation in the horizontal plane between the
above described surface derived axes and the kinematic optimal exion axis as described
by Eckhoff et al
47
, is of important clinical value and will be investigated and discussed in
Chapter 3, II.
Table II. Surgical caveats, distilled from the papers discussed, according to the technique or reference lines used.
31
REFERENCES
1. Skolnick MD, Coventry MB and Ilstrup DM: Geometric total knee arthroplasty. A two-year follow-up study.
J Bone Joint Surg 1976; 58:749-753.
2. Lotke P and Ecker M. Inuence of position of prosthesis in total knee replacement. J Bone Joint Surg. 1977;
59:77-79.
3. Coventry M: Two part total knee arthroplasty. Evolution and present status. Clin Orthop 1979; 145:29.
4. Bargren JH, Blaha JD, Freeman MZ. Alignment in total knee arthroplasty. Clin Orthop 1983; 173:178.
5. Windsor RE, Scuderi GR, Moran MC, et al. Mechanisms of failure of the femoral and tibial components in total
knee arthroplasty. Clin Orthop 1989;248:15.
6. Hamilton LR: UCI total knee replacement: a follow-up study. J Bone Joint Surg Am 1982 64:740.
7. Tew M, Waugh W.Tibiofemoral alignment and the results of knee replacement. J Bone Joint Surg Br 1985;
67:551.
8. Agglietti P, Buzzi R. Posteriorly stabilised total condylar knee replacement: three to eight years follow-up on 85
knees. J Bone Joint Surg Br1988; 70: 211.
9. Hsu HP, Garg A, Walker PS, et al. Effect of knee component alignment on tibial load distribution with clinical
correlation. Clin Orthop 1989;248: 135.
10. Insall JI, Scott WN, Ranawat CS. The total condylar knee prosthesis: a report on two hundred and twenty cases.
J Bone Joint Surg Am 1979; 61:173.
11. Ritter MA, Faris PM, Keating M, et al. Postoperative alignment of total knee replacement. Clin Orthop. 1994;
299: 153.
12. Gibbs AN, Green GA, Taylor JG. A comparison of the Freeman-Swanson (ICLH) and Walldius prostheses in
total knee replacement. J Bone Joint Surg Br 1979; 61:358
13. Bartel DL, Burstein AH, Santavicca EA, et al. Performance of the tibial component in total knee replacement.
J Bone Joint Surg Am 1982; 64: 1026.
14. Green GV, Berend KR, Berend ME, Glisson RR, Vail TP: The effects of varus tibial alignment on proximal
tibial surface strain in total knee arthroplasty. J Arthroplasty 2002; 17: 1033-1039.
15. Mochizuki RM, Schurman MD: patellar complications following total knee arthroplasty. J Bone Joint Surg.
1979; 61:879-883.
16. Berger RA, Crossett LS, Jacobs JJ, Rubash HE: Malrotation causing patellofemoral complications after total
knee arthroplasty. Clin Orthop. 1998; 356:144-153.
17. Yoshii I, Whiteside LA, White SE, Milliano MT: Inuence of prosthetic joint line position on knee kinematics
and patellar position. J Arthroplasty 1991; 6: 169-177.
18. Akagi M, Matsuse Y, Mata T, Asada Y, Horiguchi M, Iida H, Nakamura T: Effect of rotational alignment on
patellar tracking in total knee arthroplasty. Clin Orhop. 1999; 366:155-163.
19. Matsuda S, Miura H, Nagamine R, Urabe K, Hirate G, Iwamoto Y: Effect of femoral and tibial component posi-
tion on patellar tracking following total knee arthroplasty: 10-year follow-up of Miller-Galante I knees. Am J
Knee Surg. 2001; 14:152-156.
20. Rhoads DD, Noble PC, Reuben JD, Mahoney OM, Tullos HS: The effect of femoral component position on
patellar tracking after total knee arthroplasty. Clin Orthop. 1990; 260:43-51.
21. Anouchi YS, Whiteside LA, Kaiser AC and Milliano MT: The effects of axial rotational alignment of the femo-
ral component on knee stability and patellar tracking in total knee arthroplasty demonstrated on autopsy speci-
mens. Clin Orthop 1991; 287: 170.
22. Fehring TK: Rotational malalignment of the femoral component in total knee arthroplasty; Clin Orthop 2000;
380:72-79.
23. Berger RA, Rubash HE, Seel MJ, Warren HT, Crosset LS: Determining the rotational alignment of the femoral
component in total knee arthroplasty using the epicondylar axis. Clin Orthop. 1993; 286:40-47.
24. Olcott CW and Scott RD: Femoral component rotation during total knee arthroplasty. Clin Orthop 1999; 367:
39-42
25. Miller MC, Berger RA, Petrella AJ, Karmas A, rubash HE: Optimizing femoral component rotation in total knee
arthroplasty. Clin Orthop 2001; 392: 38-45.
26. Hanada H, Whiteside LA, Steiger J, Dyer P, Masatoshi N: Bone landmarks are more reliable than tensioned gaps
in TKA component alignment. Clin Orthop. 2007; 462:137-142.
27. Hungerford DS, Kenna RV: Preliminary experience with a total knee prosthesis with porous coating used with-
out cement. Clin Orthop 1983; 176:95.
28. Moreland JR, Bassett LW and Hanker GJ. Radiographic analysis of the axial alignment of the lower extremity.
J Bone Joint Surg Am 1987; 69:745-49.
32
29. Laskin RS: Flexion space conguration in total knee arthroplasty. J Arthroplasty 1995;10:657-660
30. Poilvache PL, Insall JN, Scuderi GR, Font-Rodriguez DE: Rotational landmarks and sizing of the distal femur
in total knee arthroplasty. Clin Orthop. 1996; 331: 35-46.
31. Yoshioka Y, Siu D and Cooke TD: The anatomy and functional axes fo the femur. J Bone Joint Surg. 1987;
69:873-80.
32. Grifn FM, Insall JN, Scuderi GR: The posterior condylar angle in osteoarthritic knees. J Arthroplasty 1998;
13:812-5.
33. Asano T, Akagi M, Nakamura T: The functional exion-extension axis of the knee corresponds to the surgical
epicondylar axis. J Arthroplasty 2005; 20: 1060-7.
34. Arima J, Whiteside LA, McCarthy DS, White S: Femoral rotational alignment, based on the anteroposterior
axis, in total knee arthroplasty in a valgus knee. J Bone Joint Surg. 1995;77:1331-1334
35. Whiteside LA, Arima J: The anteroposterior axis for femoral rotational alignment in valgus total knee arthro-
plasty. Clin Orthop. 1995; 321:168-172
36. Dorr LD, Boiardo : Technical considerations in total knee arthroplasty. Clin Orthop 1986; 205:5-11.
37. Insall JN: Technique of total knee replacement. AAOS Instr Course Lect 1981; 30:324
38. Freeman MA, Todd RC, Bamert P, Day WH: ICLH arthroplasty of the knee: 1968-1977. J Bone Joint Surg.
1978; 60B: 339-344
39. Grifn FM, Math K, Scuderi GR, Insall JN, Poilvache PL: Anatomy of the epicondyles of the distal femur.
J Arthroplasty 2000; 15:354-359.
40. Vandamme G, Defoort K, Ducoulombier Y, Van Glabbeek F, Bellemans J, Victor J. What should the surgeon
aim for when performing computer-assisted knee arthroplasty? J Bone Joint Surg Am 2005; 87A: 52-58
41. Scuderi GR, Insall JN: Posterior stabilized prosthesis. Orthop Clin North Am 1989; 20:71-78
42. Mantas JP, Bloebaum RD, Skedros JG, Hofmann AA. Implications of reference axes used for rotational align-
ment of the femoral component in primary and revision knee arthroplasty. J Arthroplasty 1992; 7:531-535
43. Grifn MF, Insall JN, Scuderi GR: Accuracy of soft tissue balancing in total knee arthroplasty. J Arthroplasty
2000;15:970-973
44. Matsuda S, Matsuda H, Miyagi T, Sasaki K, Iwamoto Y, Miura H: Femoral condyle geometry in the normal and
the varus knee. Clin Orthop. 1998; 349:183-188
45. Nagamine R, Miura H, Inoue Y, Urabe K, Matsuda S, Okamoto Y, Nishizawa M, Iwamoto Y: Reliability of the
anteroposterior axis and the posterior condylar axis for determining rotational alignment of the femoral compo-
nent in total knee arthroplasty J Orthop Sci. 1998; 3:194-198
46. Yoshino N, Takai S, Ohtsuki Y, Hirasawa Y: Computed tomography measurement of the surgical and clinical
transepicondylar axis of the distal femur in osteoarthritic knees. J Arthroplasty 2001; 16:493-497
47. Eckhoff DE, Hogan C, DiMatteo L, et al: Difference between the epicondylar and cylindrical axis of the knee.
Clin Orthop 2007; 461:238-244
33
CHAPTER 2: OBJECTIVES AND METHODOLOGY
I. Aims of the thesis
As outlined in the introduction, current clinical understanding of normal knee kinematics is
subject of an ongoing discussion. Moreover, it remains unclear to what extent the natural ki-
nematic patterns can be repaired after post-traumatic or degenerative disease of the knee joint
and what the specic role of a prosthetic implant should be. Variable degrees of constraint
and mechanical substitution for lost ligament function are possible. Equally unsolved is the
issue of three-dimensional alignment and geometrical anatomy of the distal femur and proxi-
mal tibia, where a common reference frame is needed for better understanding and improving
surgical outcomes. This thesis aims at revealing the relation between the surface anatomy of
the bones and the kinematic tibiofemoral patterns, leading to a rank order of reliability and
consistency of reference axes and points. Following this Cartesian exercise, our practical
ability to reproduce normal alignment during surgery will be critically evaluated with intra-
operative recording of the accuracy of locating surface derived landmarks and post-operative
evaluation of the nal position of the implants. In a next step, we want to link this quantitative
anatomical description to its functional consequences for the ligaments in dening ligament
(an)isometry for the medial collateral ligament, the lateral collateral ligament and the medial
patellofemoral ligament.
The variety of kinematic patterns that has been described in the literature will be compared
to our results, obtained in a knee kinematic simulator, and the specic effect of loading and
muscle action on knee kinematics will be examined. The changes imposed on the joint and its
soft tissues by the insertion of a prosthetic device will be described. As the comparison can
be made on a one to one basis in the same specimen, actual changes will purely be attributed
to the surgical act and implant conguration without further bias. The impact of prosthetic
design and the difference between cruciate ligament preservation and cruciate ligament sub-
stitution will be examined.
Studying patients in vivo after the insertion of the same type of prosthesis allows for compar-
ing the kinematic patterns that were observed in the ex vivo experiment and validating the
experimental set-up. An overview of the above-described objectives, linked to their clinical
relevance is shown in table I.
34
Chapter 2: Objectives and methodology
AIM RELEVANCE
Describe relation between surface ge-
ometry and kinematics
Ranking of reliability and consistency
of reference axes for surgical align-
ment
Determine 3 plane accuracy of surgical
navigation
Discrimination of plane dependent ac-
curacy of surgical navigation
Describe ligament isometry, based
upon anatomical references
Validation of previous quantitative an-
atomical descriptions
Describe normal knee kinematics Validate or refute previous models and
investigate the effect of loading and
muscle action
Describe prosthetic knee kinematics Understand the differences induced by
the prosthetic implant
Compare cruciate preservation versus
mechanical substitution
Evaluation of the quality of the me-
chanical substitution of the cruciate
ligaments
Validate experimental model Model to be used for future investiga-
tions on different types of implants
Table I: Summary of the aims of the thesis with consequent relevance
II. Hypotheses
The ex vivo measured kinematic patterns of the cadaver specimens will correlate with previ-
ously published data.
1-8
The insertion sites of supercial medial collateral ligament, lateral collateral ligament and
medial patellofemoral ligament as anatomically described by La Prade
9,10
will display a
change in overall length of less then 10% during the loaded squat.
Insertion of a knee prosthesis will alter the kinematics, mechanics and ligament insertion site
distances.
The changes induced by insertion of a prosthesis will be greater when the cruciate ligaments
are sacriced.
The ex vivo measured kinematic patterns of the cadaver specimens after insertion of pros-
thesis will correlate with in vivo measurements on patients having undergone knee replace-
ment.
35
III. Materials and Methods of the ex vivo experiment
The work to be described consists of three major parts.
In part I, the kinematics and mechanics of the native human knee joint are described, based
upon an ex vivo experiment, simulating a squat in loaded conditions on a cadaver specimen.
The workow includes geometrical analysis of bone morphology, solid body kinematics of
femur and tibia, description of quadriceps loads, and changes in length of the ligaments and
the patellar tendon.
In part II, a similar experiment with identical biomechanical analyses is performed on the
same specimens, after insertion of a total knee prosthesis. The geometry and biomechanics
of the native knee are compared to the geometry and biomechanics of the replaced knee. In
addition, a comparison is made between a prosthesis that preserves both cruciate ligaments
and a prosthesis that aims at substituting the sacriced cruciate ligaments.
In part III, outcomes are studied in vivo. In a uoroscopic study, the kinematic patterns of
the knee during a loaded squat, after implantation of bicruciate replacement prosthesis, are
examined. This working sequence meets the objectives that were put forward in the design of
the experiment and covers clinically signicant topics, as summarized in table I.
A. METHODOLOGICAL SEQUENCE
Thirteen fresh frozen cadaver specimens were used for the ex vivo testing, one for the pilot
experiment, and twelve for the subsequent measurements. The ex vivo experimental set-up
was guided by the following sequence. A general description of all steps is given. Detailed
descriptions, relevant for specic investigations are given in the respective sections of Chap-
ter 3.
1. Mounting of the reference markers on tibia and femur
The specimen consists of a complete limb, disarticulated at the level of the hip, which has
been fresh frozen. Commercially available optical reference frames (Brainlab, Feldkirchen,
Germany) were rigidly attached to the tibia, the femur and the patella (g 1a). The reference
frames carry spheres for reecting infrared light. These markers serve a triple purpose: they
can accurately be located on the pre- and post-implantation CT scan, they can be tracked by
the three dimensional motion analysis system (Vicon Motion Systems, Los Angeles, USA),
and they guide the surgical navigation system during insertion of the prosthesis for part II of
the experiment.
2. CT scanning of frozen specimen with markers
The specimens were transported in frozen state in an isolating box to the CT scan.
For all cadaveric knee specimens, we obtained volumetric CT scans on a 64-row helical
multidetector computed tomography (MDCT) scanner (General Electric Lightspeed VCT,
36
Milwaukee, WI). The images were obtained at 120 kV and 450 mA, with a slice thickness of
1.25 mm and a pitch of 0.5 mm/rev g 1b. Raw data were processed using a bone lter. The
images were stored in DICOM format onto a DVD. The CT scans were analysed using Mim-
ics 11.02 and its MedCAD module (Materialise, Haasrode, Belgium) to create the surface
reconstruction and identify the bony landmarks. (See below: section B. Data processing)
a b
Fig 1 a. Amputated fresh frozen specimen with reference frames attached
b. CT scanning of specimen
3. Thawing of the specimen
The day before the experiment, the specimen was taken out of the refrigerator to allow 24
hours for thawing.
4. Initialisation of the surgical navigation software
At this stage, the surgical navigation software (Brainlab, Feldkirchen, Germany) to be used
for later implantation of the prosthesis was initialized. As the next step included amputation
of the foot and resection of the proximal femur, important alignment landmarks for surgical
references would be lost. The navigation software determined the centre of the hip as the
femur was moved through a number of circular motions, keeping the femoral head stabilized
(g 2). In previous work, we validated this algorithm,
and demonstrated a mean accuracy of 1.6 mm, com-
paring biplanar X-Rays to the computed centre of the
hip.
11
At the level of the tibia, the medial and lateral
malleolus were marked and stored in the navigation
software for later reference.
Fig 2 Screenshot of surgical navigation interface, demonstrating
kinematic determination of the centre of the hip.
37
5. Amputation of the proximal femur and the ankle
The proximal femur was amputated 32 cm
proximal to the joint line. The foot was
amputated 28 cm distal to the joint line.
The bones were cleared of soft tissues and
muscles over a distance of 12 cm (g 3).
Fig 3 Measured amputation of hip and foot and
rigid xation of specimen in containers with
PMMA
6. Fixation of Femur and Tibia in containers
The tibia and the femur were xed in the containers with polymethylmethacrylate (PMMA).
The tibia was xed in a position parallel to the container in the coronal and the sagittal plane.
The femur was xed in the coronal plane, taking the physiologic valgus angle into account. In
the sagittal plane, the femur was xed parallel with the container (g 4). Screws were driven
into the end of the bones for providing sufcient rotational control in the PMMA.
Fig 4 alignment and xation of tibia and femur in containers
7. Preparation of tendon xation
On the medial side, the semitendinosus and semimembranosus tendon were sutured, using a
Krackow stitch with Ticron nr5 suture. On the lateral side, the biceps tendon was prepared
in a similar way. For the quadriceps, a stronger xation was needed as the pilot experiment
revealed high forces: the tendon was cut at 10 cm proximal of its attachment to the patella and
38
wrapped around a metal rod with sectional diameter of 5 mm. The tendon was sutured on its
own, using Ticron nr5 suture, and Mersilene tape reinforcement (g 5).
Fig 5 Detail of metal rod, securely xed with quadriceps tendon
wrapped and sutured
8. Calibration of the 3D motion analysis cameras (Vicon Motion Systems, Los
Angeles, USA)
Five cameras were positioned on the medial side of the specimen to allow maximal visual-
ization of the optical reference markers xed to the tibia, femur and patella. Using a special
want with calibrated distance between the reective spheres, the software was instructed to
calculate the relative position of the cameras in space.
9. Mounting of specimen in Oxford Rig
The specimen was then mounted onto the mechanical knee rig with the containers rigidly
xed to the hip actuator and the ankle load cell (g 6). A detailed description of the rig is
given in section D.
Fig 6 Specimen mounted in kinematic rig with the
containers securely xed to the machine.
The reference markers on the femur, tibia
and patella can be clearly seen, reecting in
the light of the photography ash light. The
femur is above, the tibia below.
39
10. Performance of the ex vivo squat and recording of motion and loads
The position of the rigid bodies consisting of the femur, tibia and patella, with their respec-
tive reference frames was followed as a function of time by the 3D motion analysis system.
The loads on the ankle and on the quadriceps tendon were measured with calibrated load
cells. Initial recording consisted of a positioning of the knee in full extension with the tibial
container xture loose along the vertical axis. The knee was then pulled into full extension
by exerting a progressive load on the quadriceps. Static recording of the position of the rigid
bodies was performed. A similar static measurement was performed near 10 and 20 of
exion. The knee was then initialized at a position of 20 to 30 of exion for starting the
dynamic measurements. A squat was performed and motion and loads were recorded dur-
ing the full cycle. The full set of measurements was performed with a constant load on the
hamstrings of 50N medially and laterally. In 6 specimens, additional recordings were made
in passive conditions, in loaded conditions with only the quadriceps acting, with quadriceps
and medial hamstrings acting, and with quadriceps and lateral hamstrings acting.
11. Demounting of specimen and insertion of the knee prosthesis
A standard parapatellar medial incision was used to expose the knee joint. The patella was
everted. The integrity of the cruciate ligaments and the articular cartilage was checked and
recorded. A digital photograph was taken of all specimens (g 7a). Further steps for using
the surgical navigation system were undertaken. Virtual surgical planning was executed and
bony cuts made accordingly. Insertion of the knee prosthesis (Smith and Nephew, Memphis,
USA), bicruciate stabilized with sacrice of the ligaments (BCS) for the rst six specimens
and bicruciate retaining for the following six specimens (BCR)(g 7b). The components
were xed with PMMA, selection of insert thickness was performed as in surgical practice.
The patella was not resurfaced. The arthrotomy was closed with a vicryl 2 suture.
40
Fig 7a Photographs of all specimens at the time of arthrotomy
41
Fig 7b Pictures of the two different implants. The Bicruciate retaining implant (BCR) is shown in exion (upper
left) and extension (upper right), demonstrating the clearance for the cruciate ligaments in exion and ex-
tension. The Bicruciate substituting implant (BCS) is shown in exion (lower left) and in a sagittal plane
cross section for demonstrating the cam-post relationship (lower right)
12. Mounting of the specimen with prosthesis on the Oxford rig
The specimen was mounted on the Oxford rig in identical fashion as in the native knee.
13. Repetition of ex vivo squat and recording of motion and loads
All actions of step 10 were repeated.
14. Demounting of specimen and dissection of supercial MCL
The specimen was taken out of the Oxford rig and the supercial medial collateral ligament
(sMCL) was carefully dissected. The insertion sites of the sMCL, proximally and distally on
the tibia were identied, according to the quantitative anatomical description by LaPrade.
9

42
Two 3mm aluminum nails were
driven in the bone at the insertion
sites for later identication on the CT
scan (g 8).
Fig 8 Dissection of the MCL and marking of
the proximal and distal insertion of the
supercial medial collateral ligament
15. Freezing and CAT scanning of specimen
The specimen was then frozen in extension, after the patella was cut loose and frozen sepa-
rately (to avoid scattering from the prosthesis during CT scan). The specimen was taken in
the isolating box in frozen status to the CT scan. The scan was performed, following the
identical protocol as previously described with addition of the application of scatter reduction
software for the tibia and the femur.
B. DESCRIPTION OF DATA PROCESSING
The datasets of all specimens were handled in identical sequence. The DICOM dataset of the
native knee, yielded by the CT scan and taken before the ex vivo experiment, was fed in 3D
analysis software (Mimics, Materialise, Haasrode, Belgium). All anatomical landmarks,
ligament insertion sites, planes and axes that are of interest were located on the CAT scan
images. (g 9) A detailed description of all datapoints, planes and axes is given in chapter
3, section I and II. The centres of the reective spheres were determined for the reference
frames of femur,
Fig 9 3D reconstruction of femur and
tibia and marking of surface
reference points in Mimics. The
small spheres to the right of the
anatomic specimen represent
markers, xed on the reference
frames. The big purple sphere on
top represents the best-t circle
to the femoral head. Its centre
will serve to dene the femoral
mechanical axis.
43
Fig 10 Geometrical denitions of points, axes and planes on tibia and femur
44
tibia and patella, dening a Cartesian coordinate system, based on the optical trackers. A Car-
tesian coordinate system was dened for each separate bone, based upon surgically relevant
axes (g 10). These references were identical to those used by Grood and Suntay
12
, except for
the horizontal axis of the femur and the tibia. We rened the denition of the horizontal axis,
as Grood and Suntay were conned to using landmarks identiable on bi-planar X-rays. As
the CT scan was available to us, more secure and reproducible denitions could be used in
the horizontal plane, allowing a more reliable denition of the femoral and tibial transverse
axis (Chapter 3, I).
The coordinate transformation
yielded a set of coordinates al-
lowing reconstruction of the
joint line, analysis of quantita-
tive anatomical relationships and
comparison between the surface
geometry of the native versus
the replaced knee. Also, the cal-
culation of the coordinates of an
identical pre- and post-operative
reference point allowed to con-
rm the reference frames had not
moved during the experiment or
transportation (g 11 and section
D, i).
Fig 11 Example of excel sheet with landmark coordinates
For the description of the kinematic behavior of the bones, relative to each other, the method-
ology of Grood and Suntay
12
, with the creation of a joint coordinate system was followed for
all rotational descriptions. When translations in a certain plane were of interest, orthogonal
projection on that plane was used. The dynamic measurements were presented in the form of
a datasheet with all rotations and translations, isometry descriptions and loads presented as a
function of the exion angle. This presentation allowed a clinical interpretation and statistical
analysis. Below is a summary of the order of the mathematical processing of the data:
1. Feed CT based data from DVD
into Mimics
2. Create surface model of distal fe-
mur, proximal tibia and patella
3. Locate all needed anatomical
landmarks in Mimics software and
export this list of coordinates, still
in the CAT based Cartesian coor-
dinate system
Fig 12 Coordinate transformation
45
04. Dene Cartesian coordinate systems, based on the three optical tracking frames of femur,
tibia and patella
05. Transform the coordinates of all anatomical landmarks from the CAT model into these
new Cartesian coordinate systems (g 12).
06. Calculate the relative position of each anatomical landmark relative to the reference frame
(femur, tibia and patella) within the global Vicon coordinate system
07. Describe the position of each anatomical landmark as function of time after ltering the
VICON data with Woltering lter for mechanical (vibration and resonance) or optical,
electronic noise etc.
08. Dene a single bone coordinate system for femur, tibia and patella.
09. For rotational kinematic description: Grood and Suntay protocol for joint coordinate sys-
tem using the TMAx and FTAx as body xed axes and the cross product as the oating
axis for ab- and adduction.
12
10. For translational kinematic descriptions use projection planes as dened in Chapter 3,
Section I. Medial femoral condyle translation (MFT) is dened as the distance of the
orthogonal projection of the medial condyle centre onto the horizontal plane of the tibia
to the tibial transverse axis. The lateral femoral condyle translation (LFT) is dened in
analogy (g 13).
Fig 13 Projection of medial condyle centre onto horizontal plane of the tibia for measuring medial and lateral
condylar translation (MFT and LFT).
C. KNEE KINEMATIC SIMULATOR
A dynamic knee simulator system based on the Oxford rig (g 14), was customized for this
study. This electromechanical system is designed to simulate and record the motions and
loads in a knee joint during squatting. The hip joint can move up and down and can ex
and extend. The ankle joint can move medio-laterally and has all three rotational degrees of
46
freedom (exion-extension, internal-external rotation, ab-adduction). Thus, this construction
provides the knee joint with all six degrees of freedom. Only the exion angle of the knee is
controlled directly (by programming the hip position as a function of time). All other degrees
of freedom are left free. Translations and rotations along these axes are governed by the ge-
ometry and anatomy of the joint. One actuator simulates the quadriceps muscle, and a second
actuator produces vertical hip motion. The quadriceps actuator is positioned on the upper leg
in a way that reproduces its anatomical location and thus also its moment arm with respect to
the knee joint. Two constant force springs (50 N each) have been added to the system to load
the hamstrings on the lateral and medial side of the tibia. They are xed to the metal frame
which represents the pelvis to reproduce their bi-articular function. Their position is such that
their moment arms are similar to the in vivo situation. Sensors placed in line with the actua-
tors detect the quadriceps force, the ankle forces and moments, and hip height relative to the
ankle. A real-time data acquisition and closed feedback system (Labview, National Instru-
ments, TX, USA) is used to perform a squat with a certain hip velocity while simultaneously
applying a quadriceps force on the knee to induce a vertical ankle force. The hip actuator
is controlled by error feedback from the hip position sensor using a proportional-integral-
derivative (PID) controller. If the hip position is too high with respect to the programmed
position at any instant during the exion cycle (i.e. the hip is lagging behind), the actuator
will be instructed to speed up and vice versa. Likewise, the quadriceps actuator is controlled
by error feedback from the 6-axis ankle load cell using a similar PID system. If the vertical
ankle force is too low with respect to the programmed value at any instant during the squat,
the actuator will be instructed to pull harder and vice versa. Both control loops are completely
independent. A block diagram of the control algorithm is shown in g. 15. To record the
movements of the knee joint, which are not controlled by the simulator, a gait analysis sys-
tem is used (Vicon, Oxford, UK). Four infrared emitting cameras are positioned around the
Oxford rig and pick up the reected infrared light from the markers on the frames, which are
xed to the bones. Whenever a marker is visible for at least two cameras, its position can be
determined by triangulation. The accuracy of the system for this set-up was tested and shown
to be better than 1 mm.
Fig 14 Dynamic knee simulator based
on the Oxford rig, with speci-
men installed. The hip joint
(above) has two degrees of
freedom. The ankle joint can
slide medio-laterally along the
metal bar and has all three ro-
tational degrees of freedom.
47
The system is set to measure with a frequency of 100 Hz. Since three markers are xed to
each bone, the position and orientation of the bones in space can be determined as a function
of time.

Fig 15 Process diagram of knee kinematics simulator
48
D. ACCURACY AND SENSITIVITY ANALYSIS
i. Reference frame stability
Introduction
During the tests, only the reective markers xed to the femur and the tibia were visible for
the infrared cameras. The position of the bones was derived afterwards from the measured
position of these reective markers. As the experiment involved extensive handling and po-
sitioning of the specimens, frame stability was checked to conrm the relative position of the
markers to the bones had remained unchanged in the course of the experiment.
Methods
A second CT scan of the frozen specimens was performed after all experiments were n-
ished. The relative position of selected reference points was compared to the initial position
as measured on the rst CT scan. As the second scan was performed on the amputated speci-
men with the prosthesis in place,
only few points could serve as a reli-
able landmark. The centre of the en-
try point in the cortex of one of the
bone pins was selected as a primary
reference point in both femur and
tibia. The lateral epicondyle and the
tip of the bula were selected as sec-
ondary reference points in the femur
and tibia-bula unit, respectively
(g 16).
For each of the four control points,
the distance between the point and
the three reective markers attached
to the same bone was recorded in
the rst and the second CT scan. A
change in the distance between a
control point and the markers would
indicate that the frame with the re-
ective markers had moved with re-
spect to the bone.
Fig 16 Primary and secondary reference points on femur and tibia
49
Results
Table II shows an overview of the recorded changes in distance for all specimens and all
control points. In most of the cases the recorded changes in distance are smaller than 1 mm
and thus of the same order of magnitude as the resolution of the CT image. Only the lateral
epicondyle shows an average position change of more than 1 mm. This can be explained by
the intra-observer variability of this point (reported in chapter 3, I)
Discussion
Position change of the markers relative to the reference frames is within the intra-observer
error range. We concluded the reference frames did not move relative to the bone during the
experimental tests.
50
Femur " distance from entry point to " distance from lateral epicondyle to
Marker 1 Marker 2 Marker 3 Av Marker 1 Marker 2 Marker 3 Av
Spec 1 3.0 2.0 1.7 2.2 2.0 1.3 1.4 1.6
Spec 2 0.9 0.4 0.4 0.6 1.7 2.5 2.8 2.3
Spec 3 0.4 0.9 0.5 0.6 1.0 1.8 1.4 1.4
Spec 4 0.0 0.2 0.2 0.1 4.0 1.7 2.2 2.6
Spec 5 0.1 0.4 0.3 0.3 1.6 0.1 2.0 1.2
Spec 6 0.9 0.6 0.2 0.6 0.4 0.2 0.2 0.3
Spec 7 0.3 0.1 0.5 0.3 1.8 0.3 1.4 1.2
Spec 8 0.6 0.1 0.4 0.4 2.8 0.0 2.7 1.9
Spec 9 1.2 1.1 0.4 0.9 1.7 2.0 2.3 2.0
Spec 10 0.3 0.1 0.3 0.3 0.1 0.3 0.4 0.3
Spec 11 1.3 1.4 0.9 1.2 2.7 3.2 2.0 2.6
Spec 12 0.4 0.8 0.1 0.5 0.0 0.1 0.5 0.2
Average 0.7 1.5
Intraobserver 1.3
Tibia " distance from entry point to " distance from tip bula to
Marker 1 Marker 2 Marker 3 Av Marker 1 Marker 2 Marker 3 Av
Spec 1 1.4 0.5 0.3 0.7 2.8 2.3 2.1 2.4
Spec 2 0.1 0.1 0.1 0.1 0.4 0.2 0.3 0.3
Spec 3 0.4 1.4 1.3 1.0 0.4 0.7 0.5 0.5
Spec 4 0.1 0.6 0.0 0.3 0.0 0.3 0.4 0.3
Spec 5 1.5 0.3 1.6 1.1 1.4 0.8 1.1 1.1
Spec 6 0.0 0.1 0.0 0.1 0.2 0.9 0.8 0.6
Spec 7 0.1 0.1 0.2 0.1 0.2 0.9 0.0 0.4
Spec 8 0.5 0.4 0.0 0.3 1.7 1.3 1.7 1.5
Spec 9 0.5 1.9 2.6 1.7 0.1 1.4 0.6 0.7
Spec 10 0.6 0.8 0.6 0.7 0.2 0.0 0.4 0.2
Spec 11 0.3 0.7 0.8 0.6 1.0 0.6 1.0 0.9
Spec 12 0.6 0.9 0.7 0.7 1.4 0.6 0.8 0.9
Average 0.6 0.8
Intraobserver 0.7
Table II Distances of markers to reference points in mm
51
ii. Repeatability
Introduction
A basic analysis of the repeatability of the knee kinematics simulator was done on data ob-
tained for cadaver specimens. The purpose of the analysis was to determine what effect there
was of repeated squat trials on specimen kinematics when all test conditions are the same.
Methods
Only native knee repeatability results are reported here, because initial testing showed that
native knees are more variable compared to replaced knees. Only the test conditions with
three or more squat trials were included for repeatability analysis. The repeatability of the
parameters Tibial Rotation TR, Medial Femoral Condyle Translation MFT, Lateral Femo-
ral Condyle Translation LFT, Tibiofemoral Coronal Alignment ALCor, Ankle Load AL, and
Quadriceps Load QL were measured as functions of Flexion Angle FA.
Repeatability of the system was quantied based on previously described methods. For each
set of repeated trials under the same test conditions, the mean kinematics curves were calcu-
lated. Then the residual errors between each trial curve and the mean curve were calculated,
and the standard deviations of the errors were recorded as measures of variability. The worst-
case, or largest, variability values were used to calculate coefcients of repeatability (CR) for
system measurements. Assuming normality of the residual errors, the CR was dened as two
times the standard deviation (CR = 2SD) of the residuals
13
. As such, the CR values give ap-
proximate 95% condence intervals for the kinematics parameters around their mean curve
in each test condition.
Results
Four test conditions were analysed, whose results are summarized in Table III. Each of the
test conditions had a different specimen, target ankle load, and/or exion range. Based on the
results of the most variable conditions, 95% of system measurements from repeated squat tri-
als in descent after 30 exion, is expected to fall within the following ranges about a mean
curve:
TR: 0.5
MFT: 0.8 mm
LFT : 0.7 mm
AlCor : 0.6
AL: 13 N
QL: 49 N
52
Discussion
These ranges of variability are small. The results suggest that one squat trial at any particular
condition gives a similar result as averaging 3-4 trials, for the kinematics measurements in
this study, within the ranges listed above.
Table III - Mean and variability of residuals from repeated trials*
MEASUREMENT
TEST CONDITION
Worst-
Case
CR**
Specimen
spec 1
pre-TKA
spec2
pre-TKA
spec2
pre-TKA
spec3
pre-TKA
Target AL 90 N 90 N 130 N 90 N
FA () 40-120 30-120 30-120 50-130
n = 3 3 4 3
TR ()
Mean 0.00 0.00 0.00 0.00
1SD 0.24 0.09 0.10 0.13
2SD 0.48 0.19 0.19 0.27 0.48
MFT (mm)
Mean -0.15 0.00 0.00 0.00
1SD 0.39 0.08 0.09 0.20
2SD 0.79 0.16 0.18 0.39 0.79
LFT (mm)
Mean -0.12 0.00 0.00 0.00
1SD 0.34 0.12 0.12 0.10
2SD 0.68 0.24 0.23 0.20 0.68
ALCor ()
Mean -0.13 0.00 0.00 0.00
1SD 0.30 0.03 0.03 0.06
2SD 0.59 0.05 0.06 0.11 0.59
AL (N)
Mean 0.00 0.00 0.11 0.00
1SD 6.49 2.93 2.63 2.92
2SD 12.99 5.87 5.26 5.84 12.99
QL (N)
Mean N/A N/A N/A 0.00
1SD 24.52
2SD 49.04 49.04
* For selected measurements with relatively large variability, and conditions where n ! 3
** Coefcient of repeatability (CR) = 2x standard deviation, as dened by Bland et al.
13
Table III Summary of system repeatability based on cadavers in different test conditions. The variability of the
residual errors of each trial were calculated. Coefcient of repeatability (CR) values give an approxi-
mate 95% condence interval for each measurement across the squat, assuming normal distribution
of the errors. Abbreviations: Tibial Rotation TR, Medial Femoral Condyle Translation MFT, Lateral
Femoral Condyle Translation LFT, Tibiofemoral Coronal Alignment ALCor, Ankle Load AL, and
Quadriceps Load QL.
53
iii. Effect of Ankle Load
Introduction
This analysis was performed on the knee kinematics simulator to nd the effect of target
ankle load on cadaver knee kinematics during a squat, relative to inter-specimen variability.
This information helped determine which ankle loads to study in the full experimental analy-
sis.
Methods
Measurements versus exion angle for specimen 2 and 3 were analysed pre- and post-TKA
at 90, 130, and 180 N target ankle load (AL) in descent. These loads were chosen as the ex-
tremes of possible test conditions. The variation among kinematics curves at different loads
was recorded.
Results
The measurements analysed included TR, MFT, and LFT, and their results are shown here
because they had the most variation. The other kinematics parameters showed less variation
and are not shown here. Replaced knees also showed less variation than native knees and are
not shown here.
Figure 17 below shows graphs of the results for native knees. The shapes of the curves at dif-
ferent loads were approximately the same within the same specimen, although the absolute
values changed by small increments. TR and the two translations showed maximum differ-
ences between the 90 N and 180 N curves of less than 2 and less than 4 mm, respectively.
Inter-specimen differences exceeded the inter-load differences.
Discussion
The effects of AL were small for TR, MFT, LFT, and the other parameters not shown here,
including for replaced knees. The general patterns of the curves also were similar as AL
changed. Therefore, any single AL between 90 to 180 N should be able to demonstrate pos-
sible differences between different knees. Plots of the data showed that inter-specimen vari-
ability was greater than inter-load variability in the conditions tested.
The knee simulator itself also was noted to better reach the target AL at loads greater than
90 N. Also the quadriceps load (QL) was less variable at ankle loads of less than 180 N. As
a compromise for system stability, one physiologic target AL of 130 N could be chosen for
study. This reduced the number of squat trials and also reduced the chance of damage to the
specimens.
54
Spec 2 Spec 3
Fig 17 Knee kinematics parameters vs. exion (FA) at different target ankle loads (AL). TR, MFT, and LFT versus
exion angle during squat descent at different target ankle loads (AL) for spec 2 and spec 3.

T
R

-6
-4
-2
0
2
4
30 50 70 90 110 130

FA ()
T
R

(
)

-
6
-
4
-
2
0
2
4
30 50 70 90 110 130

T
R

(
)
90 N 130 N 180 N
M
F
T

-14
-12
-10
-8
-6
-4
-2
0
2
4
30 50 70 90 11
0
130
FA ()
M
F
T

(
m
m
)
-14
-12
-10
-8
-6
-4
-2
0
2
4
30 50 70 90 110 130
FA ()
M
F
T

(
m
m
)
L
F
T

-16
-14
-12
-10
-8
-6
-4
-2
0
30 50 70 90 110 130
FA ()
L
F
T

(
m
m
)

-16
-14
-12
-10
-8
-6
-4
-2
0
30 50 70 90 110 130
FA ()
L
F
T

(
m
m
)

FA ()
55
iv. Single vs. Double Squats
Introduction
This analysis compared single and double squat measurements from the knee kinematics
simulator. This was done to determine how data from the simulator may differ in the second
consecutive squat of a double squat motion.
Methods
Double squat trials were attempted on a case-by-case basis at the researchers discretion,
based on cadaver and machine integrity. All double squat trials that completed at least half
of the 2
nd
squat were chosen for analysis. 13 double squats from 5 specimens were usable for
this study. They had a variety of test conditions: pre- or post-total knee arthroplasty, different
hamstring attachments, and different target ankle loads. These trials were pooled to increase
sample sizes. Trial data were split into the 1
st
and 2
nd
squats, and absolute and relative differ-
ences between the 1
st
and 2
nd
squats were measured.
Results
Data analysis was limited to descent due to system limitations. Limited ascent data were
available for the 2
nd
squat due to instability of the system (g 18). Each specimen had unique
motions in time, and to reduce the variability, the data were normalized according to exion
angle (FA). Adduction (AlCor), tibial rotation (TR), lateral femoral translation (LFT), me-
dial femoral translation (MFT), quadriceps load (QL), and ankle load (AL) were analysed
versus FA. Mean absolute differences between the paired squats at each FA are summarized
in Figure 19. For kinematics, mean absolute differences between the 1
st
and 2
nd
squats from
30-120 FA were within 0.5 for AlCor and TR, and 0.5 mm for MFT and LFT (g 4). Vari-
ability in the kinematic differences decreased at deeper exion.
Discussion
The knee simulator system functioned well for the 1
st
squat but not for the 2
nd
, having prob-
lems during ascent. The actual AL followed the target AL within 20% across the entire 1
st

squat, for most trials (g 18), resulting in smooth load curves. Data were more stable at deep-
est exion. During the 2
nd
squat, errors were as much as 80% away from the target AL. This
instability likely comes from limitations of the knee rig system, or stretching of the xation
on the quadriceps tendon, leading to instability in the feedback loop. The current knee simu-
lator system did not discriminate between double and single squat kinematics. After normal-
izing according to FA, mean kinematic differences (TR, AlCor, MFT, LFT) between single
and double squats in descent were small from 30-120 exion (g 19). The kinematic dif-
ferences were not statistically signicant for this system, and the mean differences remained
close to zero within the ranges shown.
56
Fig 18 Ankle load vs. time plots for 1
st
and 2
nd
squats (13 trials superimposed). Loads are expressed as a ratio to
the target ankle load, which was either 90 N or 130 N.
Fig 19 Mean paired differences between the two squats for each trial (2
nd
squat minus 1
st
squat). n = 13 for 40-
100, n = 10 for 110, and n = 6 for 30 and 120. Graphs shown for tibial internal rotation (TR), adduction
(AlCor), medial femoral translation (MFT), lateral femoral translation (LFT), quadriceps load (QL), and
ankle load (AL).

-1
-0.5
0
0.5
1
20 40 60 80 100 120
Flexion ()
A
l
C
o
r

d
i
f
f
e
r
e
n
c
e

(
)
-3
-2
-1
0
1
2
3
20 40 60 80 100 120
Flexion ()
T
R

d
i
f
f
e
r
e
n
c
e

(
)
-0.6
-0.4
-0.2
0
0.2
0.4
20 40 60 80 100 120
Flexion ()
A
L

d
i
f
f
e
r
e
n
c
e
(
F
r
a
c
t
i
o
n

o
f

t
a
r
g
e
t

a
n
k
l
e

l
o
a
d
)
-4
-2
0
2
4
20 40 60 80 100 120
Flexion ()
Q
L

d
i
f
f
e
r
e
n
c
e

(
R
a
t
i
o

t
o

t
a
r
g
e
t

a
n
k
l
e

l
o
a
d
)
-2
-1
0
1
2
20 40 60 80 100 120
Flexion ()
M
F
T

d
i
f
f
e
r
e
n
c
e

(
m
m
)
-2
-1
0
1
2
20 40 60 80 100 120
Flexion ()
L
F
T

d
i
f
f
e
r
e
n
c
e

(
m
m
)

0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
0 5 10 15 20
Time (s)
A
L

(
R
a
t
i
o

t
o

t
a
r
g
e
t

A
L
)
1
s t
squat AL
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
0 5 10 15 20
Time (s)
Q
L

(
R
a
t
i
o

t
o

t
a
r
g
e
t

A
L
)
2
nd
squat AL
57
v. Optical tracking system accuracy
Introduction
The purpose of this analysis was to determine how accurately the Vicon Motion Capture
System measures marker displacements in expected worst-case laboratory conditions.
Methods
Different Vicon system congurations were analysed for their accuracy in tracking the dis-
placements between two reective markers. The congurations varied the number of cameras
(4-8), layout of cameras (wide or narrow convergence angles), camera lens focal length (6-
12.5 mm) and target distance, target volume (~50-120 cm
3
), reective marker diameter (9.5
to 14 mm), and marker displacement magnitude (0-20 cm). The displacements measured by
the system were compared to the displacements measured by digital hand calipers accurate to
0.01 mm. Two markers were attached to the calipers (g 20), and the calipers were arbitrarily
moved by hand in the motion capture target volume with the markers spaced at a locked
distance. The cameras measured the distance over several seconds of video frames. Then the
markers were displaced by a known amount, as measured by the calipers, and motion cap-
ture system measured the new distances to nd the displacements. Trajectories were not l-
tered in order to give worst-case data. The differences, or errors, between the displacements
measured by the cameras and by the calipers were quantied over each video frame for the
unltered data. Then the worst-case raw data was reanalysed by rst ltering the trajectories
with a Woltring lter func-
tion (MSE = 10) supplied
with Vicon Nexus software,
as recommended by manu-
facturer instructions. The
motion data was analysed in
the same way as described
previously, and new error
values were calculated.
Results
The worst-case Vicon system conguration, with the largest variability in the measurement
errors, was not the conguration used for analysis of the knee kinematics simulator. Howev-
er, the worst-case conguration is presented here. It had an error of (mean SD) 0.06 0.60
mm (n = 7000 frames) without ltering the marker trajectories. When the trajectories were
ltered, the error was reduced, producing errors of 0.03 0.19 mm (n = 7000). Filtering the
data reduced the size and variability of error by approximately 50% and 70%, respectively.
Error curves were visibly smoother (g 21).
Fig 2 Digital hand calipers used as displacement reference, with reective
markers attached.
58
Discussion
In expected worst-case laboratory conditions tested, the Vicon system can measure marker
displacements within 0.60 mm for unltered trajectories and 0.19 mm for ltered trajec-
tories. This assumes that all marker trajectories can be reconstructed throughout the motion
capture trial, meaning markers are not completely blocked from view. Measurement errors
are expected to be smaller for slow movements and when there is minimal camera obstruc-
tion, such as in the camera conguration used for analyzing the knee kinematics simulator.
Use of a Woltring lter (MSE = 10) on raw trajectories was chosen for Vicon data analysis
of anatomic motions.
a b
Fig 21 Example displacement measurement error curves for the Vicon system before (a) and after (b) applying a
Woltring lter on raw marker data.
Measurement error, raw data
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
0 100 200 300 400 500
Time
E
r
r
o
r

(
m
m
)
Measurement error, filtered data
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
0 100 200 300 400 500
Time
E
r
r
o
r

(
m
m
)
59
REFERENCES
unloaded living knee studied by MRI. J Bone Joint Surg 2000; 82-B:1196-1198
femur and the tibia in the unloaded cadaver knee. J Bone Joint Surg. 2000:1189-1195
weight bearing and non-weight bearing knee kinematics using interventional MRI. J Biomech. 2005; 38:269-
276
4. Li G, Rudy T, Sakane M, Kanamori A, Ma C, Woo SL. The importance of quadriceps and hamstrings muscle
loading on knee kinematics and in-situ forces in the ACL. J Biomech 1999; 32:395-400
5. Lu TW, Tsai TY, Kuo MY, Hsu HC, Chen HL: In vivo three dimensional kinematics of the normal knee during
active extension under unloaded and loaded conditions using single-plane uoroscopy. Med Eng Phys. 2008;
doi: 10.1016/j.medengphy. 2008.03.001 in press
6. MacWilliams BA, Wilson DR, DesJardins JD, Romero J, Chao EY. Hamstrings cocontraction reduces inter-
nal rotation, anterior translation, and anterior cruciate ligament load in weight-bearing exion. J Orthop Res.
1999;17(6):817-822.
7. Most E, Axe J, Rubash H, Li G. Sensitivity of the knee joint kinematics calculation to selection of exion axes.
Journal of Biomech. 2004;37(11):1743-1748
8. Wilson DR, Feikes JD, Zavatsky AB, OConnor JJ. The components of passive knee movement are coupled to
exion angle. Journal of Biomech. 2000;33(4):465-473.
9. LaPrade RF, Engebretsen AH. Ly TV, et al: The anatomy of the medial part of the knee. J Bone Joint Surg Am.
2007; 89:2000-2010.
10. LaPrade RF, Ly TV, Wentorf FA et al: The posterolateral attachments of the knee: a qualitative and quantitative
morphologic analysis of the bular collateral ligament, popliteus tendon, popliteobular ligament, and lateral
gastrocnemius tendon. Am. J. Sports Med. 2003; 31; 854-860.
11. Victor J and Hoste D: Image-Based Computer-Assisted Total Knee Arthroplasty Leads to Lower Variability in
Coronal Alignment. Clin Orthop. 2004; 428: 131-139.
12. Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: ap-
plication to the knee. J Biomech Eng. 1983; 105:136-144.
13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measure-
ment. Lancet. 1986;307-310.
60
CHAPTER 3: RESULTS OF THE EX VIVO EXPERIMENT
I. Intra- and inter-observer variability
J. Victor, D. Van Doninck, L. Labey, B. Innocenti, P.M. Parizel and J. Bel-
lemans. How precise can bony landmarks be determined on a CT scan of the
knee? The Knee. 2009; doi:10.1016/j.knee.2009.01.001
Abstract
The purpose of this study was to describe the intra- and inter-observer variability of the regis-
tration of bony landmarks and alignment axes on a Computed Axial Tomography (CT) scan.
Six cadaver specimens were scanned. Three-dimensional surface models of the knee were
created. Three observers marked anatomic surface landmarks and alignment landmarks. The
intra- and inter-observer variability of the point and axis registration was performed. Mean
intra-observer precision ranks around 1 mm for all landmarks. The intra-class correlation
coefcient (ICC) for inter-observer variability ranked higher than 0.98 for all landmarks. The
highest recorded intra- and inter-observer variability was 1.3 mm and 3.5 mm respectively
and was observed for the lateral femoral epicondyle. The lowest variability in the determi-
nation of axes was found for the femoral mechanical axis (intra-observer 0.12 and inter-
observer 0.19) and for the tibial mechanical axis (respectively 0.15 and 0.28). In the hori-
zontal plane the lowest variability was observed for the posterior condylar line of the femur
(intra-observer 0.17 and inter-observer 0.78) and for the transverse axis (respectively 1.89
and 2.03) on the tibia. This study demonstrates low intra- and inter-observer variability in
the CT registration of landmarks that dene the coordinate system of the femur and the tibia.
In the femur, the horizontal plane projections of the posterior condylar line and the surgical
and anatomical transepicondylar axis can be determined precisely on a CT scan, using the
described methodology. In the tibia, the best result is obtained for the tibial transverse axis.
Introduction
The use of Computed Axial Tomography (CT scan) as a medical imaging tool has widespread
applications in the eld of knee surgery. It is routinely used in the diagnosis and treatment
of peri-articular fractures and patellofemoral pathology. In arthroplasty surgery, adoption of
this technology has been slower. The CT scan is nowadays considered the premium tool for
planning and evaluation of lower limb alignment
1
, and this can be attributed to the develop-
ment of technological applications like computer navigation and robotic surgery. These tech-
nological achievements put accurate medical imaging to the forefront of orthopedic surgery
and research of the knee
2-8
. In the eld of total knee arthroplasty (TKA), the CT scan serves
different applications. Surgeons use a CT scan in a conventional way during the pre-operative
stage, to plan the position of the femoral component in the horizontal plane
9-11
. In the post-
operative stage, the use of a CT scan is a routine tool in the evaluation of failed TKA
12
, as
rotational malalignment of the femoral component has been determined as a main cause of
61
poor clinical outcome after TKA
12-21
. In image-based computer-assisted surgery, the CT scan
provides three-dimensional anatomic details
2-4, 7
. Novel techniques use CT-based patient-
specic templating to achieve the desired alignment in TKA without the use of conventional
alignment jigs. Finally, in vivo kinematic research of the native knee relies on CT
6
, or MRI
6,22

derived bone models. Those are used for model registration-based three-dimensional kine-
matic measurements, computed from sequential two-dimensional X-ray images.
In all of the above-mentioned clinical applications, surface-derived anatomical landmarks
provide the link between the CT scan data and surgically relevant references that can be
found by visualization or palpation during the operation. In addition, for the surgical naviga-
tion and patient-specic templating applications, the CT scan is used to dene the common
coordinate system, providing the surgeon the frontal, sagittal and horizontal plane of the
femur and the tibia. It is fair to question the ability to accurately identify the surface-derived
anatomical references and the reference points needed to provide the common coordinate
system that denes the three above-mentioned clinical planes. Relatively few publications
addressed this issue. Most studies concentrate on the relative position of different axes
15-17,
23-26
. Only few evaluate intra- or inter-observer variability
26-31
. To our knowledge, no study has
investigated a full set of surface-derived landmarks and alignment landmarks for inter- and
intra-observer variability.
In order to avoid semantic confusion, the following denitions are used. Accuracy is dened
as the closeness of a given measurement to the actual value for the variable considered.
Precision is dened in terms of the measurement error, as the deviation of a set of repeated
measurements from an arbitrary value
32
. As such, two observers can be very precise in their
measurements (small measurements errors) but very inaccurate because of a consistent posi-
tive or negative error. Applied to this study, previous work has shown that a calibrated CT
scan is a highly accurate tool.
The objectives of this study were two-fold:
1. To evaluate the intra- and inter-observer precision in locating reference points on a sur-
face reconstruction of the femur and the tibia, based on CT scans of fresh frozen ampu-
tated leg specimens.
2. To evaluate the intra- and inter-observer precision of the corresponding axes, relevant for
surgical use.
Materials and Methods
Six unpaired fresh frozen amputated legs (3 right, 3 left) were analysed, using a helical CT
scan (General Electric Lightspeed VCT, Milwaukee, WI, USA). The specimens were ob-
tained from 1 female and 5 male Caucasian subjects, aged between 78y and 87y old when
they deceased. The images were obtained at 120 kV and 450 mA, with a slice thickness of
1.25 mm and a pitch of 0.5 mm/rev. Raw data were processed using a bone lter. The CT
scans were analysed using Mimics 11.02 and its MedCAD module (Materialise, Haas-
rode, Belgium) to create the surface reconstruction and identify the bony landmarks. Three
observers participated in the study: one experienced orthopedic surgeon (JV), one medical
student (DVD) and one engineer (LL). The surgeon dened the set of relevant landmarks and
provided the two other observers with a denition and a brief teaching session. Afterwards,
62
Chapter 3: Results of the ex vivo experiment
the three observers analysed the CT scans independently. Two observers (DVD and LL)
performed all analyses three times with a minimum interval of one week for obtaining intra-
observer repeatability. The thresholding feature in Mimics was used to dene two masks (one
for the distal femur and one for the proximal tibia and bula). Lower and higher threshold
values were dened manually. The masks were then cropped to the peri-articular areas of the
bones and edited to separate the different bones. Finally, the masks were converted into 3D
models for identication of the anatomical landmarks.
Anatomical landmarks of the femur (g 1)
Femoral Hip Centre (FHC): centre of best-t sphere to the head of the femur
Femoral Knee Centre (FKC): most anterior point in the middle of the femoral notch on
a caudal to cranial view of the femur, aligning the hip centre with the roof of the femoral
notch.
Femoral Medial Condyle Centre (FMCC): centre of the best-t sphere to the medial con-
dyle
Femoral Lateral Condyle Centre (FLCC): centre of the best-t sphere to the lateral con-
dyle.
Femoral Medial Epicondyle (FME): most anterior and distal osseous prominence over the
medial aspect of the medial femoral condyle
33
.
Femoral Medial Sulcus (FMS): depression on the bony surface slightly proximal and
posterior to FME
33
.
Femoral Lateral Epicondyle (FLE): the most anterior and distal osseous prominence over
the lateral aspect of the lateral femoral condyle
35
.
Femoral Trochlea Proximal (FTP): deepest point of the trochlear groove on the 3D model
of the femur, aligned along the femoral mechanical axis (FMAx).
Femoral Medial Condyle Posterior (FMCP): the most posterior point of the medial con-
dyle on the 3D model of the femur, aligned along the FMAx.
Femoral Lateral Condyle Posterior (FLCP): the most posterior point of the lateral condyle
on the 3D model of the femur, aligned along the FMAx.
Fig 1 Three-dimensional model of the distal femur in frontal and lateral view. Abbreviations of the relevant sur-
face and alignment points are shown on the image. For the denitions, see text.
63
Anatomical landmarks of the tibia (g 2)
Tibial Ankle Centre (TAC): the centre of the best-t circle of the tibial plafond.
Tibial Knee Centre (TKC): the midpoint between the two tibial spines projected on the
bony surface, identied by viewing the 3D model of the tibia from cranial along the tibial
shaft axis.
Tibial Medial Condyle Centre (TMCC): the centre of the best-t circle around the edge of
the cortex of the medial tibial plateau
27
.
Tibial Lateral Condyle Centre (TLCC): the centre of the best-t circle around the edge of
the cortex of the lateral tibial plateau
27
.
Tibial Medial Condyle Posterior (TMCP): the most posterior point of the medial tibial
plateau, on a cranial view, aligned along the tibias shaft axis.
Tibial Lateral Condyle Posterior (TLCP): the most posterior point of the lateral tibial
plateau, on a cranial view, aligned along the tibial shaft axis.
Tibial Tubercle Anterior (TTA): the most anterior point of the tibial tuberosity, on a cra-
nial view, aligned along the tibial shaft axis.
Fig 2 Three-dimensional model of the proximal tibia in frontal and lateral view. Abbreviations of the relevant
surface and alignment points are shown on the image. For the denitions, see text.
Consequently, we obtained seven sets of coordinates (three analyses by two of the three
observers, one analysis by one observer) for the 17 landmarks in each of the six specimens.
Intra- and inter-observer variability was expressed as the distance between the mean position
of a landmark to the observed position of the landmark
34
.
For intra-observer precision, the mean positions of the landmarks
P (x, y, z)
and the distances
D
i of the observed position to that mean position were dened as follows (subscripts 1,2 and
3 refer to the different observations with 1 week interval):
64
The mean value and standard deviation of
D
i, was then calculated for each landmark as a
measure of the overall intra-observer variability for that landmark.
For inter-observer precision, the mean positions of the landmarks
( )
, , P x y z
were calculated
using the means of the coordinates found by each observer, giving the following formulas
(subscripts 1, 2 and 3 refer now to the respective observers):
Mean value and standard deviation of the three
D
i obtained from observers 1, 2 and 3 were
calculated for each landmark as a measure of the overall inter-observer variability for that
landmark.
To be able to discriminate between precisions along the relevant anatomical axes, a coor-
dinate frame was dened for the femur and the tibia, based on the mean positions of the
selected landmarks. For the femur, the femoral mechanical axis (FMAx) was dened as the
line joining the femoral knee centre and the femoral hip centre (FHC-FKC). The frontal
plane was dened as the plane that contains the FMAx and is parallel to the line joining the
medial and lateral centers of the femoral condyles. The horizontal axis was dened as the
perpendicular line to the FMAx in the frontal plane, containing the femoral knee centre. The
horizontal plane contains the horizontal axis and is perpendicular to the frontal plane. The
sagittal axis was dened as the line perpendicular to the FMAx and the horizontal axis and
passes through the knee centre.
For the tibia the tibial mechanical axis (TMAx) was dened as the line joining the centre of
the tibial plateau and the centre of the ankle (TKC-TAC). The frontal plane of the tibia was
dened as the plane containing the TMAx and parallel to line joining the medial and lateral
tibial condylar centre. The horizontal axis of the tibia was dened as the perpendicular line to
the TMAx in the frontal plane, passing through the centre of the tibial knee centre. The hori-
zontal plane of the tibia is perpendicular to the frontal plane and contains the tibial horizontal
axis. The sagittal axis of the tibia was dened as the line perpendicular to the TMAx and the
horizontal axis, passing through the tibial knee centre. All measured coordinates of all land-
marks were transformed into these coordinate frames to evaluate reproducibility along the
three Cartesian axes of the bones.
1 2 3
1 2 3 1 2 3
1 2 3
3
3 3
3
x x x
x
P P P y y y
P y
z z z
z
+ +
=
+ + + +
= =
+ +
=

( ) ( ) ( )
2 2 2
i i i i i
D P P x x y y z z = = + +

1 2 3
1 2 3 1 2 3
1 2 3
3
3 3
3
x x x
x
P P P y y y
P y
z z z
z
+ +
=
+ + + +
= =
+ +
=

( ) ( ) ( )
2 2 2
i i i i i
D P P x x y y z z = = + +

65
In a nal step, the intra- and inter-observer variation of the femoral and tibial axes was
quantied, based on the mean deviation of their dening landmarks. It was assumed that the
errors in the coordinates of the landmarks were independent and random and that simple er-
ror propagation estimations could therefore be used. This was done for the mechanical axes
of femur and tibia (FMAx and TMAx) and for the axes with surgical relevance to rotational
alignment, with the following denitions.
Anatomical transepicondylar axis: FME-FLE.
Surgical transepicondylar axis: FMS-FLE.
Femoral posterior condylar line: FMCP-FLCP.
Femoral transverse axis: FMCC-FLCC.
Femoral trochlear antero-posterior axis: FKC-FTP.
Tibial posterior condylar line: TMCP-TLCP
Tibial transverse axis: TMCC-TLCC.
Tibial Tubercle axis: TKC-TTA.
For the measurement of intra- and inter-observer angular differences in the rotation axes of
femur and tibia, a geometrical projection on the horizontal plane of the femur and the tibia
was respectively carried out.
For each of the considered landmarks positions, we evaluated the intra-class correlation coef-
cient (ICC) for multiple measurements by different observers on different specimens
34
. By
denition, the ICC is evaluated according to the following formulation:
2
2
b
ICC

Where the total variance of measurements by different observers is

2
on different subjects,
and the variance between subjects is
2
b
. ICC values range from 0 to 1, indicating better

agreement as the value approaches 1. An ICC value higher than 0.75 indicates excellent
agreement. The statistical analysis was performed using Matlab R2008a (The MathWorks,
Natick, Massachusetts, USA). For all recorded distances and angles, mean values, maximum
values and standard deviations are reported.
Results
The magnitudes of intra-observer and inter-observer variability for each landmark are shown
in Figures 3 and 4 respectively. The observed mean values, maximum values and standard
deviations are displayed separately. Mean intra-observer variability for all landmarks is situ-
ated around 1 mm (range: 0.4 mm 1.4 mm). All joint centres (FHC, FKC, TKC, TAC) and
condyle centres (FLCC, FMCC, TLCC, TMCC) can be identied with a mean variability of
less than 1 mm. The femoral epicondyles and sulcus, as well as the posterior points on the
tibial condyles are least reliable with mean variabilities larger than 1 mm.
Inter-observer variability is larger than intra-observer variability and more different amongst
landmarks, but is still acceptable (range: 0.3 mm 3.5 mm). Again, the joint centres are
most reliable with mean inter-observer variabilities of less than 1 mm, with the exception of
the tibial knee centre. The posterior points on the tibial condyles, the tibial tubercle and the
femoral lateral epicondyle are least reliable with mean inter-observer variabilities of more
than 2 mm.
66
Fig 3 Intra-observer variability in the registration of the landmarks on the tibia and the femur, shown as mean
value, maximum value and standard deviation.
Fig 4 Inter-observer variability in the registration of the landmarks on the tibia and the femur, shown as mean
value, maximum value and standard deviation.

4 b b l h f h l d k h b d h f h l l
67
Table 1 shows the mean and the standard deviation of the distance from observed position
to the mean position for all landmarks, split along the three anatomical axes. In general,
landmarks that are located on a bony surface can be identied very reliably in the direction
perpendicular to the surface. The variability is usually almost twice as large in the directions
tangent to the surface.
The ICC values for all dened landmarks fall in a range between 0.986 and 1, showing
that observer agreement and reliability for all landmarks is excellent. The statistical results
conrm that the joint centres (with the exception of the tibial knee centre) and the posterior
points on the femoral condyles are most reliable. (ICC = 1) The epicondyles, the medial
sulcus, the posterior points of the tibia, the tibial knee centre and the tibial tubercle have a
slightly lower ranking (respective ICC values: 0.99;0.99;0.99;0,99;0,98).
Table 1 Intra- and inter-observer distances to the observed position to the mean position for all landmarks in 3D and
split along the anatomical axes.
The resulting angular variation between
the different axes could be computed,
based on the dened landmarks.(Table 2)
The mechanical axes in femur and tibia
can be determined very accurately due to
the reliability of the landmarks on which
they are based and the large distance be-
tween the dening points. (FMAx 0.05
intra- and 0.08 inter-observer, TMAx
0.15 and 0.28 respectively). Of the
axes relevant for rotational alignment,
the trochlear antero-posterior axis is least
reliable (mean inter-observer deviation
Table 2 Intra- and inter-observer variability of
angular deviations of the femoral and tibial axes in
the horizontal plane
68
of 2), while the posterior condylar line is most reliable with a mean inter-observer devia-
tion of 0.5. The anat TEA and surg TEA fall in between the two aforementioned axes. In the
tibia, the transverse axis as dened recently by Cobb et al.
27
, shows a mean intra- and inter-
observer variability of respectively 1.44 and 1.66. The two other axes that dene rotation
are less reliable: posterior condylar line respectively 1.37 and 3.16, and the tubercle axis
1.09 and 2.42. A graphical representation is shown in gure 5.
Fig 5 Graphical representation of the accuracy of the registration of the important reference points and axes for
dening rotation in the femur and the tibia. The dark area is a representation of the mean error (enlarged
for better visualization, scale in the legend), the lighter grey area represents the mean error + 1 standard
deviation (SD).
Discussion
Amongst clinicians, the CT scan is often considered the ultimate precision tool in measuring
alignment in the lower limb
1
. The outcome of a given procedure in terms of alignment or po-
sition is often described as a comparison to a reference value, obtained from a CT scan
8,36,37
.
It must be emphasized that the actual reference value (plane, axis or point) remains unknown
and determination of points and axes on a CT scan is subject to inter- and intra-observer vari-
ability. As appears from our results, the intra- and inter-observer variability of the landmarks
that dene the coordinate system of the femur and the tibia, is low. This is fundamental, as
it is the basis for all applications of CT data in the clinical setting
2-8
. However, some of the
study weaknesses have to be understood. First, surface modeling and landmark registration
occurred in optimal circumstances on dedicated computer stations, after studying the recent
anatomic literature and with anatomic drawings at hand. It is clear that this is not the real
life clinical setting where surgeons often work under substantial time constraints. Also, we
picked the tool that is most suited for imaging bone and providing Cartesian coordinates, the
CT scan. It is unclear whether the same accuracy could be achieved in using an MRI scan. As
the cartilage contours can be dened in much greater detail on MRI scans, this tool is more
suitable for patient specic templating
38
and model registration-based three-dimensional ki-
nematic measurements
6,22
. Because of the different qualities, with the CT scan being better
for dening the bony surface and the MRI being better for dening the cartilage surface,
some research groups have used the combination of both for optimal imaging
6
.
Failure to obtain correct alignment in total knee arthroplasty leads to inferior results and early
revisions
12-21, 39-43
. Errors can occur at different levels: application of a wrong reference deni-
69
tion (e.g. the direct use of the posterior condylar line for rotational alignment of the femoral
component), individual variability in the subjects (e.g. dysplasia of the lateral condyle), the
radiological or surgical location of reference landmarks (e.g. locating the epicondyles), in-
strumental errors (e.g. mechanical play), and execution errors (e.g. xation of cutting blocks
and making of the bone cuts). This study only deals with one of those items: the precision of
locating reference landmarks on a CT scan.
Nishihara et al.
7
reported the accuracy of registration in terms of position and angle to be 0.8
mm and 0.6 degrees of bias with 0.2 mm and 0.3 degrees of root-mean-square in the femur,
and 0.5 mm and 0.4 degrees of bias with 0.2 mm and 0.3 degrees of root-mean-square in
the tibia. The aim of this study was to determine the precision of intra- and inter-observer
measurement on a CT scan in a clinically relevant setting: how reproducible is the location
of relevant surface points and axes? This information can help in the development of surgical
navigation algorithms, patient specic cutting blocks, and pre-operative surgical planning.
The precision in locating certain landmarks on a CT scan cannot be extrapolated to the pre-
cision of locating landmarks intra-operatively. Several authors have emphasized the im-
portant inter- and intra-observer variability in the surgical location of the femoral epicon-
dyles
8,10,24,25,29,44
. Even in more idealized circumstances, using cadavers with or without soft
tissues there is signicant variability among observers
36, 37,45-47
. Yau et al.
37
reported high inter-
observer variability in the detection of the anatomic epicondylar axis using 5 cadavers and
surgical navigation: 9 of maximum error due to medial epicondyle registration error and 7
due to the lateral epicondyle registration error. In a similar experiment, Stckl et al.
47
reported
the inter-observer variability as a 95
th
percentile of the distances between the clinical registra-
tions, compared to the CT registrations. For the medial epicondyle, the reported distance was
14.9 mm in the antero-posterior direction and 18.7 mm in the proximal-distal direction. For
the lateral epicondyle, the reported distance was 15.7 mm and 19 mm respectively. Compared
to this reported variability in intra-operative landmark registration, the CT based registration
proves to be superior, as shown in our results.
In addition, there is clinical literature evidence that the use of a pre-operative CT scan offers
opportunities to enhance surgical precision in TKA
9-11
, and increased use of this tool is to
be expected in the future in an attempt to avoid outliers in post-operative alignment. Recent
publications
11,48
conrm the clinical trend to include pre-operative CT scans in the planning
of the procedure. Knowledge of the precision of landmark allocation is mandatory to further
improve surgical outcomes.
The mean intra- and inter-observer error for all landmarks that involve denition of the co-
ordinate system (FHC, FKC, FMCC, FLCC for the femur and TAC, TKC, TLCC, TMCC for
the tibia) in this study is less than 1mm with the exception of the TKC inter-observer value
being 1.8mm. The maximum intra- and inter-observer error for these landmarks is 2.1mm
with the exception of the TKC inter-observer error being 3.3 mm. Given the distance between
the centre of the ankle and the centre of the tibial plateau, the maximum angular error is only
0.34. It can be concluded that the CT scan is a safe tool to dene the coronal, sagittal and
horizontal plane of the femur and the tibia. The inter-observer ICC ranked higher than 0.98
for all landmarks. Of those landmarks that dene rotation of the femur, the lateral epicondyle
was least reproducible. In the pre-operative planning, the surgical transepicondylar axis is
often considered an optimal reference for horizontal plane alignment of the femoral compo-
nent
23, 40
. In the post-operative evaluation of component alignment, the epicondylar axis is the
70
only remaining landmark for dening the rotational position of the femoral component. Wai
Hung et al.
31
compared the CT registration error of the epicondylar axis to anatomic dissec-
tion and observed a mean 2.4 error with conventional CT versus a signicantly higher error
of 2.9 when a three-dimensional reconstruction was used. This could be explained by the in-
ferior quality of the reconstruction or by the fact that a third dimension is taken into account.
Consequently, it is important to consider the split error along the three Cartesian coordinate
axes. As the epicondylar axis serves as a reference in the horizontal plane, registration er-
ror of the lateral epicondyle will have the greatest impact along the sagittal (AP) axis of the
femur. The mean intra- and inter-observer error for the lateral epicondyle is respectively 0.5
and 0.8 mm along this axis. The maximum values are respectively 1.3 mm and 1.8 mm. It ap-
pears that most of the error for the lateral epicondyle is observed along the vertical axis (PD).
Mean intra- and inter-observer values along this axis are respectively 1.3 mm and 3.3 mm,
and maximum values 4.2 mm and 8.7 mm. This three dimensional analysis of error explains
why the anatomical and surgical trans-epicondylar axis show little angular intra- and inter-
observer variability when projected on the horizontal plane of the femur (Table 2).
The lowest intra- and inter-observer variability is observed for the femoral posterior condylar
line (respectively 0.17 and 0.78). The highest intra- and inter-observer variability is found
for the trochlear antero-posterior axis (respectively 1.35 and 3.26). This axis is the most
difcult to dene precisely on a CT scan and it is the only axis related to rotational alignment
that exceeds the clinically accepted 3 threshold for its maximum angular error. As such, it
cannot be regarded a reliable landmark.
At the level of the tibia, the three axes that dene rotation are the posterior condylar line, the
tibial transverse axis and the tibial tubercle axes. Of those, the tibial transverse axis shows
the least intra- and inter-observer variability, respectively 1.44 and 1.66. This conrms the
ndings of Cobb et al.
27
, who rst dened this axis as a reliable landmark for describing rota-
tion of the tibia. Both the tubercle axis and the posterior condylar line have a maximum error
exceeding 3 and cannot be recommended as reliable landmarks.
Conclusion
This study demonstrates low intra- and inter-observer variability in the CT registration of
landmarks that dene the coordinate system of the femur and the tibia. In the femur, the
horizontal plane projections of the posterior condylar line and the surgical and anatomical
transepicondylar axis can be determined precisely on a CT scan, using the described meth-
odology, and can be recommended as reliable landmarks. In addition, the posterior condylar
line is a hard reference, easily located during surgery, allowing to bridge the gap between
the CT scan and real femoral geometry. In the tibia, the least variability is found in the tibial
transverse axis. Further research is needed to determine how precise this axis can be recon-
structed on the real tibial geometry during surgery.
71
REFERENCES
1. Chauhan SK, Scott RG, Breidahl W, Beaver RJ. Computer-assisted knee arthroplasty versus a conventional jig-
based technique. A randomised, prospective trial. J Bone Joint Surg Br. 2004; 86: 372-377
2. Amiot LP, Poulin F. Computed tomography-based navigation for hip, knee, and spine surgery. Clin Orthop Relat
Res. 2004: 77-86
3. Bathis H, Perlick L, Luring C, Kalteis T, Grifka J. [CT-based and CT-free navigation in knee prosthesis implan-
tation. Results of a prospective study]. Unfallchirurg. 2003; 106: 935-940
4. Bellemans J, Vandenneucker H, Vanlauwe J. Robot-assisted total knee arthroplasty. Clin Orthop Relat Res.
2007; 464: 111-116
5. Jazrawi LM, Birdzell L, Kummer FJ, Di Cesare PE. The accuracy of computed tomography for determining
femoral and tibial total knee arthroplasty component rotation. J Arthroplasty. 2000; 15: 761-766
6. Moro-oka TA, Hamai S, Miura H, Shimoto T, Higaki H, Fregly BJ, et al. Dynamic activity dependence of in
vivo normal knee kinematics. J Orthop Res. 2008; 26: 428-434
7. Nishihara S, Sugano N, Ikai M, Sasama T, Tamura Y, Tamura S, et al. Accuracy evaluation of a shape-based reg-
istration method for a computer navigation system for total knee arthroplasty. J Knee Surg. 2003; 16: 98-105
8. van der Linden-van der Zwaag HMJ, Valstar ER, van der Molen AJ, Nelissen RGHH. Transepicondylar axis
accuracy in computer assisted knee surgery: A comparison of the CT-based measured axis versus the CAS-
determined axis. Computer Aided Surgery. 2008; 13: 200 - 206
9. Geiger F, Parsch D. Intraoperative assessment of femoral component rotational alignment in total knee arthro-
plasty. Archives of Orthopaedic and Trauma Surgery. 2008; 128: 267-270
10. Kinzel V, Scaddan M, Bradley B, Shakespeare D. Varus/valgus alignment of the femur in total knee arthroplasty.
Can accuracy be improved by pre-operative CT scanning? Knee. 2004; 11: 197-201
11. Galaud B, Beauls P, Michaut M, Abadie P, Fallet L, Boisrenoult P. Distal femoral torsion: comparison of CT
scan and intra-operative navigation instruments during total knee arthroplasty. A report of 70 cases. Rev Chir
Orthop Reparatrice Appar Mot. 2008; 94:573-579.
12. Berger RA, Crossett LS, Jacobs JJ, Rubash HE. Malrotation causing patellofemoral complications after total
knee arthroplasty. Clin Orthop Relat Res. 1998: 144-153
13. Akagi M, Matsusue Y, Mata T, Asada Y, Horiguchi M, Iida H, et al. Effect of rotational alignment on patellar
tracking in total knee arthroplasty. Clin Orthop Relat Res. 1999: 155-163
14. Anouchi YS, Whiteside LA, Kaiser AD, Milliano MT. The effects of axial rotational alignment of the femoral
component on knee stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens.
Clin Orthop Relat Res. 1993: 170-177
15. Hanada H, Whiteside LA, Steiger J, Dyer P, Naito M. Bone landmarks are more reliable than tensioned gaps in
TKA component alignment. Clin Orthop Relat Res. 2007; 462: 137-142
16. Matsuda S, Miura H, Nagamine R, Urabe K, Hirata G, Iwamoto Y. Effect of femoral and tibial component posi-
tion on patellar tracking following total knee arthroplasty: 10-year follow-up of Miller-Galante I knees. Am J
Knee Surg. 2001; 14: 152-156
17. Miller MC, Berger RA, Petrella AJ, Karmas A, Rubash HE. Optimizing femoral component rotation in total
knee arthroplasty. Clin Orthop Relat Res. 2001: 38-45
18. Mochizuki RM, Schurman DJ. Patellar complications following total knee arthroplasty. J Bone Joint Surg Am.
1979; 61: 879-883
19. Olcott CW, Scott RD. The Ranawat Award. Femoral component rotation during total knee arthroplasty. Clin
Orthop Relat Res. 1999: 39-42
20. Rhoads DD, Noble PC, Reuben JD, Mahoney OM, Tullos HS. The effect of femoral component position on
patellar tracking after total knee arthroplasty. Clin Orthop Relat Res. 1990: 43-51
21. Yoshii I, Whiteside LA, White SE, Milliano MT. Inuence of prosthetic joint line position on knee kinematics
and patellar position. J Arthroplasty. 1991; 6: 169-177
22. Defrate LE, Papannagari R, Gill TJ, Moses JM, Pathare NP, Li G. The 6 degrees of freedom kinematics of the
knee after anterior cruciate ligament deciency: an in vivo imaging analysis. Am J Sports Med. 2006; 34: 1240-
1246
23. Grifn FM, Math K, Scuderi GR, Insall JN, Poilvache PL. Anatomy of the epicondyles of the distal femur: MRI
analysis of normal knees. The Journal of Arthroplasty. 2000; 15: 354-359
24. Oberst M, Bertsch C, Lahm A, Wuerstlin S, Holz U. Regression and correlation analysis of preoperative versus
intraoperative assessment of axes during navigated total knee arthroplasty. Computer Aided Surgery. 2006; 11:
87 - 91
72
25. Talbot S, Bartlett J. The anterior surface of the femur as a new landmark for femoral component rotation in total
knee arthroplasty. Knee Surgery, Sports Traumatology, Arthroscopy. 2008; 16: 258-262
26. Yoshino N, Takai S, Ohtsuki Y, Hirasawa Y. Computed tomography measurement of the surgical and clinical
transepicondylar axis of the distal femur in osteoarthritic knees. The Journal of Arthroplasty. 2001; 16: 493-
497
27. Cobb JP, Dixon H, Dandachli W, Iranpour F. The anatomical tibial axis: reliable rotational orientation in knee
replacement. J Bone Joint Surg Br. 2008; 90-B: 1032-1038
28. Han H, Chang C, Seong S, Lee S, Lee M. Evaluation of anatomic references for tibial sagittal alignment in total
knee arthroplasty. Knee Surgery, Sports Traumatology, Arthroscopy. 2008; 16: 373-377
29. Jerosch J, Peuker E, Philipps B, Filler T. Interindividual reproducibility in perioperative rotational alignment
of femoral components in knee prosthetic surgery using the transepicondylar axis. Knee Surg Sports Traumatol
Arthrosc. 2002; 10: 194-197
30. Nofrini L, Slomczykowski M, Iacono F, Marcacci M. Evaluation of Accuracy in Ankle Center Location for
Tibial Mechanical Axis Identication. Journal of Investigative Surgery. 2004; 17: 23 - 29
31. Wai Hung CL, Wai Pan Y, Kwong Yuen C, Hon Bong L, Lei Sha LW, Ho Man SW. Interobserver and Intraob-
server Error in Distal Femur Transepicondylar Axis Measurement with Computed Tomography. The Journal of
Arthroplasty. In Press, Corrected Proof:
32. Prez-Prez A, Alesan A, Roca L. Measurement error: Inter-and Intraobserver Variability. An Empiric Study.
International Journal of Anthropology. 1990; 5: 129-135
33. LaPrade RF, Engebretsen AH, Ly TV, Johansen S, Wentorf FA, Engebretsen L. The Anatomy of the Medial Part
of the Knee. J Bone Joint Surg Am. 2007; 89: 2000-2010
34. Bland JM, Altman DG. Measurement error. BMJ. 1996; 313: 744
35. LaPrade RF, Ly TV, Wentorf FA, Engebretsen L. The Posterolateral
Attachments of the Knee: A Qualitative and Quantitative Morphologic Analysis of the Fibular Collateral Liga-
ment, Popliteus Tendon, Popliteobular Ligament, and Lateral Gastrocnemius Tendon. Am J Sports Med. 2003;
31: 854-860
36. Robinson M, Eckhoff DG, Reinig KD, Bagur MM, Bach JM. Variability of landmark identication in total knee
arthroplasty. Clin Orthop Relat Res. 2006; 442: 57-62
37. Yau WP, Leung A, Liu KG, Yan CH, Wong LLS, Chiu KY. Interobserver and Intra-observer Errors in Obtain-
ing Visually Selected Anatomical Landmarks During Registration Process in Non-Image-Based Navigation-
Assisted Total Knee Arthroplasty. The Journal of Arthroplasty. 2007; 22: 1150-1161
38. Lombardi AV Jr, Berend KR, Adams JB. Patient-specic approach in total knee arthroplasty. Orthopedics.
2008;31:927-930
39. Jeffery RS, Morris RW, Denham RA. Coronal alignment after total knee replacement. J Bone Joint Surg Br.
1991; 73: 709-714
40. Jenny JY, Boeri C. Low reproducibility of the intra-operative measurement of the transepicondylar axis during
total knee replacement. Acta Orthop Scand. 2004; 75: 74-77
41. Kessler O, Lacatusu E, Sommers MB, Mayr E, Bottlang M. Malrotation in total knee arthroplasty: effect on
tibial cortex strain captured by laser-based strain acquisition. Clin Biomech (Bristol, Avon). 2006; 21: 603-609
42. Ritter MA, Faris PM, Keating EM, Meding JB. Postoperative alignment of total knee replacement. Its effect on
survival. Clin Orthop Relat Res. 1994: 153-156
43. Werner FW, Ayers DC, Maletsky LP, Rullkoetter PJ. The effect of valgus/varus malalignment on load distribu-
tion in total knee replacements. J Biomech. 2005; 38: 349-355
44. Perrin N, Stindel E, Roux C. BoneMorphing versus freehand localization of anatomical landmarks: Conse-
quences for the reproducibility of implant positioning in total knee arthroplasty. Computer Aided Surgery. 2005;
10: 301 - 309
45. Fuiko R, Kotten B, Zettl R, Ritschl P. [The accuracy of palpation from orientation points for the navigated
implantation of knee prostheses]. Orthopade. 2004; 33: 338-343
46. Siston RA, Goodman SB, Patel JJ, Delp SL, Giori NJ. The high variability of tibial rotational alignment in total
knee arthroplasty. Clin Orthop Relat Res. 2006; 452: 65-69
47. Stoeckl B, Nogler M, Krismer M, Beimel C, Moctezuma de la Barrera J-L, Kessler O. Reliability of the Tran-
sepicondylar Axis as an Anatomical Landmark in Total Knee Arthroplasty. The Journal of Arthroplasty. 2006;
21: 878-882
48. Aglietti P, Sensi L, Cuomo P, Ciardullo A. Rotational position of femoral and tibial components in TKA using
the femoral transepicondylar axis. Clin Orthop Relat Res. 2008; 466:2751-2755.
73
II. Horizontal plane geometry of the distal femur
J. Victor, D. Van Doninck, L. Labey, F. Van Glabbeek, P.M. Parizel,
J. Bellemans. A common reference frame for describing rotation of the
distal femur. J Bone Joint Surg. 91-B; 2009: 683-690.
Abstract
The understanding of rotational alignment of the distal femur is essential in total knee re-
placement to ensure that there is correct placement of the femoral component. Many refer-
ence axes have been described, but there is still disagreement about their value and mutual
angular relationship. Our aim was to validate a geometrically-dened reference axis against
which the surface-derived axes could be compared in the axial plane. A total of 12 cadaver
specimens underwent CT after rigid xation of optical tracking devices to the femur and
the tibia. Three-dimensional reconstructions were made to determine the anatomical surface
points and geometrical references. The spatial relationships between the femur and tibia in
full extension and in 90 of exion were examined by an optical infrared tracking system.
After co-ordinate transformation of the described anatomical points and geometrical refer-
ences, the projection of the relevant axes in the axial plane of the femur were mathematically
achieved. Inter- and intraobserver variability in the three-dimensional CT reconstructions
revealed angular errors ranging from 0.16 to 1.15 for all axes except for the trochlear
axis which had an interobserver error of 2. With the knees in full extension, the femoral
transverse axis, connecting the centres of the best matching spheres of the femoral condyles,
almost coincided with the tibial transverse axis (mean difference -0.8, SD 2.05). At 90 of
exion, this femoral transverse axis was orthogonal to the tibial mechanical axis (mean dif-
ference -0.77, SD 4.08). Of all the surface-derived axes, the surgical transepicondylar axis
had the closest relationship to the femoral transverse axis after projection on to the axial
plane of the femur (mean difference 0.21, SD 1.77). The posterior condylar line was the
most consistent axis (range -2.96 to -0.28, SD 0.77) and the trochlear anteroposterior axis
the least consistent axis (range -10.62 to +11.67, SD 6.12). The orientation of both the pos-
terior condylar line and the trochlear anteroposterior axis (p = 0.001) showed a trend towards
internal rotation with valgus coronal alignment.
Introduction
Poor outcomes and major complications after total knee replacement (TKR) have been linked
directly to errors in rotational alignment of the components.
1-5

Despite the clinical importance of correct rotational alignment the dilemma of choosing the
correct surgical references remains unsolved. For the rotational alignment of the femoral
component, two systems prevail. In the dependent resection, or tensioned-gaps technique, the
surgeon performs ligamentous release to balance the knee in extension followed by resection
of the posterior femoral surfaces parallel to the prepared cut tibial surface, after applying
equal loads to the medial and lateral compartments.
6,7
In the measured resection technique,
a surface-derived reference axis of the femur is used as a guide to determine the position of
74
the femoral component in the axial plane. Several different reference axes have been pro-
posed including the posterior condylar line (PCL),
8,9
the surgical transepicondylar axis (surg
TEA),
10-12
the anatomical transepicondylar axis (anat TEA)
13
and the trochlear anteroposte-
rior axis (TRAx).
14

There are four obstacles to surgical consistency as follows: semantic confusion over the de-
nition of the correct axis, the natural variability of the reference axes, the inter- and intra-
observer variability in the intra-operative determination of these references and the practical
execution of the cuts.
There is a distinct difference between the desired alignment in the axial plane of the femoral
component and the natural alignment of the distal femur. This difference can be explained
by the fact that the normal natural tibial plateau has a varus conguration.
15
The perpendicu-
lar coronal cut of the tibia will change this angle. Consequently, the femoral component will
not be correctly aligned to provide an adequate exion gap if it follows the natural surface
anatomy. An external rotational compensation to the same degree as the correction of the
tibial cut in the coronal plane will generally be advocated. Therefore since the literature re-
fers to the correct alignment of the axial plane of the femoral component, we always refer
to the adapted alignment of the femoral component in the axial plane, which differs from
the natural situation.
The numerous studies which have compared axes, relevant for describing rotational align-
ment, have all used the PCL, surg TEA, anat TEA and TRAx in their examinations.
9-14,16-19

These axes have been employed because the surface points describing them are palpable dur-
ing surgery. From a methodological standpoint, however, the results are difcult to compare,
given the absence of a xed reference frame.
We wished to add more information to improve the understanding of the three-dimensional
relationships of the distal femur by including geometrically dened reference axes, based on
anatomical descriptions of the distal femur
20
and the proximal tibia.
21
In addition, we have
used the dependent relation between the femur and tibia in extension and exion to validate
these geometrical axes, respectively called the femoral transverse axis (FTAx) and tibial
transverse axis (TTAx).
Our aim was threefold:
1. To determine the natural angular variability of previously studied surface-derived axes
(PCL, anat TEA, surg TEA, TRAx), compared with the FTAx, projected on the axial
plane of the femur. The surface-derived axis with the smallest variability is theoretically
the most useful axis for surgical reference.
2. To measure the angle between the FTAx and TTAx projected on the axial plane of the
femur with the knee extended. Since there is no rotational freedom between the tibia and
femur in full extension, this is a reliable position to measure the rotational relation be-
tween the two bones. If they are parallel in the axial plane, these axes can be validated as
reliable rotational landmarks.
3. To determine the relationship between the aforementioned femoral axes and the tibial me-
chanical axis at 90 of exion. It is hypothesised that the FTAx will be perpendicular to
this tibial mechanical axis, indicating neutral varus/ valgus alignment at 90 of exion.
75
With the approval of the ethical committee, 12 cadaver limbs showing no signs of prior
trauma or surgery, ligament imbalance or arthritis were disarticulated at the level of the hip
in ten subjects (one matched pair, ten non-matched). The specimens were deep frozen and
optical stereotactic reference frames were rigidly attached. Volumetric CT scans on a 64-row
multidetector CT scanner (General Electric Lightspeed VCT, Milwaukee, Wisconsin) were
performed using the following settings: 120 kVp, 450 mA; rotation time 0.5; pitch 0.516/1,
speed 20.62 mm per rotation; helical thickness 1.25 mm; interval 0.8 mm. The CT scan data
were fed into a three-dimensional software system (Mimics 11.02 and its MedCAD module;
Materialise, Haasrode, Belgium) for further determination of the relevant axes and surface
points. Important surface landmarks were identied, using the quantitative morphological
description by LaPrade et al.
22,23
Geometrically important points such as the centre of the hip
or the centre of the femoral condyles were determined by tting a sphere in the outer surface
mask of the respective structures.
A repeatability analysis was done on the rst six specimens to evaluate the reliability of
localising the bony landmarks used to dene the four major rotational axes and the FTAx.
Three observers (DVD, LL and JV) used bone surface reconstructions of the scanned joints
to identify the landmarks three times at a minimum interval of one week (for intra-observer
precision) and independently (for inter-observer precision). Based on the recorded differ-
ences in position for each landmark between repeated measurements and between observers,
the angular deviation for each axis with respect to its mean orientation was calculated. Meth-
odological analysis was carried out as described by Bland and Altman.
24

The Cartesian co-ordinate system of the CT data was transformed into a new co-ordinate
system for the femur and the tibia. The axial plane of the femur (or horizontal plane in a
standing subject) was dened as the plane perpendicular to the mechanical axis of the femur
and comprising the centre of the knee. The latter was dened as the most posterior point of
the trochlear groove (top of the femoral notch). The FTAx connects the centres of the two
best matching spheres of the medial and lateral femoral condyle. The anat TEA connects
the medial epicondyle with the lateral epicondyle. The surg TEA connects the medial sulcus
with the lateral epicondyle. The TRAx, commonly referred to as Whitesides line,
19
connects
the cranial and caudal deepest point of the trochlear groove (g. 1). At the tibia, the centre
of the medial and lateral condyles were computed according to the description by Cobb et
al,
21
tting the best matching circle to the cortical outline of the proximal tibia, 20 mm below
the tibial spines. The line connecting the medial and lateral condylar centres, dened as the
anatomical tibial axis by Cobb et al,
21
was referred to as the TTAx to maintain consistent
terminology with the femur (g. 2).
76
The cadaver specimens were moved from full extension to 90 of exion, while the optical
stereotactic reference frames were followed by ve previously calibrated infrared cameras
(Vicon Motion Systems, Los Angeles, California). Since the stereotactic optical frames were
rigidly attached to the femur and the tibia, and their spatial relation to the important reference
points and axes on the bones was documented by the CT scan, the position of all points and
axes on the femur relative to the tibia could be computed.
The projection of the FTAx and TTAx on the axial plane was performed with the specimens
in full extension (g. 3). The relation between the tibial mechanical axis and the various
femoral axes was computed with the knees in 90 of exion (g. 4).
Statistical analysis. All datasets were checked for normal distribution. Pairwise comparisons
between the different axes were performed with Statistica software, (Statsoft Inc., Tulsa,
United States) using Students t-test. Correlations were evaluated using Pearsons test. Statis-
tical signicance was set at a p-value " 0.05.
Fig 1 Diagram showing the axes which have been
studied in relation to the rotational alignment
of the distal femur. The posterior condylar
line (PCL) is the tangent of the posterior
part of the medial and lateral condyle. The
anat TEA connects the medial epicondyle to
the lateral epicondyle and the surg TEA the
medial sulcus to the lateral epicondyle. The
trochlear axis (Whitesides line) connects the
deepest point of the trochlea to the highest
point in the notch, with the femur in caudal
view.
Fig 2 Diagram showing the TTAx as dened
by Cobb et al.
21
77
Results
Intra- and interobserver variability.
We found a high level of reproducibility for both intra- and inter-observer error in our meth-
od. A repeatability analysis was done on the rst six specimens to evaluate the precision
of marking the four surface derived rotational axes and the femoral transverse axis. Three
observers used bone surface reconstructions of the scanned joints to identify the landmarks
three times (for intra-observer precision) and independently (for inter- observer precision).
Based on the recorded differences in position for each landmark between repeated measure-
ments and between observers, the mean angular difference for each axis can be calculated.
24
The results are shown in Table I.
FTAx PCL Anat TEA Surg TEA !TRAx
"#
intra
1.15 0.16 0.52 0.57 1.24
"#
inter
1.15 0.57 1.15 1.15 2.07
Table I Mean intra- and interobserver differences for the ve femoral axes in degrees.
Fig. 3 Diagram showing the projection of the mean
TTAx and the mean FTAx on the axial plane
of the femur. The projection of both axes in
the extended knee virtually coincide.
Fig. 4 Diagram showing the relation between the
tibial mechanical axis and the distal femur at
90 of exion.
78
Variability and discrimination of the surface-derived axes.
The projections on the axial plane of the resulting angles between the four major rotational
axes and the FTAx per specimen are given in Table II. The surg TEA was almost parallel to
the FTAx with a mean relative external rotation of 0.21 (SD 1.77). The anat TEA had a mean
external rotation relative to FTAx of 3.40 (SD 1.66). The PCL had a mean relative internal
rotation of -1.41 (SD +0.77). The perpendicular to the trochlear axis had a mean external ro-
tation of 1.39 (SD 6.12). The angle between the PCL and the FTAx had the least variability
with an SD of 0.77 (-2.96 to -0.28). The highest variability was found in the angle between
the perpendicular to the TRAx and the FTAx with an SD of 6.12 (10.62 to +11.67) (Table
II). Comparison of the mean values for the relative angles between the FTAx and the PCL,
the surg TEA and the anat TEA showed clear distinctions with p-values of <0.005 for the
PCL/anat TEA and anat TEA/surg TEA and p-values = 0.008 for the PCL/surg TEA (t-tests).
By contrast, comparison between the FTAx and the perpendicular to TRAx showed poor dis-
crimination from the other axes because of the high SD (TRAx/PCL p = 0.012; TRAx/anat
TEA p = 0.251; TRAx/surg TEA p=0.512, all t-test) (g. 5).
Angle relative to FTAx in degrees
SPECIMEN PCL Anat TEA Surg TEA ! TRAx
1 -0.28 3.44 1.10 -0.53
2 -1.23 4.44 1.52 1.23
3 -1.00 6.16 3.31 -0.06
4 -1.02 2.22 -1.73 4.69
5 -2.96 1.73 0.95 -7.21
6 -1.18 5.23 -0.76 -0.71
7 -0.39 3.21 0.42 5.65
8 -2.01 0.35 -2.91 0.03
9 -2.22 1.86 -2.00 -10.62
10 -1.44 3.26 0.81 6.08
11 -1.93 4.54 1.59 6.41
12 -1.22 4.34 0.24 11.67
MIN -2.96 0.35 -2.91 -10.62
MAX -0.28 6.16 3.31 11.67
MEAN -1.41* 3.40* 0.21* 1.39
SD 0.77 1.66 1.77 6.12
Table II Angles in degrees, formed between the four major rotational axes of the distal femur and the femoral
transverse axis (FTAx).
Negative values indicate a relative internal rotation of the variable in relation to FTAx, positive values a
relative external rotation of the variable to FTAx. Abbreviations: Posterior Condylar Line = PCL, Anatomic Tran-
sepicondylar Axis = Anat TEA, Surgical Transepicondylar axis = Surg TEA, Perpendicular to the Trochlear Axis
=! TRAx . *p<0,005
79
A strong correlation was found between the coronal alignment in full extension (dened as
the angle between the femoral mechanical and tibial mechanical axis in the frontal plane) and
the trochlear axis orientation (r = 0.81, p = 0.001) indicating an increasing internal rotation
of the TRAx orientation with increasing valgus. A weak correlation was found with the PCL
orientation (r = 0.36, p = 0.257) and no correlation with the surg (r = 0.02, p = 0.95) or anat
TEA (r = -0.01, p = 0.97).
Angle between the projected FTAx and TTAx at full extension.
The projected centre of the medial condyle on the axial plane was found to be a mean of 1.6
mm (-2.6 to -5.9) anterior to the tibial transverse axis, compared with a mean of 2.83 (-6.1 to
+0.66) posterior to the projected centre of the lateral femoral condyle. This resulted in a mean
angle between the FTAx and TTAx of 0.8 (-1.8 to +3) (g. 3).
Relation between the femoral axes and the tibial mechanical axis at 90 of exion.
For the analysis of the tibiofemoral alignment in relation to the distal femoral geometry, the
angle between the tibial mechanical axis and the femoral rotation axes was computed. A
graphical illustration of the statistics is shown in Figure 6 and the results are summarised in
Table III.
The tibial mechanical axis was close to the perpendicular to the FTAx, surg TEA and parallel
to the TRAx at 90 of exion. The perpendicular to the PCL of the femur was at a mean of
2.17 of internal rotation relative to the TMAx. As expected, based upon the distal femoral
Fig. 5 Box-and-whisker plots of the four axes,
relative to FTAX clearly showing the small
spread of the posterior condylar line (PCL)
and the large interindividual variability of the
trochlear axis.
Fig. 6 Box-and-whisker plots of perpendiculars to
the FTAx (pFTAx). The posterior condylar
line (PCL), the anat TEA (point TEA, the surg
TEA (psurg TEA) and the TRAx (pTRAx)
Relative to the tibial mechanical axis at 90
of exion.
D
e
g
r
e
e
s
D
e
g
r
e
e
s
80
geometrical values, the greatest variability was found in the angle between TMAx and TRAx
(SD 5.96, range -6.86 to +12.7). There was a signicant difference at p < 0.005 between
the means for FTAx/PCL, FTAx/anat TEA, PCL/anat TEA and surg TEA/anat TEA (t-test,
all p < 0.005). No signicant difference was found between FTAx/surg TEA (p = 0.93) and
between perpendicular to TRAx/surg TEA (p = 0.52).
Angle relative to TMA in degrees at 90 of exion
SPECIMEN ! FTAx ! PCL ! anat TEA ! surg TEA TRAx
1 -3.96 -4.24 -0.52 -2.86 -4.49
2 -4.48 -5.71 -0.04 -2.96 -3.25
3 -1.12 -2.12 5.04 2.19 -1.18
4 -3.37 -4.39 -1.15 -5.10 1.32
5 2.01 -0.95 3.74 2.96 -5.20
6 -0.55 -1.73 4.68 -1.31 -1.26
7 -3.10 -3.49 0.11 -2.68 2.55
8 6.11 4.10 6.46 3.20 6.14
9 3.76 1.54 5.62 1.76 -6.86
10 2.72 1.28 5.98 3.53 8.80
11 -8.24 -10.17 -3.70 -6.65 -1.83
12 1.03 -0.19 5.37 1.27 12.70
MIN -8.24 -10.17 -3.70 -6.65 -6.86
MAX 6.11 4.10 6.46 3.53 12.70
MEAN 0.77 -2.17 2.63 -0.55 0.62
SD 4.08 3.79 3.46 3.48 5.96
Table III Angles in degrees, formed between TRAx and the perpendiculars to the FTAx, PCL, anatTEA and
surgTEA and the tibial mechanical axis. Negative values indicate a relative internal rotation of the
variable in relation to the TMAx, positive values a relative external rotation of the variable to the
TMAx.
Discussion
One of the most important errors leading to revision in TKR is malalignment.
25
Small errors
may be acceptable and may not interfere with the function of a TKR.
4
In regard to rotational
malalignment of the femoral component, excessive internal rotation has been related to pain,
stiffness and patellar instability
1,2,4,26-29
whereas excessive external rotation has been associ-
ated with exion instability,
30
increased shear forces on the patella
5
and varus alignment in
exion.
16
81
Unfortunately, the surgical process of positioning the implant is subject to errors. Failure to
align a component properly with a desired axis can occur at three different levels. Occasion-
ally, the desired reference axis such as the femoral mechanical axis, is not visible in vivo and
a secondary axis, generally the centre of the intramedullary canal, is chosen to serve as a
guide during surgery. Ideally, this secondary axis has a reliable angular relation, in statistical
terms a small SD, to the desired axis. A rst level of error is the individual variability in the
angular relation between the desired axis and the surgical secondary axis. The second level
of error is related to the intra-operative determination of the secondary surgical axis, in other
words, the ability of the surgeon to locate accurately and reproducibly anatomical landmarks
which lead to the secondary surgical axis. Previous studies have emphasised the difculties
surgeons face in this area.
31,32
The third level of error is related to the positioning and xa-
tion of the cutting block and the execution of the cut with the saw. The subject of our study
is at the rst level, namely, the denition of the desired axis and its relation to the secondary
surgical axis.
The main weakness of our study was the small sample size, caused by the extensive techni-
cal set-up and lengthy duration of the experiment. However, the distribution of the coronal
alignment of the specimens fell within the range of that of previously published series,
15
sug-
gesting that it was a representative sample.
We decided to compare the position of the proposed desired axis of the femur (FTAx) to the
tibia, since a description of the rotation of the distal femur is intrinsically related to the tibia.
In the extended knee, we hypothesised parallelism in the axial plane between the projected
FTAx and TTAx. In the exed knee at 90, we hypothesised that the tibial mechanical axis
was perpendicular to the FTAx, resulting in neutral varus/valgus alignment.
16
As shown in the
results and illustrated in Figure 3, the FTAx and TTAx virtually coincide as they are projected
on the axial plane of the femur with the knee in extension. This is a strong indication that
the TTAx, the anatomical tibial axis of Cobb et al,
21
is a reliable landmark for describing the
rotation of the tibia and the FTAx for the femur. In addition, at 90 of exion, the mean value
for the extension of the tibial mechanical axis crossing the FTAx was almost a right angle,
indicating neutral alignment. This was a second argument for accepting the FTAx as a correct
anatomical reference to describe the neutral rotation of the femur and to guide the surgeon in
the placement of the femoral component during TKR.
With this FTAx as a reference, the surface derived secondary surgical axes can be exam-
ined. The results presented in Table I and Figure 5 indicate that the PCL, the anat TEA and
the surg TEA can be discriminated in the axial plane projection, based on their different ori-
entation in individuals. The PCL displays the smallest interindividual variability with an SD
of less than 1. It is internally rotated relative to the anat and surg TEA. This is in line with
previous published ndings.
9-14,17,19

Of the two transepicondylar axes, the axial plane projection, the surg TEA, is almost parallel
to the FTAx. The anat TEA is consistently more externally rotated than the surg TEA, in our
results relatively by a mean of 3.18. This again conrms previously published work.
11,12,19

Concerns can be raised about the trochlear axis (Whitesides line). Despite the fact that
the mean value of this axis shows an almost orthogonal orientation to the FTAx, the range
(-10.62 to +11.67) is considerable. If used as a single reference for the femoral component,
it could lead to unacceptable outliers in rotational alignment. In addition, a signicant corre-
82
lation was found between the amount of internal rotation of the trochlear axis and the valgus
alignment in the coronal plane, which could limit the value of this reference in the valgus
setting. This contrasts with the original description of this trochlear axis by Arima et al,
14
in
which they proposed this axis for the knee with valgus deformity. However, that study did
not use valgus knees but
30
normal femora obtained from cadavers. No further information on
coronal alignment was provided. They reported a range for the angle between the PCL and
TRAx as -1 to +10. In our study we found a range between the PCL and TRAx of -8.4 to
12.9. Three knees with valgus alignment form the lower end of this range (specimen 1, 5
and 9, Table II), indicating they had, relative to FTAx, more internal rotation of TRAx than
of the PCL. Talbot and Bartlett
31
recently compared the orientation of the anterior surface of
the femur, just cranial of the trochlea, with the TRAx in the axial plane and reported a reli-
able correlation between them but a signicant variability with the PCL. They deduced that
the TRAx was a reliable axis for rotational alignment. Without devaluating their contribu-
tion, their conclusion clearly illustrates how the nding of a strong correlation between two
surface derived axes of the distal femur can lead to a false deduction of an optimal reference
axis. Since the relationship to the tibia is lacking, conclusions cannot be drawn on the opti-
mal rotational alignment, but only on the morphological description of the distal femur. We
believe that the distinction between these two concepts is crucial in the discussion of axial
alignment in TKR.
The consequences of our results for surgical practice are as follows. It is obvious that a com-
mon frame of reference is to be used if we wish to discuss the achievement of optimal rota-
tion of the femoral component. The denition of the femoral transverse axis as a geometrical
reference can support future analysis of deformity in patients. Our study deals only with the
rst level of error, being the individual variability in angular relationships. The second level
of error, the inter- and intrasurgeon variability of nding the landmarks during surgery may
be even more important, and should be considered in the overall evaluation of the surgical
value in using these landmarks. Siston et al,
32
studying surgical reliability in determining
rotational alignment of the femoral component found that the use of the TRAx (Whitesides
line) produced the largest variability with an SD of 7.6 and a range of - 12 to +15, with
little improvement when the epicondyles were referenced. The lowest range was reported
when surgeons used the PCL with a posterior referencing guide (-9 to +11). Middleton
and Palmer
33
reported on the reliability and reproducibility of the TRAx in the distal femora
in 50 cadavers. A wide range of rotation of the TRAx relative to the surg TEA was reported
from -10 to +12 with an SD of 4.7. The ability to mark accurately visually selected ana-
tomical landmarks in a surgical navigation system with a CT scan as a baseline reference has
been investigated revealing maximum errors in the TEA in the axial plane of 9.1 because
of registration errors of the medial femoral epicondyle and of 7.2 due to registration errors
of the lateral femoral epicondyle.
34
This conrms the ndings of other reports.
35,36
Efforts to
reduce surgical incisions compromise surgical consistency in locating the TRAx and anat
TEA further.
37

Based on our ndings, we believe that the variability of the TRAx is too wide to allow it to
serve as a reliable secondary reference during TKR. The additional intra- and inter-surgeon
variability in the surgical determination of this axis strengthens this. Unfortunately, the latter
argument on the difcult intra-operative localisation of landmarks equally applies to the anat
83
TEA and the surg TEA.
32,34-38
Despite the relatively consistent relationship to the FTAx our
ndings indicate that surgical precision is insufcient to rely on these axes for positioning the
femoral component in the axial plane.
By contrast, the PCL, which had the closest and most constant relationship to the FTAx in
our study, is a landmark which can be located accurately and reliably during surgery.
9,38
It
could be used to align the femoral component in the axial plane during surgery, provided that
the mean compensation of 1.41 of external rotation, according to our results, was applied.
However, it is important that the surgeon recognises that the articular surface may be eroded
due to wear and may induce considerable changes to the described morphology. This would
require a further intraoperative compensation if the surgeon relied on the PCL. Recently,
Aglietti et al
39
described a relationship between the orientation of the PCL and coronal align-
ment, showing more internal rotation in the severe valgus knee and more external rotation
in the severe varus knee. The different wear patterns which are observed in the severely
deformed varus or valgus knee can certainly explain this. In our series, we found a weak and
non-signicant correlation (r=0.36; p = 0.257) between the PCL and coronal alignment, al-
lowing us to defend the role of the PCL as a reliable landmark for knees with no or moderate
deformity, and without posterior condylar cartilage or bone wear.
REFERENCES
1. Mochizuki RM, Schurman MD. Patellar complications following total knee arthroplasty. J Bone Joint Surg
[Am] 1979;61-A:879-83.
2. Rhoads DD, Noble PC, Reuben JD, Mahoney OM, Tullos HS. The effect of femoral component position on
patellar tracking after total knee arthroplasty. Clin Orthop 1990;260:43-51.
3. Sikorski JM. Alignment in total knee replacement. J Bone Joint Surg [Br] 2008;90-B:1121-7.
4. Berger RA, Crossett LS, Jacobs JJ, Rubash HE. Malrotation causing patellofemoral complications after total
knee arthroplasty.
5. Miller MC, Berger RA, Petrella AJ, Karmas A, Rubash HE. Optimizing femoral component rotation in total
knee arthroplasty. Clin Orthop 2001;392:38-45.
6. Freeman MA, Todd RC, Bamert P, Day WH. ICLH arthroplasty of the knee: 1968-1977. J Bone Joint Surg [Br]
1978;60-B:339-44.
7. Scuderi GR, Insall JN. Posterior stabilized prosthesis. Orthop Clin North Am 1989;20:71-8.
8. Hungerford DS, Kenna RV. Preliminary experience with a total knee prosthesis with porous coating used with-
out cement. Clin Orthop 1983;176:95-107.
9. Laskin RS. Flexion space conguration in total knee arthroplasty. J Arthroplasty 1995;10:657-60.
10. Grifn FM, Insall JN, Scuderi GR. The posterior condylar angle in osteoarthritic knees. J Arthroplasty
1998;13:812-15.
11. Berger RA, Rubash HE, Seel MJ, Warren HT, Crosset LS. Determining the rotational alignment of the femoral
component in total knee arthroplasty using the epicondylar axis. Clin Orthop 1993;286:40-7.
12. Asano T, Akagi M, Nakamura T. The functional exion-extension axis of the knee corresponds to the surgical
epicondylar axis. J Arthroplasty 2005;20:1060-7.
13. Poilvache PL, Insall JN, Scuderi GR, Font-Rodriguez DE. Rotational landmarks and sizing of the distal femur
in total knee arthroplasty. Clin Orthop 1996;331:35-46.
14. Arima J, Whiteside LA, McCarthy DS, White S. Femoral rotational alignment, based on the anteroposterior
axis, in total knee arthroplasty in a valgus knee: a technical note. J Bone Joint Surg [Am] 1995;77-A:1331-4.
15. Moreland JR, Bassett LW, Hanker GJ. Radiographic analysis of the axial alignment of the lower extremity. J
Bone Joint Surg [Am] 1987;69-A:745-9.
16. Hanada H, Whiteside LA, Steiger J, Fyer P, Naito M. Bone landmarks are more reliable than tensioned gaps in
TKA component alignment. Clin Orthop 2007;462:137-42.
84
17. Yoshioka Y, Siu D, Cooke TD. The anatomy and functional axes to the femur. J Bone Joint Surg [Am] 1987;69-
A:873-80.
18. Grifn MF, Insall JN, Scuderi GR. Accuracy of soft tissue balancing in total knee arthroplasty. J Arthroplasty
2000;15:970-3.
19. Yoshino N, Takai S, Ohtsuki Y, Hirasawa Y. Computed tomography measurement of the surgical and clinical
transepicondylar axis of the distal femur in osteoarthritic knees. J Arthroplasty 2001;16:493-7.
20. Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement. 1: the shapes and relative movements of the
femur and the tibia in unloaded cadaver knee. J Bone Joint Surg [Br] 2000;82-B:1189-95.
21. Cobb JP, Dixon H, Dandachli W, Iranpour F. The anatomical tibia axis: reliable rotational orientation in knee
replacement. J Bone Joint Surg [Br] 2008;90-B:1032-8.
22. LaPrade RF, Engebretsen AH, Ly TV, et al. The anatomy of the medial part of the knee. J Bone Joint Surg [Am]
2007;89-A:2000-10.
23. LaPrade RF, Ly TV, Wentorf FA, Engebretsen L. The posterolateral attachments of the knee: a qualitative and
quantitative morphologic analysis of the bular collateral ligament, popliteus tendon, popliteobular ligament,
and lateral gastrocnemius tendon. Am J Sports Med 2003;31:854-60.
ment. Lancet 1986;1:307-10.
25. Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM. Why are total knee arthroplasties failing today?
Clin Orthop 2002;404:7-13.
26. Yoshii I, Whiteside LA, White SE, Milliano MT. Inuence of prosthetic joint line position on knee kinematics
and patellar position. J Arthroplasty 1991;6:169-77.
27. Anouchi YS, Whiteside LA, Kaiser AC, Milliano MT. The effects of axial rotational alignment of the femoral
component on knee stability and patellar tracking in total knee arthroplasty demonstrated on autopsy specimens.
Clin Orthop 1991;287:170-7.
28. Akagi M, Matsuse Y, Mata T, et al. Effect of rotational alignment on patellar tracking in total knee arthroplasty.
Clin Orthop 1999;366:155-63.
29. Matsuda S, Miura H, Nagamine R, et al. Effect of femoral and tibial component position on patellar tracking fol-
lowing total knee arthroplasty: 10-year follow-up of Miller-Galante I knees. Am J Knee Surg 2001;14:152-6.
30. Olcott CW, Scott RD. Femoral component rotation during total knee arthroplasty. Clin Orthop 1999;367:39-
42.
31. Talbot S, Bartlett J. The anterior surface of the femur as a new landmark for femoral component rotation in total
knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2008;16:258-62.
32. Siston RA, Patel JJ, Goodman SB, Delp SL, Giori NJ. The variability of femoral rotational alignment in total
knee arthroplasty. J Bone Joint Surg [Am] 2005;87-A:2276-80.
33. Middleton FR, Palmer SH. How accurate is Whitesides line as a reference axis in total knee arthroplasty? Knee
2007;14:204-7.
34. Yau WP, Leung A, Liu KG, Wong LLS, Chiu KY. Interobserver and intra-observer errors in obtaining visually
selected anatomical landmarks during registration process in non-image based navigation-assisted total knee
arthroplasty. J Arthroplasty 2007;22:1150-61.
35. Jenny JY, Boeri C. Low reproducibility of the intra-operative measurement of the trans-epicondylar axis during
total knee replacement. Acta Orthop Scand 2004;75:74-7.
36. Kinzel V, Ledger M, Shakespeare D. Can the epicondylar axis be dened accurately in total knee arthroplasty?
Knee 2005;12:293-6.
37. Yau WP, Leung A, Liu KG, et al. Errors in the identication of the transepicondylar and anteroposterior axes
of the distal femur in total knee replacement using minimally-invasive and conventional approaches: a cadaver
study. J Bone Joint Surg [Br] 2008;90-B:520-6.
38. Galaud B, Beauls P, Michaut M, et al. Distal femoral torsion: comparison of CT scan and intra-operative navi-
gation instruments during total knee arthroplasty: a report of 70 cases. Rev Chir Orthop Reparatrice Appar Mot
2008;94:573-9 (in French).
39. Aglietti P, Sensi L, Cuomo P, Ciardullo A. Rotational position of femoral and tibial components in TKA using
the femoral transepicondylar axis. Clin Orthop 2008;466:2751-5.
85
III. Horizontal plane geometry of the proximal tibia
The resulting angles between the projections of the two major rotational axes on the hori-
zontal plane and the Tibial Transverse Axis (TTAx) per specimen are displayed in table I.
A graphical illustration is shown in g 1. The posterior tangent line of the tibia has a mean
internal rotation relative to TTAx of 3,52. The perpendicular to the tibial tubercle axis has a
mean external rotation of 18,81.
The lowest variability is found for the posterior tangent line (PTLtib) with a standard devia-
tion of 2,77 (range -7,68/2,04) versus 3,88 for the tibial tubercle axis (TibTubAx) (range
12,4/22,9)(Table I). As distribution of the angles for TibTubAx was not normal, non-para-
metric statistics were applied to investigate correlation. According to the Spearman rank
order correlation test, there is a very weak correlation between the PTLtib angle en TibTubAx
angle (r=0,111, p>0,05), indicating an important natural variability of the latter.
Spec PTLtib ! Tib Tub Ax
1 -5.17 22.46
2 -4.04 18.35
3 -4.76 20.76
4 -4.82 21.50
5 -2.28 19.34
6 -1.30 22.42
7 -0.19 22.90
8 -6.18 21.80
9 -7.68 13.12
10 -2.30 12.40
11 -5.51 16.83
12 2.04 13.88
Min -7.68 12.40
Max 2.04 22.90
Mean -3.52 18.81
Std 2.77 3.88
Table I Angular relation between the Tibial Transverse Axis on one hand and the posterior tangent line of the tibia
(PTLtib) and the perpendicular to the tibial tubercle axis (!TibTubAx) on the other hand, in degrees. Nega-
tive values display relative internal rotation, positive values external rotation
Fig 1 Top view on the tibial plateau for graphical representa-
tion of the important rotation axes. The posterior tangent
line connects the most posterior parts of the tibial con-
dyles, the tibial transverse axis connects the centres of
the two best t circles to the tibial condyles and the tibial
tubercle axis connects the knee centre to the tip of the
tibial tubercle.
86
IV. Component Position
All components were positioned with the help of a surgical navigation system (Brainlab,
Feldkirchen, Germany). The orientation in the coronal and sagittal plane was executed as
planned on the virtual planning station
(g1), aiming at orthogonal position to
FMAx and TMAx in the coronal plane
and orthogonal to FMAx for the femur
and 3 exion to TMAx for the tibia, in
the sagittal plane. The rotational align-
ment of the femoral component was
based on the posterior condylar line
(PCL), adding 3 of external rotation.
The rotation of the tibia component was
based on the femoral component, with
the knee in full extension. The positions
of the components were calculated,
based on the post-operative CT scan
and dened in the separate coordinate
system of the femur and the tibia.
Fig 1 Screenshot of alignment planning and actual
update after making the bone cuts
A. CORONAL PLANE
The measured positions (varus/valgus) of the femoral and tibial component on the post-op-
erative CT scan are shown in Table I. The mean actual position of the tibial and femoral
component in the coronal plane was within one degree of error from the desired position (0).
The maximum deviation was 3.2 on the tibia and 2.5 on the femur.
The mean composite tibiofemoral position was 0.2. This is very close to the measured mean
tibiofemoral coronal alignment of the specimens post-TKA in the Oxford rig, being 0.
The reported differences betweent the composite component alignment in the coronal plane
and the measured alignment of the bones in the coronal plane reect the following cumula-
tive errors:
Surgical navigation system error
Surgical block position error
Surgical bone cutting execution error
Measurement error on post-operative CT scan
Positioning error on Oxford rig (unwanted varus- or valgus stress due to incorrect ce-
mented position of femur or tibia in the containers)
Measurement error with VICON system
Calculation error with coordinate transformations
87
As the reported deviations are small, it can be concluded that the methodological set-up and
sequence of measurements and calculations is robust. In addition, all specimens were posi-
tioned within three degrees of varus/valgus error in the coronal plane, clinically considered
as perfect alignment.
SPEC FEMUR TIBIA Composite ALCor #AlCor-Comp
1 -2.3 2.8 0.5 2.0 1.5
2 -2.0 3.2 1.3 -0.4 -1.6
3 1.3 -1.5 -0.2 -1.0 -0.8
4 2.5 0.4 2.9 2.9 0.0
5 -0.9 -0.2 -1.1 -2.1 -1.0
6 -0.6 1.8 1.3 1.3 0.0
7 0.2 -3.2 -2.9 -2.7 0.2
8 -0.9 1.4 0.5 -0.9 -1.4
9 N/A N/A N/A -0.6 N/A
10 -0.8 -0.6 -1.4 -1.6 -0.2
11 -0.5 0.1 -0.4 1.8 2.2
12 -0.5 2.1 1.7 0.9 -0.8
MIN -2.3 -3.2 -2.9 -2.7 -1.6
MAX 2.5 3.2 2.9 2.9 2.2
MEAN -0.4 0.6 0.2 0.0 -0.2
STD 1.4 1.9 1.6 1.9 1.2
Table I Post-operative position of the femoral and tibial components in the coronal plane in degrees. Negative
values mean valgus, positive values varus. Column 1 shows the position of the femoral component, col-
umn 2 the position of the tibial component, column 3 (composite) the computed theoretical composite
tibiofemoral alignment based on the values of column 1 and 2. Column AlCor shows the actual measured
tibiofemoral coronal alignment with the specimen mounted in the Oxford rig and the nal column shows
the difference between the computed compound alignment and the measured tibiofemoral coronal align-
ment. Values for specimen 9 are incomplete as a fracture of the femur occurred at the end of the kinematic
experiments and the postoperative CT based alignment values are not to be considered reliable anymore.
88
B. SAGITTAL PLANE
As shown in table II, the desired amount of exion in the sagittal plane relative to FMAx
and TMAx exceeded with 2.4 exion on the femur and 1.6 on the tibia. The results are ac-
ceptable from a clinical standpoint, however, clearly not as good as the reported values for
the coronal plane. The standard deviation goes up to 3.2 on the femur and 2.9 on the tibia.
This can be explained in part by the greater technical difculty to determine the position of
the component on a post-operative CT scan due to the shorter dimensions of the components
in the sagittal plane. Also, small differences between the orientation of the bone cuts and the
nal position of the cemented component are more likely to occur in the sagittal plane than in
the coronal plane, due to the specic geometry of total knee components.
SPEC FEMUR TIBIA
1 4.9 2.1
2 10.1 -0.6
3 1.4 4.4
4 2.3 -0.6
5 2.3 1.1
6 -1.1 -3.8
7 4.1 1.0
8 -1.1 7.2
9 N/A N/A
10 1.8 3.2
11 0.4 -0.9
12 1.5 2.3
MIN -1.1 -3.8
MAX 10.1 7.2
MEAN 2.4 1.4
STD 3.2 2.9
Table II Positions of the femoral and tibial component in the sagittal plane in degrees, relative to
respectively FMAx and TMAx. Positive values represent relative exion of the component,
negative values extension.
C. HORIZONTAL PLANE
As discussed in the introduction, horizontal plane alignment poses the biggest challenge for
the femoral component position.
In the experiment, the posterior condylar line (PCL) was used as a reference for rotational
alignment of the femoral component. During the process of datapoint acquistion with the
89
surgical navigation system, the posterior condylar line, the anatomic transepicondylar axis
and the trochlear AP axis were dened and the data were stored.
The accuracy of using the PCL is reected in the postoperative rotational alignment of the
component as determined with the CT scan. As shown in table III, the mean position is 0.3,
approaching closely the goal of 0. The range is small (-4.1 to 2.5), reected in a small
standard deviation of 1.8.
SPEC Rotation FC
1 2.5
2 -1.4
3 0.7
4 -0.1
5 -0.1
6 0.1
7 1.5
8 -4.1
9 N/A
10 1.9
11 1.0
12 1.1
MIN -4.1
MAX 2.5
MEAN 0.3
STD 1.8
In contrast, the surgical accuracy in determining the anatomic transepicondylar axis and the
trochlear AP axis was far less accurate and more variable. Both axes would have induced in-
ternal rotation (respectively -6,2and -4,5, table IV) of the femoral component if they would
have been selected as a reference. The values obtained are signicantly different (Students
t test, p<0,001) from the values obtained from the CT scan. It can be concluded that surgical
navigation does not allow to locate those axes reliably during surgery. This is in line with
previous publications, as discussed in chapter 3, section II.
SPEC ANAT TEA TRAX
1 -1.4 -7.5
2 -3.6 -6.2
3 -9.3 -4.2
4 -5.5 -4.8
5 -10.2 -6.5
6 -9.0 -6.7
7 -6.6 -6.3
8 -5.2 -5.3
9 0.2 -2.1
10 -5.8 -7.0
11 -8.9 -0.3
12 -8.9 3.5
MIN -10.2 -7.5
MAX 0.2 3.5
MEAN -6.2 -4.5
STD 3.3 3.3
Table III Rotational alignment of the
femoral component in degrees,
relative to FTAx, as measured
on the postoperative CT scan
Table IV Angular relation of Anatomic transepi-
condylar axis and trochlear AP axis with
FTAx as determined with surgical naviga-
tion
90
V. Tibiofemoral Kinematics of the Native and the Replaced Knee
A. THE INFLUENCE OF MUSCLE LOAD ON NATIVE TIBIOFEMORAL
KNEE KINEMATICS
J. Victor, L. Labey, P. Wong, B. Innocenti, J. Bellemans. The inuence
of muscle action on tibiofemoral knee kinematics. J Orthop Res. 2009
Abstract
A comparative kinematic study was carried out on six cadaver limbs, comparing tibiofemoral
kinematics in ve different conditions: unloaded, under a constant 130 N ankle load with a
variable quadriceps load, with and without a simultaneous constant 50 N medial and lateral
hamstrings load. In passive conditions, the tibia rotated internally with increasing exion,
to -16 (range -12/-20, SD 3.0). Between 0 40 exion, the medial condyle translated
forwards 4 mm (range 0.8/5.5 mm, SD 2.5 mm), followed by a gradual posterior translation,
totaling -9 mm (range -5.8/-18.5 mm, SD 4.9 mm) between 40 140 exion. The lateral
femoral condyle translated posteriorly with increasing exion completing -25 mm (range
-22.6 -28.2 mm, SD 2.5 mm). Dynamic, loaded measurements mimicking a deep knee
bend, were carried out in a knee rig. Under a xed ankle load of 130 N and variable quad-
riceps loading, tibial rotation (TR) was inverted, mean TR 4.7 (range -3.3/11.8 SD 5.4),
Medial Femoral Translation (MFT) -0.5 mm (range = -4.3/2.4 mm, SD = 2.4 mm), Lateral
Femoral Translation (LFT) 3.3 mm (range = -3.6/10.6 mm, SD = 5.1 mm). As compared
to the passive condition, all excursions were signicantly different: p=0.015, p=0.013, and
p=0.011 for TR, MFT and LFT respectively. Adding hamstrings force, reduced TR, MFT and
LFT signicantly as compared to the passive condition. In general, loading the knee with
hamstrings and quadriceps reduces rotation and translation as compared to the passive condi-
tion. Lateral hamstring action is more inuential on knee kinematics than medial hamstrings
action.
Introduction
Knee kinematics are complex and intriguing and have been studied extensively
1
. The de-
duced model of knee kinematics describes posterior translation of the femoral condyles rela-
tive to the tibia with increasing exion. This posterior translation is greater on the lateral than
on the medial side, leading to relative internal rotation of the tibia with increasing exion.
However, different methodologies reveal different kinematic patterns and the existing litera-
ture is not unanimous in the description of normal knee kinematics. The differences can be
attributed to a number of intrinsic and extrinsic factors. Intrinsic factors relate to the inter-
individual differences. As most studies included very small numbers of specimens, patients
or volunteers, bias cannot be excluded, even with a normal distribution of anatomic features
or kinematic patterns in the population. Extrinsic factors include the practical set-up of the
91
experiment with differences in quadriceps force, hamstrings co-contraction, triceps surae
co-contraction, loads and mechanical constraints imposed on the knee joint. In addition to
experimental variability, the mathematical model used for describing knee kinematics will
inuence the results
2-5
.
It makes sense to discriminate between studies describing passive or unloaded knee kinemat-
ics
6-12
and studies describing loaded kinematics. Loaded ex vivo experiments typically use
loading frames, knee kinematics simulators or robots where the loads applied to the joint are
controlled
13-25
. Loaded in-vitro experiments can use a feedback-loop between ankle load and
the applied quadriceps force, but are arbitrary when it comes to co-contraction of important
muscle groups. The loads exerted by the co-contraction of hamstrings and triceps surae re-
main largely uncertain, which is reected in the large range of applied loads in the existing
literature on in-vitro kinematic experiments
14,16,20,21,24,25,
.
Loaded in- vivo research uses direct MRI imaging
10,26,27
, roentgen stereo photogrammetric
analysis
28
, and 2D uoroscopy with shape matching techniques, based on CT models or
combined CT/MRI models
29-32
. Loaded in-vivo studies are limited due to unknown loading
conditions and variation in the activity performed. In contrast, they mimic better physiologic
activity with correct tuning of muscle co-contractions as the subject adapts to the external
forces imposed by the experimental locomotor task.
The aim of this study is to isolate the impact of muscular action on knee kinematics by keep-
ing all other variables constant during the experiment. In-vitro knee kinematics are described
in passive conditions, with variable quadriceps loads, with and without hamstrings loads on
the medial and on the lateral side. The null hypothesis to be tested is muscle action will not
change knee kinematics.
Six cadaver specimens from 3 male and 3 female subjects underwent full active and passive
testing. Passive motion was performed rst, after attaching the femoral container to the free
rotating hip joint of the knee rig. The specimens were cycled manually ve times from full
extension to maximum exion. After completion of passive measurements, the specimen
was attached to the ankle part of the knee rig through its holding container. Testing was per-
formed at constant speed with a constant vertical ankle load of 130N from about 25 to 120
exion. The higher extension positions could not be tested dynamically as the knee could be
pulled in hyperextension by the quadriceps mechanism, leading to irreversible damage to
the specimen. Quadriceps loaded simulated squats were recorded without hamstrings loads,
and with sequential loading of the medial and lateral hamstrings separately. These conditions
were obtained by releasing or attaching the hamstrings sutures to the constant force springs.
The dynamic knee kinematics data were plotted as functions of exion angle in 10 intervals
from 30-120, except for the passive exion trials which were available from 0-140. The
kinematics from different muscle action conditions was compared using general linear model
analysis of variance (ANOVA) for unequal sample sizes, across the muscle action conditions
and the exion angles. Pairwise comparisons of data were made with the Tukey method. All
statistical tests were performed with computer software (Minitab, State College, PA, USA)
with signicance set at p"0.05.
92
Results
Inter- and intra observer variability of marking the geometrical reference points for den-
ing planes and axes has been published previously
33
. Accuracy and precision of the motion
analysis system, which was used for the kinematic recordings of the markers, was shown to
be on the order of 0.2 to 0.3 mm.
Passive motion
In passive conditions, from 0-130 exion, the tibia rotated internally with increasing ex-
ion, to an average of -16 (range -12 -20, SD 3.0). Between 0 40 exion, the medial
condyle translated forwards 4 mm (range 0.8 5.5 mm, SD 2.5 mm), followed by a gradual
posterior translation, totaling -9 mm (range -5.8 -18.5 mm, SD 4.9 mm) between 40 140
exion. The lateral femoral condyle translated posteriorly with increasing exion completing
-25 mm (range -22.6 -28.2 mm, SD 2.5 mm).(g 1 and Table 1)
93
Fig 1 Translation (mm) and rotation () plots as a function of the exion angle for the ve conditions studied.
MFT: medial femoral condyle centre translation, LFT: lateral femoral condyle centre translation, TR: tibial
rotation, negative values represent internal rotation, positive values external rotation
94

TR []
Average (stdev)
Min / Max
MFT [mm]
Average (stdev)
Min / Max
LFT [mm]
Average (stdev)
Min / Max
Passive
(between 30 and 120 of exion)
-12.3 (3.6) -8.5 (3.0) -21.0 (3.4)
-18.5 / -9.7 -12.9 / -5.2 -24.3 / -17.1
Quad. Only
4.7 (5.4) -0.5 (2.4) 3.3 (5.1)
-3.3 / 11.8 -4.3 / 2.4 -3.6 / 10.6
Quad. & Med. Hamstrings
5.1 (5.1) -2.2 (2.4) 2.2 (4.7)
0 / 13.1 -5.0 / 0.6 -1.2 / 10.4
Quad. & Lat. Hamstrings
-1.2 (5.5) 0.9 (4.0) -0.3 (6.5)
-7.8 / -1.6 -3.5 / 6.1 -6.1 / 8.1
Quad. & both Hamstrings
-0.5 (5.8) -3.5 (2.3) -4.3 (5.0)
-6.2 / 8.8 -6.1 / 0 -7.9 / 4.6
rotation: - = internal, + = external
translation: - = posterior, + = anterior
Table 1 Ranges of movement for tibial axial rotation and femoral condyle centres
Loaded quadriceps only, no hamstrings
From 30 to 70 exion, the tibia was more anterior than during the passive motion.(g 1)
When only the quadriceps was active, tibial axial rotation was inverted. (g 2) Between 30
and 60 of exion, the tibia rotated slightly internally over 2, but beyond 60 of exion, the
tibia rotated externally over 6.5 and ended in neutral axial orientation at 120 of exion.
This behaviour was the result of anterior translation of the medial femoral condyle centre
between 30 and 60 of exion, while the lateral femoral condyle centre remained relatively
stable. Beyond 60 of exion, the lateral femoral condyle centre started moving anteriorly,
while the medial femoral condyle centre remained stable (between 60 and 80) or moved
posteriorly (between 80 and 120, where the movement of this point was almost identical to
its movement in passive conditions).
Overall, the position of the medial femoral condyle was relatively stable with an average
total excursion of -0.5 mm (range = -4.3 +2.4 mm, SD = 2.4 mm). The lateral femoral
condyle translated anteriorly with increasing exion with an average total excursion of 3.3
mm (range = -3.6 +10.6 mm, SD = 5.1 mm). All parameters (TR, MFT and LFT) differ
signicantly from their values during passive exion. (Table 2) Average LFT total excursion
during passive motion in the same exion range was signicantly greater than in this loaded
case without hamstrings (p = 0.011). Average excursion of MFT during passive motion was
also signicantly greater than in this case (p = 0.013).
95
Fig 2 Average tibial rotation for the ve conditions, superimposed
Effect of the hamstrings
Active exion with both hamstrings attached showed little rotation. (g 2 and 3) However,
when no hamstrings were attached or when only the medial hamstrings were attached, the
tibia was more internally rotated. (g 2) Passive exion was signicantly different from ac-
tive exion in all muscle load conditions, with greater difference at deep exion. (Table 2)

Quad. only
Quad. & Med.
Hamstr.
Quad. & Lat.
Hamstr.
Quad. & both
Hamstr.
TR
Passive 0.0147 0.0024 0.0113 0.0148
Quad. Only - 0.9712 0.0171 0.0092
Quad. & Med. Hamstr. - 0.0037 0.0048
Quad. & Lat. Hamstr. - 0.3014
MFT
Passive 0.0129 0.0572 0.0718 0.0317
Quad. Only - 0.1124 0.0263 0.4321
Quad. & Med. Hamstr. - 0.2484 0.0347
LFT
Passive 0.0116 0.0001 0.0005 0.0020
Quad. Only - 0.4308 0.0119 0.1396
Quad. & Med. Hamstr. - 0.0004 0.2811
Table 2 p values of pair-wise comparison across the different load conditions for tibial rotation (TR), medial con-
dyle centre translation (MFT) and lateral condyle centre translation (LFT). Bold values are signicant, grey
font non signicant
96
For the translations, ANOVA detected signicant differences among the conditions for both
MFT and LFT. However, less pair-wise differences among conditions could be detected for
MFT than for LFT. In active exion with the lateral hamstring attached, LFT was closest to
zero. Compared to this case, LFT was 2,2 mm more anterior when only the medial hamstrings
were attached and 3.3 mm when no hamstrings were attached (p =0.0004 and p=0.0119 re-
spectively.). For LFT, passive exion was signicantly different from active exion, under
all load combinations, also with greater difference at deep exion.
Fig 3 Projection of medial and lateral femoral condyle centers on the horizontal tibial plane for the different load
cases
97
Patellar tendon angle
The patellar tendon angle, dened as the angle between the patellar tendon and the tibial
mechanical axis in the sagittal plane was positive between 0-65 exion and negative from
65 to maximum exion (g 4a and b).
Fig 4 a. Surface reconstruc-
tions illustrating the measure-
ment of the patellar tendon angle
to the tibial mechanical axis in
the sagittal plane
b. Angle between patellar
tendon and tibial mechanical axis
in the sagittal plane of the tibia as
a function of exion angle
Discussion
This study aimed at investigating the effect of muscle action on knee kinematics. Based on
the results, the null hypothesis is refuted. Muscle action signicantly changes knee kinemat-
ics and affects the medial and lateral side of the joint differently. Signicant differences are
shown between the passive condition and all loaded conditions, and between the loaded
condition with quadriceps only and loaded condition with quadriceps+both hamstrings or
quadriceps+lateral hamstrings. Adding or removing medial hamstrings action does not in-
duce signicant kinematic changes.
98
Several limitations must be considered when interpreting the results. The specimens could be
weakened by old age or chronic illness. Six specimens are relatively few, given the previous-
ly reported variability in knee kinematics. Tissue quality can degrade over time in the course
of the experiment. The loaded squat is a reproducible motion but does certainly not reect the
full spectrum of knee motions of daily life. The intricate interaction between hamstrings and
quadriceps cannot be fully modeled and the actual forces exerted in-vivo by the hamstring
muscles are unknown. The hamstrings load of 50 N medially and laterally was chosen, based
upon previous biomechanical cadaver studies
14,21,24,25,34
Churchill et al studied knee kinematics based on a loaded rig setting with 100 N ankle load
and 30 N combined hamstrings load
21
. Kwak et al used the relative physiological cross-
sectional area to simulate physiologic muscular co-contraction. For a given quadriceps force
of 534 N they applied 111 N to the biceps femoris, 111 N to the semimembranosus and 45 N
to the pes anserinus muscles
25
. Other authors used forces ranges from 45 N to 90 N for the
hamstrings.
14,34
As we assumed antagonist co-contraction should be inferior to the agonist ac-
tion in physiologic motion, we considered the Kwak hamstrings values to high and decided
to stay in the range proposed by the other authors.
The data obtained in the loaded setting (Quadriceps + Hamstrings) of this in-vitro experi-
ment seem to be at odds with several in-vivo experiments, studying loaded kinematics. A
partial explanation can be found in the weaknesses of the in-vitro set-up, not making use of
the triceps surae action and ankle forces for replicating a physiologic deep knee bend. Also,
the timing and magnitude of hamstrings co-contraction is chosen arbitrarily and might not
resemble the in-vivo setting. As the targeted ankle load in this study was limited to 130 N,
the mean quadriceps load observed at 30 was 263 N, rising to 1149 N at 110 exion, cre-
ating a relatively low quadriceps contraction/hamstrings co-contraction ratio at low exion
angles. In addition, a close reading of the loaded in-vivo studies learns that the activities
studied are often arbitrary because of methodological and technical constraints. In the study
by Nakagawa et al, measurements were made with the subjects lying on the side with the
investigated leg supported by the table. External forces are likely to be small and muscle
forces were probably relatively low.
26
Johal et al reported kinematics during a loaded squat in
an open coil MRI. They described the subjects body position during the experiment in detail.
It appears that the volunteers performed a wall sit with the back supported by a board that
was inclined backwards at 10. Depending on the friction between the back of the subject and
the wall, part of the subjects weight is taken away from the knee joint. Moreover, the contact
force from the wall has an important horizontal component affecting the forces acting on the
knee joint.
27
Moro-Oka et al described three different activities. The greatest tibial internal
rotation was seen in the knee and lunge activities. According to their description, either the
contralateral limb or the upper limbs contributed signicantly in load sharing during these
locomotor tasks, reducing the load on the studied knee. The activity that resembled best the
in-vitro simulated knee bend as performed in our study, was the stair activity. For this in-vivo
activity, the least tibiofemoral rotation and translation was reported.
99
Authors Materials Motion Measurement technique
Blankevoort 4 cadavers series of static positions RSA
Hill 6 cad. & 13 subj. series of static positions MRI
Iwaki 6 cadavers series of static positions MRI
Johal 10 subjects series of static positions MRI
Li 13 cadavers least resistance path, robot MRI
Lu 8 subjects dynamic tests Fluoroscopy CT models
Most 6 cadavers least resistance path, robot Digitised models
Wilson 15 cadavers dynamic tests with rig MRI
Table 3 Overview of published studies on passive knee kinematics
Several authors described kinematics of the knee in the passive setting
6-12
. An overview of
these studies and their different methodologies is shown in Table 3. Figure 5a shows a plot
of the tibial rotations of the current study, superimposed on the previously published data.
Despite the trend being the same for all studies, the actual values show variation depending
on the methodology used. The very stable position of the medial condyle over the exion arc
from full extension to 110 of exion, as shown by the group of Freeman
7,10,27
, is not fully
conrmed in this study. Our kinematics parameters are close to those reported by Most
6
, with
initial forward sliding of the medial femoral condyle to a maximum anterior position at 40 of
exion, followed by a posterior translation, completing -12 mm. On the lateral side, posterior
translation starts immediately, totalling -25 mm. The tibia undergoes internal rotation of 16,
relative to the femur over the full exion arc. Most internal rotation occurs between 0-20
(screw-home) and between 70-110. From 110 to full exion, no further rotation is taking
place, because of equal posterior translation on the medial and lateral side.
Considering the results of this study, the inuence of muscle action on knee kinematics is
undeniable. Recognition of this phenomenon is not new, and the clinical consequences can be
extensive. Our ndings regarding the effect of quadriceps loading conrm previous experi-
ments, based on different methodologies. Beynnon et al reported that quadriceps contraction
strained the ACL at 30 knee exion but not at 90, using in-vivo techniques
35
. Arms et al
reported ACL strain to increase to 45 of exion and to decrease beyond 60 knee exion
36
.
This corresponds with our ndings of anterior translation of the tibia up to 60 knee exion,
followed by posterior translation, if the quadriceps is loaded in isolation. As the isolated
quadriceps contraction can produce forces beyond those required for ACL tensile failure
37
,
the role of hamstrings co-contraction is likely to be of signicant clinical importance. The
hamstring co-contraction has indeed been recognized as a stabilizing factor
38
and a protago-
nist of the anterior cruciate ligament as it limits anterior tibial translation and tibial internal
rotation.
24,25,39
. Renstrm showed in a cadaveric study, using a hydraulic testing machine and
strain transducers an increased strain within the ACL upon quadriceps loading between 0 to
45 in comparison to the passive condition. Simultaneous hamstrings action was not able to
counteract this phenomenon unless the knee exion angle exceeded 30
39
. Mac Williams et al
tested the effect of hamstrings force in a loaded dynamic knee rig experiment on 8 cadavers.
100
They tested two conditions: one loaded condition (hip vertical load and variable quadriceps
load) without hamstrings load and with hamstrings load, equivalent to the vertical load. They
found the hamstrings to signicantly reduce internal tibial rotation, anterior tibial translation
and caused a reversed direction of shear force on the tibia
24
. Similar results were reported
by More et al, using a closed chain experimental set-up combining quadriceps load with a
constant hamstrings load of 0, 45 or 90N
14
. Kwak et al built an knee-joint testing machine
and tested ve cadaveric specimens statically in open chain between 0 and 90 of knee ex-
ion. They tested different muscular loading modes, quadriceps loading, in combination with
or without hamstrings and iliotibial band loading. Despite a greater force being applied on
the medial hamstrings (156N) than on the biceps tendon (111N), the tibia rotated externally
with hamstrings force. The authors explained this by a larger effective moment arm about
the tibial shaft because of the eccentric attachment of the biceps on the bular head. Because
of the important observed effects of the hamstrings, the authors considered the hamstrings
the most effective stabilizers (with respect to anterior cruciate ligament function) of the
knee and warned against the prelevation of hamstrings grafts for ACL reconstruction for that
reason
25
. They did, however, not distinguish between the separate role of the medial ham-
strings versus the biceps. To our knowledge, no study examined sequentially passive motion
versus loaded conditions separating quadriceps, medial and lateral hamstrings force. The
available literature data on tibiofemoral translation obtained in kinematic knee rigs in loaded
conditions are summarized in gure 5b and 5c. The strength of this study is the controlled
measurement of the kinematic effects of sequential modes of loading on the knee joint in a
continuous closed chain model.
Because of the specic methodology used, we were able to calculate the patellar tendon
angle. This is important in analysing the effect of quadriceps loading as anterior or posterior
translation of the tibia is governed by force components in the sagittal plane. The results are
shown in gure 4 and demonstrate how the patellar tendon pulls the tibia forward between
0- 65 exion. In deeper exion, the patellar tendon load will cause the tibia to translate in
a posterior direction. These results closely match the data previously published by Herzog
et al
41
.
101
a.
b.
c.
Fig. 5 Plots of tibial rotation as a function of the exion angle based on previous publications on passive kinemat-
ics.
a. Rotation. Curves were constructed based on data available in the paper. All curves were to pass through 0
at 0 of exion. In case no data was available at 0 exion, the curve was offset to pass through the average
TR value of all other papers at the lowest exion angle where data was available.
b. Medial condyle translation
c. Lateral condyle translation
102
Fig 6. Projection of medial and lateral femoral condyle centres on the horizontal tibial plane for the different load
cases grouped for four selected exion angles
In order to help understand the effects of muscle loading on tibiofemoral kinematics, the re-
sults were reformatted as shown in gure 6. The projections of the medial and lateral femoral
condyle centres on the tibial horizontal plane for the different load cases are superimposed
for four exion angles: 30, 60, 100 and 120. At 30 of exion, the effect of the different
muscle loads is explained as follows:
The quadriceps (Q) induces anterior translation of the tibia. Consequently, the projections
of the femoral condyle centres shift posteriorly on the tibia. Since the medial compart-
ment is inherently more stable than the lateral compartment, translation is less on the
medial side, inducing internal tibial rotation.
Adding the 50N medial hamstrings (Q+mH) load only affects the medial compartment.
The hamstrings action counteracts the translation induced by the quadriceps, but not to its
full extent: medial hamstrings co-contraction of 50N fails to fully compensate the anterior
shear and increases internal rotation initiated by the quadriceps load.
Adding the 50N lateral hamstring load (Q+lH) only affects the lateral compartment. The
biceps action counteracts the translation induced by the quadriceps, but not to its full
103
extent: biceps co-contraction of 50N fails to fully compensate the anterior shear but de-
creases internal tibial rotation.
Adding medial and lateral hamstrings co-contraction (Q+H) translates the tibia poste-
riorly, bringing it close to its relative position in the passive condition: both hamstrings
effectively counteract shear and rotation induced by the quadriceps.
At 60 of exion, the effects are similar to the translations and rotations observed at 30 but
with smaller absolute values.
In deeper exion (100 and 120 exion), all investigated combinations of muscle loads cause
posterior translation of the tibia as compared to passive conditions. Quadriceps loading re-
duces tibial internal rotation and tibial anterior translation (also referred to as femoral roll-
back). Biceps loading enhances this reduction in tibial internal rotation.
These results conrm earlier work by MacWilliams et al and Kwak et al
24,25
: among the ac-
tive exion trials, the action of both hamstrings together limits femoral AP translation and
rotation. As such, the hamstrings action is protective against ACL loading. The greater effect
of the biceps as compared to the medial hamstrings can be explained by the larger effective
moment arm
25
as well by the greater natural laxity of the lateral compartment
42
. We disagree
with Kwak et al when it comes to the clinical consequences of ACL graft choice. The greatest
muscle induced shear force on the ACL occurs in early exion when the quadriceps translates
the tibia forwards and induces tibial internal rotation. Our results clearly demonstrate the
greater effect of the biceps as compared to the medial hamstrings in reducing tibial anterior
translation. In addition, at 30 exion, the medial hamstrings co-contraction increase tibial
internal rotation, in contrast to the biceps which decreases tibial internal rotation. As far as
graft choice goes, there seems to be no biomechanical reason not to select the semitendinosus
or gracilis tendon for ACL reconstruction.
Conclusion
This study describes and compares the inuence of loading and muscle action on knee ki-
nematics. The changes induced are signicant. The widely accepted model of passive knee
kinematics is conrmed but it is clear that this traditional model is not the only valid de-
scription of how the knee moves. Quadriceps loading does provoke an anterior translation
of the tibia up to 60 exion, beyond 60 it is inverted to a posterior translation. Hamstrings
co-contraction reduces rotation and translation of the tibia. The biceps has greater inuence
than the medial hamstrings. These ndings may have consequences for the understanding of
pathogenesis and treatment of ACL ruptures and the design of prostheses.
104
REFERENCES
1. Pinskerova, V, Maquet, P, Freeman, MA. 2000. Writings on the knee between 1836 and 1917. J Bone Joint Surg
Br 82: 1100-1102.
2. Bull, AM, Amis, AA. 1998. Knee joint motion: description and measurement. Proc Inst Mech Eng [H] 212:
357-372.
3. Grood, ES, Suntay, WJ. 1983. A joint coordinate system for the clinical description of three-dimensional mo-
tions: application to the knee. J Biomech Eng 105: 136-144.
4. Kinzel, GL, Gutkowski, LJ. 1983. Joint models, degrees of freedom, and anatomical motion measurement. J
Biomech Eng 105: 55-62.
5. Walker, PS, Haider, H. 2003. Characterizing the motion of total knee replacements in laboratory tests. Clin
Orthop Relat Res: 54-68.
6. Most, E, Axe, J, Rubash, H, Li, G. 2004. Sensitivity of the knee joint kinematics calculation to selection of
exion axes. Journal of Biomechanics 37: 1743-1748.
7. Iwaki, H, Pinskerova, V, Freeman, MAR. 2000. Tibiofemoral movement 1: the shapes and relative movements
of the femur and tibia in the unloaded cadaver knee. J Bone Joint Surg Br 82-B: 1189-1195.
8. Wilson, DR, Feikes, JD, Zavatsky, AB, OConnor, JJ. 2000. The components of passive knee movement are
coupled to exion angle. Journal of Biomechanics 33: 465-473.
9. Grood, ES, Stowers, SF, Noyes, FR. 1988. Limits of movement in the human knee. Effect of sectioning the
posterior cruciate ligament and posterolateral structures. J Bone Joint Surg Am 70: 88-97.
10. Hill, PF, Vedi, V, Williams, A, et al. 2000. Tibiofemoral movement 2: the loaded and unloaded living knee stud-
ied by MRI. J Bone Joint Surg Br 82-B: 1196-1198.
11. Blankevoort, L, Huiskes, R, de Lange, A. 1990. Helical axes of passive knee joint motions. J Biomech 23: 1219-
1229.
12. Bull, AM, Earnshaw, PH, Smith, A, et al. 2002. Intraoperative measurement of knee kinematics in reconstruc-
tion of the anterior cruciate ligament. J Bone Joint Surg Br 84: 1075-1081.
13. Clary, C, Mane, A, Reeve, A, et al. 2007. Knee Kinematics During An Ex vivo Simulated Deep Flexion Squat.
Proceedings of the ASME 2007 Summer Bioengineering Conference.
14. More, RC, Karras, BT, Neiman, R, et al. 1993. Hamstrings--an anterior cruciate ligament protagonist: An ex
vivo study. Am J Sports Med 21: 231-237.
15. Browne, C, Hermida, JC, Bergula, A, et al. 2005. Patellofemoral forces after total knee arthroplasty: effect of
extensor moment arm. The Knee 12: 81-88.
16. Lo, J, Mller, O, Wnschel, M, et al. 2008. Forces in anterior cruciate ligament during simulated weight-bearing
exion with anterior and internal rotational tibial load. Journal of Biomechanics 41: 1855-1861.
17. Coughlin, KM, Incavo, SJ, Churchill, DL, Beynnon, BD. 2003. Tibial axis and patellar position relative to the
femoral epicondylar axis during squatting. The Journal of Arthroplasty 18: 1048-1055.
18. Bull, AM, Kessler, O, Alam, M, Amis, AA. 2008. Changes in knee kinematics reect the articular geometry
after arthroplasty. Clin Orthop Relat Res 466: 2491-2499.
19. Patil, S, Colwell, CW, Jr., Ezzet, KA, DLima, DD. 2005. Can Normal Knee Kinematics Be Restored with
Unicompartmental Knee Replacement? J Bone Joint Surg Am 87: 332-338.
20. Li, G, Zayontz, S, DeFrate, LE, et al. 2004. Kinematics of the knee at high exion angles: an ex vivo investiga-
tion. Journal of Orthopaedic Research 22: 90-95.
21. Churchill, DL, Incavo, SJ, Johnson, CC, Beynnon, BD. 1998. The transepicondylar axis approximates the opti-
mal exion axis of the knee. Clin Orthop Relat Res: 111-118.
22. Baldwin, M, Clary, C, Maletsky, L, Rullkoetter, P. 2008. Specimen-Specic Verication of Predicted TKR
Mechanics during Simulated Deep Flexion Loading Conditions. In, Proceedings of the ASME 2008 Summer
Bioengineering Conference. Marriott Resort, Marco Island, Florida, USA.
23. DLima, DD, Trice, M, Urquhart, AG, Colwell, CW, Jr. 2000. Comparison between the kinematics of xed and
rotating bearing knee prostheses. Clin Orthop Relat Res: 151-157.
24. MacWilliams, BA, Wilson, DR, DesJardins, JD, et al. 1999. Hamstrings cocontraction reduces internal rotation,
anterior translation, and anterior cruciate ligament load in weight-bearing exion. J Orthop Res 17: 817-822.
25. Kwak SD, Ahmad CS, Gardner TR, et al, 2000. Hamstrings and iliotibial band forces affect knee kinematics and
contact pattern. Journal of Orthopaedic Research 18: 101-108.
26. Nakagawa, S, Kadoya, Y, Todo, S, et al. 2000. Tibiofemoral movement 3: full exion in the living knee studied
by MRI. J Bone Joint Surg Br 82: 1199-1200.
105
27. Johal, P, Williams, A, Wragg, P, et al. 2005. Tibio-femoral movement in the living knee. A study of weight
bearing and non-weight bearing knee kinematics using `interventional MRI. Journal of Biomechanics 38: 269-
276.
28. Karrholm, J, Brandsson, S, Freeman, MA. 2000. Tibiofemoral movement 4: changes of axial tibial rotation
caused by forced rotation at the weight-bearing knee studied by RSA. J Bone Joint Surg Br 82: 1201-1203.
29. Banks, SA, Hodge, WA. 1996. Accurate measurement of three-dimensional knee replacement kinematics using
single-plane uoroscopy. IEEE Trans Biomed Eng 43: 638-649.
30. Komistek, RD, Dennis, DA, Mahfouz, M. 2003. In vivo uoroscopic analysis of the normal human knee. Clin
Orthop Relat Res: 69-81.
31. Lu, TW, Tsai, TY, Kuo, MY, et al. 2008. In vivo three-dimensional kinematics of the normal knee during ac-
tive extension under unloaded and loaded conditions using single-plane uoroscopy. Med Eng Phys 30: 1004-
1012.
32. Moro-oka, TA, Hamai, S, Miura, H, et al. 2008. Dynamic activity dependence of in vivo normal knee kinemat-
ics. J Orthop Res 26: 428-434.
33. Victor, J, Van Doninck, D, Labey, L, et al. 2009. How precise can bony landmarks be determined on a CT scan
of the knee? The Knee, in press doi:10.1016/j.knee.2009.01.001
34. Li, G, Rudy, TW, Sakane, M, et al. 1999. The importance of quadriceps and hamstring muscle loading on knee
kinematics and in-situ forces in the ACL. Journal of Biomechanics 32: 395-400.
35. Beynnon, BD, Johnson, RJ, Fleming, BC, et al. 1997. The Strain Behavior of the Anterior Cruciate Ligament
During Squatting and Active Flexion-Extension. The American Journal of Sports Medicine 25: 823-829.
36. Arms, SW, Pope, MH, Johnson, RJ, et al. 1984. The biomechanics of anterior cruciate ligament rehabilitation
and reconstruction. Am J Sports Med 12: 8-18.
37. Hewett, TE, Myer, GD, Ford, KR. 2006. Anterior cruciate ligament injuries in female athletes: Part 1, mecha-
nisms and risk factors. Am J Sports Med 34: 299-311.
38. Wojtys, EM, Huston, LJ, Schock, HJ, et al. 2003. Gender differences in muscular protection of the knee in tor-
sion in size-matched athletes. J Bone Joint Surg Am 85-A: 782-789.
39. Renstrom, P, Arms, SW, Stanwyck, TS, et al. 1986. Strain within the anterior cruciate ligament during hamstring
and quadriceps activity. Am J Sports Med 14: 83-87.
40. Kumagai, M, Mizuno, Y, Mattessich, SM, et al. 2002. Posterior cruciate ligament rupture alters ex vivo knee
kinematics. Clin Orthop Relat Res: 241-248.
41. Herzog, W, Read, LJ. 1993. Lines of action and moment arms of the major force-carrying structures crossing
the human knee joint. J Anat 182 ( Pt 2): 213-230.
42. Van Damme, G, Defoort, K, Ducoulombier, Y, et al. 2005. What Should the Surgeon Aim for When Performing
Computer-Assisted Total Knee Arthroplasty? J Bone Joint Surg Am 87: 52-58.
106
B. THE INFLUENCE OF MUSCLE LOAD ON REPLACED KNEE
KINEMATICS
Introduction
This analysis determined the tibiofemoral kinematics of the replaced knee (bicruciate retain-
ing) on the same specimens as in section A.
Methods
The knee was passively cycled through 5 double exion-extension arcs, by manual control
of the tibia, carefully avoiding rotational constraints. The resulting knee kinematics param-
eters were plotted versus exion angle from 0-130. The kinematics pre- and post-TKA were
compared using general linear model ANOVA for unequal sample sizes, also across exion
angles. This ANOVA was performed because not all the specimens saw the same exion
ranges, due to their individual limits in range of motion. Pair-wise comparisons of data were
made with the Tukey method. For all statistical tests signicance was set at p" 0.05.
Results
Plots versus exion angle of anteroposterior translations (MFT, LFT) and tibial rotation (TR)
comparing the native and replaced knees are shown in gure 1 and 2.
ANOVA found signicant differences between the native and replaced knee for TR (p =
0.001) and MFT (p = 0.008), where the native knee was on average 2 more externally ro-
tated and MFT was 2 mm more posterior. No signicant pair-wise differences were found at
individual exion angles; however, the graphs show trends of greater differences at less than
90 exion.
Discussion
Both the native and replaced knees exhibited the expected screw-home mechanism in this
unloaded and passive motion, as the tibia is more externally rotated near full extension. This
also is shown in the more posterior movement of the lateral femoral condyle compared to
the medial condyle. The native knee showed more slightly more external rotation than the
replaced knee at low exion, due to the medial condyle being more posterior. The differences
between native and replaced knees diminished as the knee went into deeper exion. The
impact of muscle action on knee kinematics was similar as in the native knee. The changes
induced by the implantation of a knee prosthesis are discussed in section D.
Fig 1 Average tibial rotation for different
muscle action conditions, superim-
posed, for the replaced knee.
Tibial rotation vs. flexion angle in the replaced knee
-12
-8
-4
0
4
20 40 60 80 100 120
Flexion angle ()
T
R

(
)
Passive
Loaded, no
hamstrings
Loaded, both
hamstrings
Loaded, med.
hamstrings
Loaded, lat.
hamstrings
107
Fig 2 Replaced knee kinematics. AP translation (MFT = Black, LFT = white) and external tibial rotation versus
knee exion, for different muscle load conditions
108
C. EFFECT OF QUADRICEPS LOAD ON THE KNEE NEAR FULL
EXTENSION
Introduction
When testing cadavers in the knee kinematics simulator, running a dynamic squat near full
extension risks damaging both the specimen and the simulator. This is due to the inherent in-
stability of the system near full extension. At full extension, a change in quadriceps load does
not change the ankle load. The specimen also may experience destructive hyperextension if
the cadaver is forced in the wrong direction. Because of this, a subgroup of six specimens
was tested quasi-statically in the simulator near full extension. With the hip xed in height,
the effect of changing quadriceps load on knee kinematics was measured.
Methods
Specimens 7-12 underwent static trials in the knee rig at angles targeting 0-20 exion. Knee
kinematics were measured at these angles. Each cadaver was tested in different surgical con-
ditions. 43 static trials had both load and motion data and were usable for this study. Load and
motion data were synchronized in time by resampling to 1 Hz and matching peak quadriceps
loads with minimum exion angles. Eleven kinematics and soft-tissue isometry parameters
were calculated.
Linear regressions of each parameter versus quadriceps load were calculated for each static
trial (g. 1). The slopes S of the regressions were considered sensitivity values. Sensitivities
(S-values) were plotted for each parameter versus mean exion angle (g. 2) and then com-
pared using general linear model ANOVA for unbalanced sample sizes (Minitab 15) across
surgical condition, actual exion, and their interaction, with actual exion as a covariate.
Then all the sensitivities across different surgical conditions were pooled and their distribu-
tions plotted (g. 3). Further one-way ANOVA tests were performed comparing variance in
sensitivity across the knee rotations (Flexion angle FA, Tibial rotation TR, Coronal alignment
AlCor), translations (MFT, LFT), and isometry measurements. Signicance was set at p<
0.05, and paired signicance tests were performed using Tukeys method.
Results
ANOVA showed no signicant differences in parameter sensitivity to quadriceps load among
different exion angles or surgical conditions, except for LMPFL, which showed signicant
differences in mean sensitivities among native and replaced knees (-4.39 and -22.1, mm/kN,
respectively; p = 0.024). Because all other parameters showed no signicant relationships
with surgical condition and exion angle, the sensitivity values for each parameter were
aggregated and their distributions analysed (g. 3). Comparison of the three knee rotations
showed signicant differences in sensitivity among FA, TR, and AlCor (p = 0.004). FA was
signicantly more sensitive than TR by 30/kN. AlCor was largely insensitive to QL. Of all
the soft-tissue lengths, LPT was most sensitive to QL as expected. Femorotibial soft-tissues
showed the least variability in sensitivity, while patellar soft-tissues had the largest variabi-
lity.
109
Discussion
The static data largely did not show signicant differences in sensitivity between native and
replaced knee kinematics. However, it did show which parameters in general are more sen-
sitive to QL, such as LPT or FA. The regression equations for all the static data also were
aggregated to create a simple linear model for predicting average knee parameters for the six
cadavers, for native and replaced knees. The results of the model were plotted for FA, TR,
MFT, and LFT for native and replaced knees (g. 4). Note that the native knees show more
sensitivity to target FA, although not necessarily to QL.
Example FA vs. QL graph:
BCR2 native, 0 target flexion
y = -0.016220x - 2.906838
R
2
= 0.815383
-7
-6
-5
-4
-3
-2
-1
0
0 50 100 150 200 250
QL (N)
F
A

(
)
Example sensitivity vs. static FA plot:
Sensitivity of FA to QL
-80
-40
0
40
-10 0 10 20 30 40
Static flexion angle
A
F
A
/
A
Q
L

(
/
k
N
)
Nativ
e Replace
d
Fig 1 Example graph of a knee kinematics param-
eter versus quadriceps load during one static
(or quasi-static) trial for specimen 2. Each data
point represents a 1 s time interval. The slope
of the linear regression approximates sensitiv-
ity to quadriceps load.
Fig 2 Example plot of sensitivities versus exion
angle. Each data point is the sensitivity value
of one static trial, representing how much a
parameter changes with respect to quadriceps
load. The x-axis coordinates of the points are
the midrange exion angles measured in the
static trials.
110

Fig 3 Boxplots of parameter sensitivities to quadriceps load (QL) during static loading, aggregated across all
trials for each parameter, minus outliers. Boxes show the interquartile range (IQR) of measurements, with
whiskers extending to the furthest points away from the boxes still within 1.5x the IQR. Circles indicate
means. Plots are grouped by knee rotations, translations, and soft-tissue isometry. Llatret=length lateral
retinaculum, LLCL= length lateral collateral ligament, LMPFL=length medial patellofemoral ligament,
LPT= length patellar tendon,LSMCLdist/prox=length distal/proximal medial collateral ligament
111
Fig 4 Average kinematics curves versus QL based on static data linear regressions.
Average FA vs. QL
-10
0
10
20
30
0 50 100 150 200
QL (N)
F
A

(
)
Average FA vs. QL
-10
0
10
20
30
0 50 100 150 200
QL (N)
F
A

(
)

Average TR vs. QL
-4
-2
0
2
4
6
0 50 100 150 200
QL (N)
T
R

(
)
Average TR vs. QL
-4
-2
0
2
4
6
0 50 100 150 200
QL (N)
T
R

(
)

Average MFT vs. QL
-10
-5
0
5
10
0 50 100 150 200
QL (N)
M
F
T

(
m
m
)
Average MFT vs. QL
-10
-5
0
5
10
0 50 100 150 200
QL (N)
M
F
T

(
m
m
)

Average LFT vs. QL
-10
-5
0
5
10
0 50 100 150 200
QL (N)
L
F
T

(
m
m
)
Average LFT vs. QL
-10
-5
0
5
10
0 50 100 150 200
QL (N)
L
F
T

(
m
m
)

0
10
20
Target FA
Native
Replaced
P
a
s
s
i
v
e

A
c
t
i
v
e
,

N
o

h
a
m
s
t
r
i
n
g
s

112
D. CHANGES INDUCED BY JOINT REPLACEMENT
Introduction
This analysis reports the kinematics of the twelve cadaver knee specimens before and after
joint replacement during simulated loaded squat, where six received a bicruciate-stabilized
(BCS) prosthesis design, and the other six received a bicruciate-retaining (BCR) design.
Methods
The twelve specimens underwent tests in the knee kinematics simulator as described previ-
ously, where data from 0-30 were from static trials and data from 30-120 were dynamic tri-
als. Static data were obtained at exactly 0, 10, and 20 exion angles for each specimen by
linear interpolation of the existing data collected around these angles. Both hamstrings were
attached during tests. The resulting knee kinematics parameters were plotted versus exion
angle from 0-120. Within each group of BCR and BCS specimens, the kinematics pre- and
post-TKA were compared using general linear model ANOVA for unequal sample sizes, also
across exion angles. This ANOVA was performed because not all the specimens saw the
same exion ranges, due to uncertainty in cadaver xation before testing. Pair-wise compari-
sons of data were made with the Tukey method. An additional comparison was made between
the BCS data versus the passive, native data (see Chapter 3, Section V-A). Due to limitations
in sample sizes, the six native cadavers were not the same six that received the BCS knees.
However, this comparison was made to determine how well the BCS knees reproduced the
average passive kinematics of another unrelated set of native knees.
Results
Plots versus exion angle of anteroposterior femoral translations (MFT, LFT), external tibial
rotation (TR), and quadriceps load (QL) for the native and replaced knees are shown in Fig-
ures 1 and 2.
Native vs. BCR
Figure 1 shows kinematics plots for the BCR specimen group in native and replaced condi-
tions. ANOVA showed that on average the replaced knee TR was 3.4 more internally rotated
(p < 0.001) than the native knee, while the MFT was 3.0 mm more anterior (p < 0.001) than
in the native knee. QL was 48 N larger for the replaced knee (p = 0.037). LFT was not signi-
cantly different between the native and replaced knees. No pair-wise differences between na-
tive and replaced knees at individual exion angles were detected. A signicant relationship
between exion angle and LFT was found among native and replaced knees (p = 0.002) but
not for MFT. From 0-120 exion, the mean ranges of motion for TR, MFT, and LFT were,
respectively, 4.3, 5.0 mm, and 6.9 mm for the native knee; and 4.8, 4.3 mm, and 9.3 mm
for the BCR replaced knee.
113
Native vs. BCS
Figure 2 shows kinematics plots for the BCS specimen group in native and replaced condi-
tions. ANOVA showed that on average the replaced knee TR was 3.6 more internally rotated
(p < 0.001) than the native knee, while the MFT was 2.5 mm more anterior (p < 0.001) than
in the native knee. No signicant differences in QL or LFT were found between the native
and replaced knees. No pair-wise differences between native and replaced knees at individual
exion angles were detected. Both MFT and LFT were found to be related to exion angle,
among the native and replaced knees (p < 0.001). From 0-120 exion, the mean ranges of
motion for TR, MFT, and LFT were, respectively, 4.8, 6.9 mm, and 6.8 mm for the native
knee; and 9.7, 12.9 mm, and 16.3 mm for the BCS replaced knee, being signicantly higher
(p<0.001). The additional comparison between passive, native knee kinematics and active,
BCS-replaced knee kinematics are summarized in Figure 3. Active, replaced knee kinemat-
ics was on average 3.3 more internally rotated than the passive, native knee kinematics (p
< 0.001), and the medial femoral condyle was 2.4 mm more anterior (p = 0.011). The lateral
condyle translation in both cases followed approximately the same absolute values, and no
signicant differences were detected.
Discussion
Knees replaced with BCR produced loaded kinematics curves with shapes similar to the
native knees. However, they also both showed some differences in the absolute kinematics
values compared to the native knees, on the order of 3-4 more tibial internal rotation and 2-3
mm more anterior translation of the medial femoral condyle. Overall the BCR knee kinemat-
ics showed smaller deviations from the native knee kinematics than BCS knee kinematics.
The BCS knee forced more femoral rollback and posterior translation than the native knee,
in loaded conditions. The mechanical substitution with cam and post, of the intricate cruciate
ligament system, does not allow the knee to follow the changes induced by muscle loading
that are seen in the native knee.
Replaced knee kinematics in general was not more variable than native knee kinematics.
In the case of BCR tibial rotation, the kinematics was even less variable than in the native
knee.
Regarding quadriceps loading, the BCR knees showed a 48 N higher average load during the
squat than the native knees, whereas the BCS knees did not show any signicant difference.
This may be explained by the different quadriceps moment arms in the two implants. The
BCS knee showed more femoral rollback, which would lengthen the distance between the
tibial insertions of the quadriceps and the tibiofemoral contact points. Effectively this would
give the quadriceps a larger moment arm with which to pull the knee into extension. Then a
smaller quadriceps force would be needed in a BCS knee compared to the BCR knee, for the
same motions.
114
Fig 1 Native and BCR knee kinematics during loaded squat. AP translation, external tibial rotation, and quad-
riceps load versus knee exion during active exion (loaded squat) for native and BCR replaced knees. n =
6 specimens, although not all specimens saw the same exion ranges. Error bars are 1SD.

R
e
p
l
a
c
e
d
,

B
C
R

-10
-5
0
5
10
-10 40 90 140
Flexion angle ()
T
R

(
)
-15
-10
-5
0
5
10
-10 40 90 140
Flexion angle ()
T
r
a
n
s
l
a
t
i
o
n

(
m
m
)
N
a
t
i
v
e

Quadriceps Load
R
e
p
l
a
c
e
d
,

B
C
R

0
0.4
0.8
1.2
-10 40 90 140
Flexion angle ()
Q
L

(
k
N
)
0
0.4
0.8
1.2
-10 40 90 140
Flexion angle ()
Q
L

(
k
N
)
differences were detected.

.
N
a
t
i
v
e

Femoral AP Translation Tibial Ext. Rotation
-10
-5
0
5
10
-10 40 90 140
Flexion angle ()
T
R

(
)
-15
-10
-5
0
5
10
-10 40 90 140
Flexion angle ()
T
r
a
n
s
l
a
t
i
o
n

(
m
m
)
MFT
LFT
115
Fig 2 Native and BCS knee kinematics during loaded squat. AP translation, external tibial rotation, and quad-
riceps load versus knee exion during active exion (loaded squat) for native and BCS replaced knees. n =
6 specimens, although not all specimens saw the same exion ranges. Error bars are 1SD.

N
a
t
i
v
e

R
e
p
l
a
c
e
d
,

B
C
S

N
a
t
i
v
e

Quadriceps Load
R
e
p
l
a
c
e
d
,

B
C
S

-15
-10
-5
0
5
10
-10 40 90 140
Flexion angle ()
T
R

(
)
-25
-15
-5
5
15
-10 40 90 140
Flexion angle ()
T
r
a
n
s
l
a
t
i
o
n

(
m
m
)
-25
-15
-5
5
15
-10 40 90 140
Flexion angle ()
T
r
a
n
s
l
a
t
i
o
n

(
m
m
)
MFT
LFT
-15
-10
-5
0
5
10
-10 40 90 140
Flexion angle ()
T
R

(
)
0
0.4
0.8
1.2
-10 40 90 140
Flexion angle ()
Q
L

(
k
N
)
0
0.4
0.8
1.2
-10 40 90 140
Flexion angle ()
Q
L

(
k
N
)
116
Fig 3 Passive native knee kinematics vs. Active (loaded squat) BCS-replaced knee kinematics. AP transla-
tion and external tibial rotation versus knee exion during passive exion for native knees and loaded
exion for BCS-replaced knees. Specimen groups had different cadavers, with n = 6 specimens for each
group. Error bars are 1SD.

N
a
t
i
v
e
,

P
a
s
s
i
v
e

R
e
p
l
a
c
e
d
,

B
C
S
,

L
o
a
d
e
d

S
q
u
a
t

-20
-10
0
10
-10 40 90 140
Flexion angle ()
T
R

(
)
-30
-20
-10
0
10
-10 40 90 140
Flexion angle ()
T
r
a
n
s
l
a
t
i
o
n

(
m
m
)
-20
-10
0
10
-10 40 90 140
Flexion angle ()
T
R

(
)
-30
-20
-10
0
10
-10 40 90 140
Flexion angle ()
T
r
a
n
s
l
a
t
i
o
n

(
m
m
)
MFT
LF
T
117
E. CASE STUDY: ONE SPECIMEN WITH BCR FOLLOWED BY BCS
REPLACEMENT
Introduction
One cadaver specimen was tested with two different knee replacement designs for a direct
comparison of kinematics. This is a longer and more complex procedure than analyzing just
one implant, and hence it was only practical for one specimen. The Bi-Cruciate-Stabilized
(BCS) knee and the Bi-Cruciate-Retaining (BCR) knee both were separately implanted into
the same specimen and tested in the knee simulator, using the appropriate surgical tools and
procedures. Although much of the articular geometry of the BCS and BCR designs is the
same, the kinematics of the two knees may differ because of their different ways of stabiliz-
ing the knee. The BCS relies on the post-cam mechanisms of the implant while the BCR
relies on the native anterior and posterior cruciate ligaments. This case study explored the
resulting kinematic differences.
Methods
One specimen was tested in a knee kinematics simulator as previously described, in native
condition. Then the knee was replaced with a BCR knee system and tested. After that the
knee was converted to a BCS knee and tested. The BCR surgical procedure conserved bone
and tissue that later was removed in the BCS procedure, and so in this order, with BCR
rst, testing the two implants did not cause problems.
In each of the three surgical conditions, the specimen performed ve types of motions de-
scribed previously: passive exion, and active exion with four different hamstring load
conditions. The dynamic active exion squats started away from full extension, targeting the
range 30-120, for reasons previously explained. Static data also were obtained from 0-30
only for the case with both hamstrings attached. The knee kinematics parameters under all
these conditions were calculated and plotted.
Results
Selected plots of parameters versus exion angle are shown in Figures 1-3. Figure 1 shows
external tibial rotation (TR) for passive exion and the four hamstring load conditions, for
native and replaced knees. Figure 2 shows the AP translations (MFT, LFT) for only the load-
ed exion condition with both hamstrings attached for the native knee, as this most closely
resembles the in vivo situation. Figure 3 shows the quadriceps load (QL) for the four ham-
string conditions, for native and replaced knees.
Discussion
In this single specimen the BCR knee more closely resembled the native knee kinematics, as
seen in the TR and AP translation data. This may be due to the preservation of the cruciate
ligaments, which may allow more physiologic motion. The BCS and BCR knees also repro-
duce the medial pivot and roll-back intended in their design, though more motion is seen
post-TKA than in the native knee.
118
In terms of quadriceps loading, in this case the BCS knee more closely resembled the load
situation in the native knee, while the BCR knee had a greater load than in the native knee.
This may be related to the kinematics, since the BCS knee exhibits more roll-back and theo-
retically has a greater moment arm for the quadriceps extension force. This can reduce the
quadriceps load necessary in a squat using a BCS knee.
Finally this is only one specimen, so no general conclusions can be made. The phenomena
seen here also may be partly due to non-implant factors. The BCS knee was required to be
tested last, which may bias the data if the specimen changed naturally over several hours of
testing. Furthermore each of the single trials could have been inuenced by unique conditions
in the knee simulator at the time of testing. However, the tests performed here are a proof-
of-concept that two different implants can be tested on the same cadaver. The surgical pro-
cedures and implants must be compatible, going from the most to the least bone-conserving
implant.
119
Fig 1 Tibial rotation vs. exion angle for the native, BCR, and BCS knees, for passive exion and the four active
exion conditions.
TR, Passive
-25
-20
-15
-10
-5
0
5
0 50 100 150
Flexion angle ()
T
i
b
i
a
l

r
o
t
a
t
i
o
n

(
)

TR, Both Hamstrings Attached
-20
-15
-10
-5
0
5
10
0 50 100 150
Flexion angle ()
T
i
b
i
a
l

r
o
t
a
t
i
o
n

(
)
TR, No Hamstrings Attached
-20
-15
-10
-5
0
5
10
0 50 100 150
Flexion angle ()
T
i
b
i
a
l

r
o
t
a
t
i
o
n

(
)

TR, Lat. Hamstrings Attached
-20
-15
-10
-5
0
5
10
0 50 100 150
Flexion angle ()
T
i
b
i
a
l

r
o
t
a
t
i
o
n

(
)
TR, Med. Hamstrings Attached
-20
-15
-10
-5
0
5
10
0 50 100 150
Flexion angle ()
T
i
b
i
a
l

r
o
t
a
t
i
o
n

(
)
Native
BCR
BCS

120
Fig 2 Medial and lateral femoral translations vs. exion angle for the native, BCR, and BCS knees, for
active exion with both hamstrings attached, which most closely resembles in vivo squats.
AP Translations, Native
-30
-20
-10
0
10
0 50 100 150
Flexion angle ()
A
P

t
r
a
n
s
l
a
t
i
o
n

(
m
m
)

AP Translations, BCR
-30
-20
-10
0
10
0 50 100 150
Flexion angle ()
A
P

t
r
a
n
s
l
a
t
i
o
n

(
m
m
)
AP Translations, BCS
-30
-20
-10
0
10
0 50 100 150
Flexion angle ()
A
P

t
r
a
n
s
l
a
t
i
o
n

(
m
m
)
MFT
LFT
121
Fig 3 Quadriceps load vs. exion angle for the native, BCR, and BCS knees, for the four active exion
conditions. Target ankle load = 130 N.

Native
BCR
BCS
122
VI. Ligament Isometry in the Native and in the Replaced Knee
A. THE NATIVE KNEE
J. Victor, P. Wong, E. Witvrouw, J. Vander Sloten, J. Bellemans. How
isometric are the medial patellofemoral, supercial medial collateral
and lateral collateral ligaments of the knee? Am J Sports Med. 2009 in
press DOI: 10.1177/0363546509337407
Introduction
The concept of ligament isometry is based on the assumption that ligament bers do not
change in length when the joint goes through an arc of motion. Ligament isometry is at
the heart of the model that describes normal knee motion and applies directly to surgical
repair of chronic or acute knee ligament injuries. This concept of isometry has been ex-
perimentally tested, mainly in passive conditions, in the absence of external load and muscle
action.
2,3,16,29,31,36,37
Ligament strains have been studied both directly and indirectly. Indirect
techniques measure the distance between the insertion sites of the ligament.
3,18,29,31,36,37
Dis-
tances have been measured with the help of pins,
3,36,37
dial callipers,
31
and electromagnetic
3-dimensional tracking sensor systems.
18,29
Direct techniques use in-site strain measurements
and require the implantation of external devices called strain gauges in the ligaments.
2,16,20,27

Both techniques have signicant limitations. The indirect technique, using insertion sites, re-
quires to a certain degree anatomic dissection and can only be applied in a static environment.
The direct strain measurement requires dissection as well, in addition to the insertion of a for-
eign body. This could potentially change the properties of the ligament. Dynamic measure-
ments are possible, but the adequate xation of the strain gauge with repetitive motion cycles
remains a practical concern. In this study, we propose a novel technique, using computed
axial tomography surface modeling in conjunction with optical tracking of the femur, tibia,
and patella, to examine the isometric properties of the supercial medial collateral (sMCL)
and the lateral collateral (LCL) ligament. In addition, we aimed to determine at which angle
in the exion arc of the knee the medial patellofemoral ligament (MPFL) would be most taut
and thus have the greatest contribution to resisting patellar subluxation.
The aim of this study is 2-fold:
(1) to validate this novel technique by comparing our results with those of a recent quantita-
tive description of ligament insertion-site anatomy
18,19
; and (2) to determine the length of the
sMCL, LCL, and the MPFL in function of the knee exion angle, both during passive mo-
tion as well as in a dynamic setting including physiologic quadriceps and hamstring muscle
loads.
123
Materials and methods
Twelve cadaveric limbs from 8 male and 4 female donors were disarticulated at the level of
the hip and frozen at 20C. The donors were between 78 and 87 years of age when they died.
Passive optical reective markers were rigidly attached to the femur, tibia, and patella. Volu-
metric CT scans on a 64-row multidetector CT scanner (Lightspeed VCT, GE Healthcare,
Milwaukee, Wisconsin) were performed. All CT data were loaded in a 3-dimensional visual-
ization software system (Mimics 11.02 and its MedCAD module, Materialise, Haasrode, Bel-
gium) for further analysis. After a bone surface reconstruction mask was created, all relevant
surface landmarks were identied and the assumed centers of the ligament insertions on the
femur and the tibia were marked using the quantitative morphologic description by LaPrade
et al (g. 1a).
18,19
An error analysis was done on the rst 6 specimens to evaluate the reliability
of localizing these landmarks on a reconstructed CT scan. Three observers used bone surface
reconstructions of the scanned joints to identify the landmarks 3 times with a minimum inter-
val of 1 week (for intraobserver precision) and independently (for interobserver precision).
The mean error and the maximum error for each landmark were computed along the coronal,
sagittal, and horizontal axes, based on the recorded differences in position for each landmark
between repeated measurements and between observers. Methodologic analysis was carried
out as described by Bland and Altman.
4
a. b.
Fig 1 a. reconstructed surface model showing the medial side of the knee with markings on the medial epicon-
dyle, the supercial medial collateral ligament (sMCL) insertion on the femur, and the medial patellofemo-
ral ligament (MPFL) insertion on the femur and the patella.
b. reconstructed surface model showing the fan shape of the MPFL. The gray dot on the femur represents
the narrow femoral insertion of the MPFL, distal and anterior to the adductor tubercle. On the patella, the
blue dot is marked at 20% of the patellar length from the proximal pole and is found at the superomedial
corner of the patella. The red dot is the central insertion at 40% and the green dot is the caudal insertion of
the MPFL at 60% from the proximal pole of the patella.
The distance between the insertion points on the femur and tibia (sMCL), femur and patella
(MPFL), and femur and bula (LCL) was used to deduce the length of the ligament at any
given position in the exion arc of the knee joint, assuming the bers connect to the insertion
124
sites in a straight line. The femoral attachment of the MPFL was marked in a bony depres-
sion proximal/posterior to the medial epicondyle and anterior/distal to the adductor tubercle.
The femoral insertion of the sMCL was marked in a depression, just posterior/proximal to
the medial epicondyle (g. 1a). On the medial side of the patella, a bony depression with
longitudinal form was observed. As described by LaPrade et al,
18
the midpoint of the MPFL
insertion is located 41% of the length from the proximal tip of the patella along the total
patellar length (proximal to distal). We marked this point as the central anatomic insertion
of the MPFL. In addition, we marked 3 points along the broad and fan-shaped patellar inser-
tion of the MPFL,
1,18,28,31,32
at 20%, 40%, and 60% of the patellar length from the proximal
tip (g. 1b). The corresponding structures of the MPFL are referred to as the cranial, central,
and caudal parts of the MPFL. The tibial insertion of the sMCL could not be located directly
on the CT scan. The proximal attachment is primarily to soft tissues (anterior arm of the
semimembranosus tendon) and the distal attachment is broadly based and fans out over the
medial side of the tibia. For that reason, the sMCL was dissected at the end of the experiment
and the center of the proximal and distal attachment was marked with an aluminum pin. The
proximal and distal sMCL insertions were dened as the structures connecting the femoral
insertion to the proximal and distal tibial insertion, respectively.
18
The LCL attachment on the
femur was marked on the CT scan, slightly proximal and posterior to the lateral epicondyle.
The bular attachment was marked on the anterior margin of the bular head, distal to the
tip of the bular styloid process.
19
The centers of the optical reective markers were marked
by tting a sphere on the image and taking the coordinates of the centers of the spheres. Be-
fore the experiment, the specimens were thawed over 24 hours. The hip was amputated 32
cm cranial to the knee joint line, and the foot was amputated 28 cm caudal to the knee joint
line. The femur and tibia were rigidly xed with polymethylmethacrylate in containers. The
quadriceps tendon was dissected and looped around a metal bar, 7 cm proximal from its at-
tachment to the patella and securely xed with Ticron No. 5 sutures (Covidien, Manseld,
Massachusetts) and Mersilene tape (Ethicon, Johnson & Johnson, Somerville, New Jersey).
The biceps tendon was dissected and attached to a Ticron No. 5 suture with a whipstitch. In
a similar fashion, the semimembranosus and semitendinosus tendons were prepared. The
construct was mounted on a dynamic knee simulator system, based on the Oxford rig, which
was customized for this study. This electromechanical system was designed to simulate and
record the motions and loads in a knee joint during squatting.
Testing was performed at constant speed with an ankle load of 130 N, from full extension to
120. The ankle load was set at this value, as 130 N was the maximum load for consistent
testing of the specimens. Higher ankle loads required quadriceps loads that exceeded the
strength of the extensor mechanism and threatened integrity of the specimens. Five previ-
ously calibrated infrared cameras (Vicon Motion Systems, Los Angeles, California) recorded
the motion of the femur, tibia, and patella. As the optical reective markers were rigidly
attached to the femur and the tibia (the bula was considered a solid body with the tibia),
and the CT scan documented their spatial relation to the marked surface points, the relative
position of all points on the femur, tibia and patella could be computed. In a subset of 6
specimens, additional passive recordings were made, with the investigator cycling each knee
through a full range of motion 5 times. Finally, the specimens were scanned again to check
rigidity and unchanged position of the optical markers relative to a preset reference point on
the femur and the tibia. The aluminium markers were detected, and their coordinates were fed
into the Mimics 3-dimensional software for additional determination of the respective tibial
ligament attachments.
125
Statistical analysis was performed with Statistical Package for the Social Sciences (version
16.0; SPSS Inc., Chicago,IL). The data were assessed for normality using the Kolmogorov-
Smirnov test. A two-way analysis of variance with repeated measures design was conducted
to investigate the effects of load (passive vs. loaded) and knee exion angle (the different
knee exion angles) and their interactive effect on the ligament length. Post-hoc comparison
tests with Bonferroni adjustments were used to analyse signicant main effects or interac-
tions. Multiple regression analysis was performed to investigate the inuence of the knee
exion angle and the load on the length change of the patellar tendon. Statistical signicance
for all tests was accepted at the 5% level.
The intra and inter-observer variability was assessed using the methodology of Bland and
Altman
4
. For each of the considered landmarks positions, we evaluated the intra-class cor-
relation coefcient (ICC) for multiple measurements by different observers on different spec-
imens. By denition, the ICC is evaluated according to the following formulation: .
2
2
b
ICC

.
Where the total variance of measurements by different observers is
2
on different subjects,
and the variance between subjects is
2
b
. ICC values range from 0 to 1, indicating better

agreement as the value approaches 1. An ICC value higher than 0.75 indicates excellent
agreement. The statistical analysis was performed using Matlab R2008a (The MathWorks,
Natick, Massachusetts, USA). For all recorded distances and angles, mean values, maximum
values and standard deviations are reported.
Results
The intra- and interobserver errors in the determination of the center of the ligament insertion
sites are shown in Table 1 and the ICCs in Table 2. Insertion site distances in function of the
exion angle, for the proximal and distal sMCL insertions, LCL, and MPFL in the loaded
setting are shown in Figure 2. The mean lengths in extension for the given ligaments were
50.0 mm for the proximal sMCL insertion (range, 39.7-69.1 mm), 104.6 for the distal sMCL
insertion (range, 96.7 mm-115.7 mm), 65.3 mm for the LCL (range, 53.8 mm- 72.8 mm), and
55.5 mm for the MPFL (range, 45.4- 63.5 mm). The sMCL was found to be near isometric
over the full exion arc, with changes in length of less than 2%. Between 0 and 90 of knee
exion, the maximum change in length for the proximal sMCL insertion was 1 mm, due to
some discrete slackening of the sMCL in the midexion range (30-50 of knee exion). For
the distal sMCL insertion, the pattern is similar, showing approximation of 1.5 mm between
the femoral insertion site and the distal tibial insertion site. From 90 to 120 of knee exion,
both the proximal and distal sMCL insertions slacken, due to 1.9-mm approximation of the
respective femorotibial insertion sites. These length changes were not signicant between the
different exion angles for the proximal and distal sMCL insertions(P>.05).
126
Intra-observer error
Mean Stdev Max
AP ML PD AP ML PD AP ML PD
ins LCL F 0,5 0,3 0,7 0,4 0,3 1,2 1,4 1,4 4,8
ins LCL 0,5 0,2 0,4 0,6 0,2 0,3 2,8 0,8 1,2
ins MPFL F 0,9 0,4 1,1 0,9 0,7 1,5 3,6 3,9 7,8
ins sMCL F 0,4 0,2 0,8 0,3 0,1 0,6 1,6 0,6 2,4
ins MPFL P 0,7 0,2 0,4 0,5 0,2 0,4 1,6 0,6 1,0
Inter-observer error
Mean Stdev Max
AP ML PD AP ML PD AP ML PD
ins LCL F 2,0 0,7 3,4 1,5 0,6 1,8 6,3 2,0 8,1
ins LCL 0,7 0,3 0,8 0,8 0,2 0,6 3,2 0,9 2,0
ins MPFL F 3,5 0,8 2,3 2,0 0,6 1,6 7,6 2,2 6,1
ins sMCL F 0,8 0,2 1,0 0,8 0,1 0,6 2,5 0,6 2,0
ins MPFL P 1,0 0,4 0,7 0,1 0,4 0,1 1,1 0,7 0,8
Table 1 Intra- and inter-observer error in the determination of the centre of the ligament insertion sites. The errors
are displayed along the three axes; AP: antero-posterior, ML: medio-lateral, PD:proximal-distal.
Point ICC-Intra-ObserverValues ICC-Inter-ObserverValues
ins LCL Femur 0.806 0.5526
ins LCL Fibula 0.9809 0.9398
ins MPFL Femur 0.908 0.7878
ins sMCL Femur 0.9563 0.9541
ins MPFL Patella 0.9444 0.8662
Table 2 ICC values for the different landmarks. ICC, intraclass correlation coefcient; Ins, insertion; LCL, lateral
collateral ligament; MPFL, medial patellofemoral ligament; sMCL, supercial medial collateral ligament.
127
Fig 2 Graph of the proximal supercial medial collateral ligament insertion (sMCLprox), distal supercial me-
dial collateral ligament insertion (sMCLdist), lateral collateral ligament (LCL), and medial patellofemoral
ligament (MPFL) with ligament length plotted against the knee exion angle. The sMCLprox and the
sMCLdist are isometric over the full exion arc with less than 2% change in length. The LCL is isometric
between 0 and 70 of knee exion; beyond 70 of knee exion there is linear shortening of 1.2 mm/10
of knee exion (P > .05). The MPFL is isometric between 0 and 40 of knee exion; beyond 40 of knee
exion, there is signicant linear shortening of 0.5 mm/10 of knee exion (P < .05).N = 12 for sMCL and
LCL; N = 11 for MPFL.
Fig 3 Distinct length changes of the cranial (MPFLp), central (MPFLc), and caudal (MPFLd) parts of the medial
patellofemoral ligament, corresponding to the patellar insertion points as shown in Figure 1B. N = 11.
The mean length of the LCL in extension was 65.3 mm (range, 53.8-72.8 mm). The LCL re-
mains almost isometric between 0 and 70 of knee exion, with insertion site approximation
of less than 2% (1.1 mm). From 70 to 120 of knee exion, further slackening of the LCL
occurs due to a trend of femorobular insertion site approximation of 1.2 mm/10 of exion,
without being signicant (P > .05).
As noted in the legend for Figure 3, N = 11 for the MPFL data, due to the fact that readings
of the patellar motion were lost in 1 specimen. Loosening of the optical reference frame on
128
the patella was the cause. The mean length of the MPFL (anatomic center of the ligament,
according to LaPrade et al
18
) in extension was 55 mm (range, 45.4-63.5 mm). The MPFL
remains almost isometric between 0 and 40 of knee exion; the femoropatellar insertion
sites approximate with only 1 mm. Between 0 to 90 of knee exion, there is a signicant
approximation of the insertion sites of 2.6 mm (P = .034). Between 90 and 120 of knee
exion, the insertion site approximation is 2.4 mm (P = .022). As shown in the graph, the
insertion site approximation is linear from 40 to 120 of knee exion, the total change being
4 mm (0.5 mm/10 knee exion, P = .017).
In addition, the change in length of the proximal, central, and distal parts of the MPFL were
compared. According to the marking of the patellar insertion points at 20%, 40%, and 60%
from the proximal pole of the patella (g. 1b), the length changes in the cranial, central, and
caudal bundles are displayed in Figure 3. In contrast to the initial (between full extension
and 40 of exion) isometric behavior of the central part of the MPFL, the most cranial and
caudal parts of the MPFL behave differently.
The cranial part is most taut at full extension. In early exion, slackening occurs due to
approximation of the insertion sites of 2 mm. This change in distance between extension
and 20 of exion is signicant (P = .007). Beyond 30 of exion, the pattern of the cranial
bundle is very similar to the central bundle, with gradual slackening due to linear approxima-
tion of the insertion sites (0.6 mm/10 knee exion). The most caudal part of the ligament is
most taut at 30 of exion. It starts from a more slack position in full extension and gradually
tightens with increasing exion due to distraction of the insertion sites. This change in length
is signicant between extension and 30 of exion (P = .005). Past 30 of exion, the caudal
part of the MPFL gradually slackens at a slower pace than the proximal and central part, with
approximation of its insertion sites of 0.25 mm/10 of knee exion.
We noted a difference between the active and the passive measurements. For the active squat,
the ankle load was kept at 130 N (5 N), with increasing quadriceps loading as the knee ex-
es. The evolution of the quadriceps load is shown in Figure 4a. The highest load was reached
at 110 of exion and approached 1200 N. With increasing quadriceps load from 263 N to
1149 N, we found the patellar tendon to lengthen from 57.7 to 63.6 mm (P = .001 for load
and P = .013 for knee exion angle) (g. 4b). On average, the distance between the femoro-
tibial insertion sites of the distal sMCL was 2.1 mm (range, 1.5 to 2.9 mm) shorter in the
loaded squat than during the passive motion arc (P = .035) (g. 5a). The distance between the
femorobular insertion sites of the LCL decreased with loading as compared with the passive
measurements from 1.6 mm in extension to 4.9 mm at 120 of exion (P = .007) (g. 5b).
For the MPFL length, no signicant difference between the passive and the loaded squat was
found (mean, 0.1 mm; range, 1.3 to 0.7; P > .05) (g. 5c).
129
a. b.
Fig 4 a. the quadriceps load in Newtons plotted against knee exion angle in degrees. The load gradu-
ally increases up to almost 1200 N. b. the gradual increase in length of the patellar tendon (PT),
following the quadriceps load. N = 11.
a. b.
c.
Fig 5 Active versus passive length changes in function of knee exion angle for the distal supercial medial
collateral ligament (sMCLdist), lateral collateral ligament (LCL), and medial patellofemoral ligament
(MPFL). a. and b. the sMCL and LCL are signicantly more relaxed during the loaded squat than during
the passive motion arc, due to compression of the articular cartilage. c. this difference between active and
passive is not found for the MPFL due to the stretch effect on the patellar tendon during quadriceps loading.
N = 6.
130
Discussion
Functional insight into the 3-dimensional anatomy of knee ligaments is crucial in the di-
agnosis and treatment of patients with ligament injuries. Several authors reported on the
anatomic substrate of injuries to the medial and lateral sides of the knee. Detailed descrip-
tions are available for the medial tibiofemoral structures,
5,18,36,37
the lateral corner,
19
and the
MPFL.
1,18,23,24,27
Recently, the anatomy of the medial and lateral part of the knee has been described quan-
titatively.
18,19
The main addition to previous anatomic work lies in the exact description of
the ligament attachment sites in relation to certain anatomic landmarks that can be palpated
during surgery. To our knowledge, no validation of these important anatomic studies has been
reported so far. As the insertion sites on the femur, tibia, bula, and patella will determine lig-
ament strain over the exion arc, we can use ligament strain to validate an anatomic model.
In other words, our marking of ligament insertion sites should conrm previous experimental
work on isometry and strain of the ligaments. Assuming the quantitative anatomic descrip-
tion by LaPrade et al is correct, our experiment should conrm the near-isometric nature of
the medial collateral ligament,
3,5,36
and slackening of the LCL and MPFL with increasing ex-
ion.
1,24,31-34
The methods we used have the advantage of linking detailed anatomic information
to a kinematic analysis. However, the following weaknesses have to be understood. The set-
up of a cadaveric limb in a knee rig is nonphysiologic. The quadriceps muscles are treated as
a single unit and the hamstrings muscle load of 50 N medially and laterally was chosen based
on previous biomechanical cadaver studies,
7,17,21,22
but this has not been validated in vivo.
Also, 2 different methods of marking ligament insertions have been used: anatomic dissec-
tion for the tibia and CT scan for the femur and patella.
Warren et al
36,37
described the sMCL as the primary stabilizer of the medial side of the knee
against valgus stress. They divided the ligament into 2 areas, the anterior and posterior parts,
for studying strain. As compared with full extension, they found discrete lengthening of the
most anterior part of the ligament at 90 (1 mm) of exion versus discrete shortening (2 mm)
of the most posterior part of the ligament. They explained this phenomenon in describing the
location of the most anterior part of the femoral insertion anterior to the instant center of rota-
tion and the most posterior insertion of the ligament posterior to the instant center of rotation.
In half the specimens, the distance between the ligament insertion sites was at its longest at
90 of exion. Bartel et al, working on a small sample of 3 cadaveric specimens, reported
similar ndings. They found a change in length of 3% to 4% in the exion arc between 0 and
120. Our results are close to those earlier articles and conrm near isometry of the sMCL
as the changes in length stay below 2 mm for the 0 to 120 exion arc. The reason why we
found slightly less strain than did Warren and Bartel can be explained by the location of the
femoral insertion site in the center of the ligament attachment of the sMCL. We did not divide
the ligament into an anterior or posterior part but selected the most central bers. It is logi-
cal that those bers have the least strain during the motion arc of the knee.
2
In contrast to the
concept of the stable and isometric medial side of the knee, the lateral side displays greater
mobility,
14
and the stabilizing ligaments tend to become slacker with increasing exion.
33,35

Van Damme et al
34
reported 5.9 mm of lateral joint opening with a moment load of 9.8 Nm
at 30 of exion versus 8.1 mm at 90 of exion. Our results indicate that the LCL remains
131
isometric only between 0 to 70 of knee exion. Beyond, a trend toward approximation of
the insertion site distance is observed, without being signicant.
A signicant difference between the insertion site distance in the loaded versus the passive
setting, for both the sMCL and LCL, was observed in our study. This can be partially ex-
plained by the compression of the articular cartilage in the loaded setting. The observation
that this difference was greater on the lateral than on the medial side can be due to the greater
natural laxity on the lateral side of the knee joint.
33,34
Where some distraction could occur dur-
ing passive mobilization of the knee, the articular surfaces are pushed together in the loaded
setting, mitigating the laxity effect. Reduced femoral roll-back on the lateral side, caused
by the 50-N hamstrings force, could also contribute to the reported difference between the
loaded and the passive experiment.
17

These results clearly show that some ligament relaxation must occur as soon as the knee goes
from unloaded modus with relaxed quadriceps to the loaded state with active quadriceps. In
the passive, relaxed setting, the ligaments provide stability to the joint. As soon as the joint
is loaded, some relaxation of the ligament occurs and stability must be provided by joint
congruity and muscle action.
The observation that this difference was not found for the MPFL can be explained by the fact
that the loading of the quadriceps lengthens the patellar tendon, putting the patella in a more
cranial position with increasing exion (g. 4). This patellar tendon strain of 10% may seem
surprisingly high. A potential error that could lead to this nding is the initial position of the
patella with the knee in full extension. With no load on the quadriceps and an amputated
specimen, one could simply measure the effect of resting length convolution. Therefore, the
reported length measurements start at 10 of exion when there is already initial load on the
quadriceps system. The magnitude of strain we reported seems to fall in line with other in
vivo studies and falls below the strains found at failure in other in vitro studies.
Hansen et al studied a 10-s ramp isometric knee extension in 12 male subjects and showed
7% strain of the patellar tendon (with 3.7% between-day measurement error), going from no
quadriceps contraction to maximal voluntary quadriceps contraction. They dened tendon
length as we did, but used ultrasound for measurement. In a similar in vivo study on maximal
total strain of the patellar tendon during ramped isometric knee extension, Onambele et al
25

found values of 15.4% in women and 16.5% in men. In vitro studies have determined the
strain of the patellar tendon at failure. Johnson et al,
15
using fresh-frozen patellar tendons,
found the strain to be 15% at failure. In a later study, Hashemi et al
12
reported a strain at
failure of 18%.
This results in a relative lengthening of the MPFL, compensating for the compression of the
articular cartilage. Our strain data of the anatomic central part of the MPFL are similar to
those reported by Steensen et al.
31
The authors performed a cadaveric dissection study on 11
knees (of which 8 were paired), measuring the distance between the inferior patellar attach-
ments to the superior femoral attachment. They described the MPFL as being 17 mm wide at
its patellar insertion and 15.4 mm at its femoral insertion, and divided the insertion area into 3
parts. The best isometry was found between the inferior patellar attachment and the superior
femoral attachment. These described insertion sites coincide on the femur with the deni-
tion of the MPFL insertion given by LaPrade et al.
18
On the patella, the inferior insertion
used by Steensen et al was at 50% of the patellar length. From 0 to 90 of knee exion, the
132
reported average change in length by Steensen et al was 1.1 mm and from 0 to 120 of knee
exion, 2.9 mm. Our values showed a similar trend of shortening with increasing exion,
respectively 2.6 mm and 2.4 mm. The slightly higher value from 0 to 90 of exion can be
due to the different methodology. Steensen et al used initial manual loading and then passive
natural tension on the extensor mechanism in intact cadavers, whereas we used controlled
quadriceps loading on amputated specimens in this study.
Our ndings on the MPFL have clinical consequences. Instability of the patella is a multi-
factorial problem as patellar stability relies on limb alignment, surface geometry of the distal
femur and patella, integrity of the soft tissue constraints, and the interplay of the surround-
ing muscles. In the initial phase of knee exion, between full extension and 30 of exion,
the patella is not yet fully engaged in the trochlea and the quadriceps force is relatively low.
Stability is mainly provided by muscle action and soft tissue restraints. The MPFL is the
primary passive soft tissue restraint to lateral patellar displacement from 0 to 30 of knee
exion.
1,8,13,23
In a cadaver study using sequential cutting of the medial structures, Desio et al
8
found the MPFL to provide 60% of lateral restraint. This was conrmed by Hautamaa et al
13

and later by Nomura et al,
24
who found a signicant increase in lateral patellar displacement
at 30 of knee exion after transecting the MPFL. Amis et al
1
found the MPFL to contribute
most to resisting lateral patellar translation at full extension. They did not measure MPFL
strain but reported subjective observations of the natural MPFL being tight in the extended
knee and slackening as the knee exes. In the same study, the in vitro tests showed that the
patella subluxated most easily at 20 of knee exion. Earlier biomechanical studies support
this nding. As the knee exes beyond 20, the patellofemoral joint reaction force rapidly
increases,
26
pushing the patella into the conforming geometry of the trochlea.
Surgical techniques to reconstruct the MPFL vary signicantly. In a recent systematic review
of the literature, Smith et al
30
could only nd 8 clinical papers reporting techniques and
outcomes with an appropriate methodology. Despite the low number of recurring disloca-
tions (6 of 155 cases), the authors did not draw rm conclusions from those papers because
of methodologic weaknesses. Three later papers
6,28,32
described an anatomic reconstruction
using the gracilis tendon with dual anchoring in the patella, mimicking the native fan shape
of the MPFL with a narrow femoral insertion and a wider patellar insertion. Schottle et al
28

reported a higher stability during exion and decreased patellar rotation in comparison with
a single point xation. The authors used the superomedial corner and the midpoint of the me-
dial margin as insertion sites for the gracilis graft on the patella. Tensioning was performed at
the femoral xation site, with the knee positioned in 30 of exion. Christiansen et al
6
used 2
anchor points in the proximal two thirds of the medial border of the patella, spaced 10 to 15
mm apart, and the medial epicondyle as the femoral anchor point for xation of the gracilis
graft. Fixation was carried out in 45 of exion, still allowing 10 mm of manual lateralization
of the patella. Thaunat and Erasmus
32
used the same femoral insertion point, but a different
patellar insertion. The 2 ends of the gracilis graft were attached on the proximal third of the
medial patellar border, spaced 10 mm apart. The graft was tensioned in full extension, as the
authors observed a favourable anisometry with progressive slackening of the graft as the
knee exed. This is an important concept as overconstraining can lead to medial subluxation,
graft failure, or medial cartilage overload.
9
As the native MPFL is more elastic than the ham-
strings tendons,
10
errors in graft placement or knee positioning during graft tensioning and
xation are certainly to be avoided. On the basis of our ndings, we support the observation
133
by Thaunat and Erasmus,
32
as well as the earlier descriptions by Nomura et al
24
and Amis et
al.
1
With a cranial patellar anchor point (often called the superomedial corner of the patella),
the graft will be most taut in full extension and will gradually slacken during exion. This is
clearly shown in Figure 3 (cranial bundle). When the most caudal (often called distal) part
of the MPFL is considered, initial tightening will occur from full extension to 30 of exion.
From these data, it can be concluded that a double-bundle reconstruction of the MPFL can
benet from differential tensioning. To avoid overconstraining, the cranial (or proximal) bun-
dle should be tightened in full extension, versus the caudal (or distal) bundle to be tightened
at 30 of exion. If only 1 bundle is reconstructed, the least strain is to be expected from the
caudal insertion of the patella (60% point, Figure 1B) to the anatomic insertion on the femur
as described by LaPrade et al.
18

Conclusion
We validated a novel experimental technique for the study of dynamic ligament strains. The
model demonstrates that it is possible to dene the femoral, bular, and patellar insertion
sites of the sMCL, the LCL, and the MPFL on a CT scan. Our experimental model demon-
strated near-isometric behavior of the sMCL. The LCL is isometric between 0 and 70 of
exion, followed by a nonsignicant trend toward slackening with deeper exion. The MPFL
is most taut in extension and gradually starts to slacken in a linear fashion from 40 of knee
exion onward to 120 of exion. In considering the broad patellar insertion of the MPFL,
the cranial part is most tight in full extension and the caudal part is most tight at 30 of ex-
ion. Based on these ndings, differential tensioning of the graft bundles in double-bundle
MPFL reconstruction is advocated.
REFERENCES
1. Amis AA, Firer P, Mountney J, Senavongse W, Thomas NP. Anatomy and biomechanics of the
medial patellofemoral ligament. Knee. 2003;10:215-220.
2. Arms S, Boyle J, Johnson R, Pope M. Strain measurement in the medial collateral ligament of the human knee:
an autopsy study. J Biomech. 1983;16:491-496.
3. Bartel DL, Marshall JL, Schieck RA, Wang JB. Surgical repositioning of the medial collateral ligament. J Bone
Joint Surg Am. 1977;59:107-116.
ment. Lancet. 1986;1:307-310.
5. Borden PS, Kantaras AT, Caborn DNM. Medial collateral ligament reconstruction with allograft using a double-
bundle technique. Arthroscopy. 2002;18:E19.
6. Christiansen SE, Jacobsen BW, Lund B, Lind M. Reconstruction of the medial patellofemoral ligament with
gracilis tendon autograft in transverse patellar drill holes. Arthroscopy. 2008;24:82-87.
7. Churchill DL, Incavo SJ, Johnson CC, Beynnon BD. The transepicondylar axis approximates the optimal ex-
ion axis of the knee. Clin Orthop Relat Res. 1998;356:111-118.
8. Desio SM, Burks RT, Bachus KN. Soft tissue restraints to lateral patellar translation in the human knee. Am J
Sports Med. 1998; 26:59-65.
9. Elias JJ, Cosgarea AJ. Technical errors during medial patellofemoral ligament reconstruction could overload
medial patellofemoral cartilage: a computational analysis. Am J Sports Med. 2006;34:1478-1485.
10. Feller JA, Amis AA, Andrish JT, Arendt EA, Erasmus PJ, Powers CM. Surgical biomechanics of the patel-
lofemoral joint. Arthroscopy. 2007;23:542-553.
134
11. Hansen P, Bojsen-Moller J, Aagaard P, Kjaer M, Magnusson SP. Mechanical properties of the human patellar
tendon, in vivo. Clin Biomech (Bristol, Avon). 2006;21(1):54-58.
12. Hashemi J, Chandrashekar N, Slauterbeck J. The mechanical properties of the human patellar tendon are cor-
related to its mass density and are independent of sex. Clin Biomech (Bristol, Avon). 2005;20(6):645-652.
13. Hautamaa PV, Fithian DC, Kaufman KR, Daniel DM, Pohlmeyer AM. Medial soft tissue restraints in lateral
patellar instability and repair. Clin Orthop Relat Res. 1998;349:174-182.
14. Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement, 1: the shapes and relative movements of the
femur and the tibia in unloaded cadaver knee. J Bone Joint Surg Br. 2000;82:1189-1195.
15. Johnson GA, Tramaglini DM, Levine RE, Ohno K, Choi NY, Woo SL. Tensile and viscoelastic properties of
human patellar tendon. J Orthop Res. 1994;12(6):796-803.
16. Kennedy JC, Hawkins RJ, Willis RB. Strain gauge analysis of knee ligaments. Clin Orthop Relat Res.
1977;129:225-229.
17. Kwak SD, Ahmad CS, Gardner TR, et al. Hamstrings and iliotibial band forces affect knee kinematics and
contact pattern. J Orthop Res. 2000;18:101-108.
18. LaPrade RF, Engebretsen AH, Ly TV, Johansen S, Wentorf FA, Engebretsen L. The anatomy of the medial part
of the knee. J Bone Joint Surg Am. 2007;89:2000-2010.
19. LaPrade RF, Ly TV, Wentorf FA, Engebretsen L. The posterolateral attachments of the knee: a qualitative and
quantitative morphologic analysis of the bular collateral ligament, popliteus tendon, popliteobular ligament,
and lateral gastrocnemius tendon. Am J Sports Med. 2003;31:854-860.
20. Lew WD, Lewis JL. The effect of knee-prosthesis geometry on cruciate ligament mechanics during exion. J
Bone Joint Surg Am. 1982;64:734-739.
21. Li G, Rudi TW, Sakane M, Kanamori A, Ma CB, Woo SLY. The importance of quadriceps and hamstring muscle
loading on knee kinematics and in-situ forces in the ACL. J Biomech. 1999;32:395-400.
22. More RC, Karras BT, Neiman R, Fritschy D, Woo SLY, Daniel D. Hamstrings: an anterior cruciate ligament
protagonist. Am J Sports Med. 1993;21:231-237.
23. Nomura E, Horiuchi Y, Kihara M. Medial patellofemoral ligament restraint in lateral patellar translation and
reconstruction. Knee. 2000;7:121-127.
24. Nomura E, Inoue M, Osada N. Anatomical analysis of the medial patellofemoral ligament of the knee, espe-
cially the femoral attachment. Knee Surg Sports Traumatol Arthrosc. 2005;13:510-515.
25. Onambele GN, Burgess K, Pearson SJ. Gender-specic in vivo measurement of the structural and mechanical
properties of the human patellar tendon. J Orthop Res. 2007;25(12):1635-1642.
26. Reilly DT, Martens M. Experimental analysis of the quadriceps muscle force and patello-femoral joint reaction
force for various activities. Acta Orthop Scand. 1972;43:126-137.
27. Renstrm P, Arms SW, Stanwyck TS, Johnson RJ, Pope MH. Strain within the anterior cruciate ligament during
hamstring and quadriceps activity. Am J Sports Med. 1986;14:83-87.
28. Schottle PB, Romero J, Schmeling A, Weiler A. Technical note: anatomical reconstruction of the medial patel-
lofemoral ligament using a free gracilis autograft. Arch Orthop Trauma Surg. 2008;128: 479-484.
29. Sidles JA, Larson RV, Garbini JL, Downey DJ, Matsen FA 3rd. Ligament length relationships in the moving
knee. J Orthop Res. 1988;6:593-610.
30. Smith TO, Walker J, Russell N. Outcomes of medial patellofemoral ligament reconstruction for patellar instabil-
ity: a systemic review. Knee Surg Sports Traumatol Arthrosc. 2007;15:1301-1314.
31. Steensen RN, Dopirak RM, McDonald WG. The anatomy and isometry of the medial patellofemoral ligament:
implications for reconstruction. Am J Sports Med. 2004;32:1509-1513.
32. Thaunat M, Erasmus PJ. The favourable ansiometry: an original concept for medial patellofemoral ligament
reconstruction. Knee. 2007;14:424-428.
33. Tokuhara Y, Kadoya Y, Nakagawa S, Kobayashi A, Takaoka K. The exion gap in normal knees: an MRI study.
J Bone Joint Surg Br. 2004;86:1133-1136.
34. Van Damme G, Defoort K, Ducoulombier Y, Van Glabbeek F, Bellemans J, Victor J. What should the surgeon
aim for when performing computer assisted total knee arthroplasty? J Bone Joint Surg Am. 2005;87(suppl
2):52-58.
35. Wang CJ, Walker PS. The effects of exion and rotation on the length patterns of the ligaments of the knee. J
Biomech. 1973;6: 587-596.
36. Warren LF, Marshall JL, Girgis F. The prime static stabilizer of the medial side of the knee. J Bone Joint Surg
Am. 1974;56:665-674.
37. Warren LF, Marshall JL. The supporting structures and layers on the medial side of the knee: an anatomical
analysis. J Bone Joint Surg Am. 1979;61:56-62.
135
B. THE REPLACED KNEE
Ligament lengths were calculated in similar way, after insertion of the prosthesis. Statistical
analysis was performed as described earlier. The following research questions were raised:
1. Is there a difference in ligament length before and after insertion of a knee prosthesis?
(N=12)
2. Is there a difference in ligament length before and after insertion of a Bicruciate Substitut-
ing Prosthesis (BCS, N=6)?
3. Is there a difference in ligament length before and after insertion of a Bicruciate Retaining
Prosthesis (BCR, N=6)?
4. Is the induced change from the native to the replaced knee different between BCS versus
BCR?
Results.
1. The results for all specimens pooled (N=12) are shown in g 1. It appears that insertion
of a knee prosthesis lengthens the sMCLdist on average 2 mm in mid-exion range in the
loaded setting, in a one to one comparison against the native knee. The difference is not
signicant from extension to 50 (p>0.05). In further exion, the difference is signicant:
60:p=0.03; 70:p=0.03; 80:p=0.05; 90:p=0.03; 100:p=0.03. Beyond 100, the differ-
ence is non-signicant again: p>0.05. The length of the LCL is increased with 3.6 mm
on average, over the full exion cycle. The difference is signicant at all exion angles
(p=0.001). Also, the change in length of the ligaments over the exion arc (anisometric
behaviour) was greater in the replaced group than in the non-replaced group. The maxi-
mum change in length in the replaced group was 3.1 mm for the sMCLprox, 3.9 mm for
the sMCLdist (strain 3.8%), and 8.7 mm for the LCL (strain 12%). For the native knees
these values were respectively 1 mm, 1.5 mm, and 7 mm.
No signicant change could be detected for the MPFL, nor for the length of the lateral
retinaculum (p>0.05).
2. For the BCS separately (N=6), the length of sMCLdist was increased signicantly at
110: p=0.035 and 120: p=0.029. The increase in length of the LCL was again signi-
cant: p=0.027. As for the pooled group, no signicance was found for the length of the
MPFL and the lateral retinaculum. (g 2)
3. For the BCR separately (N=6), the length of sMCL dist did not change signicantly. The
increase in length of the LCL was again signicant: p=0.026. No signicant changes were
detected for for the length of the MPFL and the lateral retinaculum. (g 2)
4. No signicant differences could be detected between the change in length of the liga-
ments induced by BCS or BCR: p>0.05 for all ligament length changes at all knee exion
angles.
136
Discussion
The insertion of a prosthesis changes the length of the medial and lateral ligaments of the
knee.
Despite this change being signicant for all exion angles for the LCL, the difference is
small being 3.6 mm on average. The MCL undergoes a similar effect, being signicant in
the mid-exion range. As the alignment of the components was correct (Chapter 3, IV) this
change cannot be attributed to malalignment. It is related to the soft tissue release, intrinsi-
cally related to the surgical process of exposure, resection of the menisci, and the bone cuts,
all leading to a small but noticeable change. As the change occurs both in the presence or
absence of the cruciate ligaments, it can be concluded that preservation of the cruciate liga-
ments does not prevent this phenomenon. The differences between the changes for MCL and
LCL can be attributed to the specic prosthesis geometry.
On the medial, isometric side of the knee, concerns could be raised over the length changes
of the MCL. One can assume that the lateral, dynamically stabilized side of the joint could
handle some strain changes, but the stiffer medial side is more prone to being harmed by the
induced effects. The reported strain for the sMCLdist is 3,8% over the full motion range.
Compared to the reported 2% strain in the native knee, this is higher. However this higher
value is still situated within physiologic boundaries.
1-3
The unchanged length of the MPFL and lateral retinaculum is an indication that soft tissue
tension on the medial and lateral side of the patella remained equal to the pre-operative sett-
ting, suggesting the surgical procedure leads to restoration of normal patellofemoral track-
ing.
REFERENCES
1. Bartel DL, Marshall JL, Schieck RA, Wang JB. Surgical repositioning of the medial collateral ligament. J Bone
Joint Surg Am. 1977; 59:107-116.
2. Borden PS, Kantaras AT, Caborn DNM. Medial collateral ligament reconstruction with allograft using a double
bundle technique. Arthroscopy 2002; 18:1-6.
3. Warren LF, Marshall JL, Girgis F. The prime static stabilizer of the medial side of the knee. J Bone Joint Surg
Am. 1974; 56:665-674.
137
Fig 1 Length of the ligaments, expressed as the distance between the insertion sites in mm, plotted against the
exion angle for the native knee and the prosthetic knee. N=12, except for MPFL, where N= 11.
Fig 2 Length changes of sMCLdist and LCL, induced by insertion of knee prosthesis, plotted for BCS and BCR
against exion angle.

138
Chapter 4: Results of the in vivo experiment
CHAPTER 4: RESULTS OF THE IN VIVO EXPERIMENT
I. Kinematics of the bicruciate substituting TKA
J. Victor, J.K.P. Mueller, R.D. Komistek, A. Sharma, M.Nadaud, J. Bel-
lemans. In vivo kinematics after a cruciate substituting TKA. A com-
parative kinematic analysis study. Clin Orthop. 2009; in press.
Abstract
Kinematic patterns in the native knee show signicant variability. Guided motion prosthetic
designs offer stability but may limit natural variability. The objective of this multi-center
study was to determine the in-vivo kinematic patterns for patients having a cruciate substi-
tuting total knee arthroplasty and to study inter-surgeon variability of a guided-motion pros-
thetic design. Three-dimensional femorotibial contact positions were evaluated for 86 TKAs in
80 subjects from three different surgeons, using uoroscopy during a weight bearing deep knee
bend. The average posterior femoral roll back of the medial and lateral condyles for all TKAs
from full extension to maximum exion, was -14.0 mm and -23.0 mm, respectively. The
average axial tibiofemoral rotation from full extension to maximum exion for all TKAs was
10.8. The average weight bearing range of motion was 109 (range 60 to 150; standard
deviation 18.7). Overall, TKA demonstrated axial rotation patterns similar to that of the nor-
mal knee, although less in magnitude than the normal knee. Surgeon to surgeon comparison
revealed dissimilarities, demonstrating the surgical technique and soft tissue handling inu-
ence kinematics in a guided-motion prosthetic design.
Introduction
In vivo kinematic patterns in subjects undergoing TKA vary considerably from the normal knee.
This is supported with ndings from uoroscopy, studies using in vitro analyses, and those
using external markers associated with gait laboratory systems.
1,2,4,8-10,12,13,16,18,19,21,22,24,25,34-37

With increasing knee exion, the normal knee reportedly experiences more posterior motion
of the lateral condyle leading to internal rotation of the tibia with respect to the femur.
18,22,25

In contrast to the normal knee, in vivo kinematic analyses suggest subjects undergoing TKA
often experience a motion pattern opposite of the normal knee where the condyles slide
in the anterior direction with increasing knee exion.
2,8,9,24,28,31,34-37
Also, in vivo kinematic
studies involving patients undergoing TKA have documented reverse rotational patterns
2,10

and lateral condylar liftoff.
13,21,34,37
There is evidence these abnormal kinematic patterns lead
to decreased range of motion.
3,11,38
In addition, these abnormal kinematic patterns possibly
relate to inferior functional performance of daily activities by patients undergoing TKA.
Noble et al
33
reported signicant differences in functional capacities between patients having
undergone successful TKA and their peers in a similar age group, never having suffered knee
pathology. The activities causing most trouble in the TKA group were related to loaded ex-
139
ion and stability: squatting, kneeling, gardening and turning/cutting. Our current inability to
restore physiologic function is multifactorial and related to irreversible damage caused by the
arthritic process, surgical damage caused by insertion of the implant, loss of proprioception
and kinematic changes induced by the prosthesis.
3,32,34
Within the spectrum of kinematic abnormalities, several studies suggest subjects having a
posterior stabilized (PS) TKA have less abnormal knee kinematics in deeper exion and
greater exion than subjects having a posterior cruciate-retaining (PCR) TKA.
8,38
However,
measurement in weight bearing conditions typically reduces the observed range of motion.
11
Both PCR and PS TKA demonstrate similar kinematic patterns in early exion activities such
as gait.
9,10
Two important features inuence stability and kinematics in the TKA implant: surface geom-
etry of the bearing surfaces and a mechanical interaction between the polyethylene and the
femoral component, eg, a cam and post mechanism. In an-vitro study, Bull et al related the
changes in knee kinematics to the articular geometry of a single radius design. They observed
only a small tibial posterior translation between 40 and 90 of exion.
6
A cruciate-retaining
implant with differing geometries on the lateral and medial condyles and a changing radius
of curvature of the femoral condyle demonstrated improved kinematic patterns.
26
Recently,
an assumption was made and implemented into a new TKA design that attempts to substitute
for the anterior cruciate ligament (ACL) and posterior cruciate ligament function by using an
asymmetric tibial plateau geometry and a dual-cam substitution attempting to produce more
normal-like kinematic patterns.
39
The aim of this study is twofold: (1) to describe the in vivo kinematics for a TKA that sub-
stitutes for the ACL and PCL and attempts to guide the motion using dual-cam constraints
function using an in vivo, uoroscopic analysis during a weightbearing, deep knee bend; and
(2) to determine and compare the differences between patient groups operated by different
surgeons . We presume if the implant behaves as a mechanically constraint-guided motion
system, no inter- and intrasurgeon differences will be observed.
For all subjects included in this study, the average age was 66.5 years (range, 40-82 years;
standard deviation [SD], 7.6). The average height, weight and BMI for all subjects in this
study was 169.1 cm (range 150-188, SD 9.3 cm), 82 kg (range 56126, SD 13.6 kg) and
28.6 kg/m
2
(range 21.2-40.7, SD 3.9 kg/m
2
), respectively (Table 1). We assessed in vivo knee
kinematics for 80 subjects implanted with 86 Journey Cruciate substituting (BCS) TKAs
(Smith and Nephew, Memphis TN) by three surgeons (JB, JV and MN) at three different
hospitals. Surgeon 1 supplied 40 TKAs from 38 patients, Surgeon 2 provided 35 TKAs from
34 patients, and Surgeon 3 provided 11 TKAs from eight patients. All surgeons used an
anteromedial surgical exposure and a measured resection surgical technique with the same
instrument set provided by the manufacturer. A chronological consecutive patient list of pa-
tients with a well-functioning TKA and judged clinically successful (Hospital for Special
Surgery scores greater than 90)
20
,was used to contact patients asking for their consent. The
list was taken from a prospective database containing follow-up of all patients undergoing
TKA. Ethical Committee or Institution Review Board approval was obtained for each of the
centers involved as well as informed consent for all patients participating in the study.
140
Fluoroscopic examinations were performed in Pellenberg, Belgium (Surgeon 1); Brugge,
Belgium (Surgeon 2), and Knoxville, TN (Surgeon 3). We asked each subject to perform suc-
cessive deep knee bends to maximum weightbearing exion. Patients were examined using a
C-arm-type uoroscopic unit, after an initial trial squat without uoroscopy. As soon as good
imaging was obtained from the full squat, the exercise was nished. The uoroscopic images
were stored on a digital video recorder for subsequent analysis.
G
r
o
u
p
N
u
m
b
e
r

o
f
s
u
b
j
e
c
t
s
N
u
m
b
e
r

o
f

T
K
A
s
A
g
e

(
y
e
a
r
s
)
H
e
i
g
h
t

(
c
m
)
W
e
i
g
h
t

(
k
g
)
B
o
d
y

m
a
s
s

i
n
d
e
x

(
k
g
/
m
2
)
All TKAs 80 86 66.5 (range,
4082; SD,
7.6)
From
77 patients
168.4 (range,
150.1188.0;
SD, 8.9)
From
77 patients
90.0 (range,
56.2 125; SD,
12.7)
From
77 patients
28.5 (range,
21.240.7; SD,
3.8)
Surgeon 1 38 40 67.1 (range,
4082; SD,
8.9)
169.9 (range,
151.9188.0;
SD, 9.4)
81.6 (range,
63.0 115.0;
SD, 11.6)
28.3 (range,
21.635.9; SD,
3.0)
5178; SD,
6.4)
166.6 (range,
150.1183.9;
SD, 8.1)
80.5 (range,
57.0125.0;
SD, 13.1)
29.0 (range,
21.2 38.6; SD,
3.8)
5874; SD,
4.9)
From 5 patients
170.7 (range,
160.0185.4;
SD, 10.2)
From 5 patients
79.4 (range,
56.2104.3;
SD, 19.4)
From 5 patients
27.4 (range,
21.340.7; SD,
7.7)
Table 1 Demographic information for all patients and groups by surgeon
SD = standard deviation.
Using a three-dimensional (3D) model-tting approach, we determined the relative position
of knee implant components in 3D from a single-perspective uoroscopic image by manipu-
lating a CAD model in 3D space using a previously described model-tting process that was
validated.
29,30
Individual uoroscopic frames were captured and analysed at full extension
(0), 30, 60, 90 (if obtained), 120 (if obtained), and maximum (Max) knee exion. The
correct 3D t was achieved when the silhouettes of the femoral and tibial implant compo-
nents best matched the corresponding components in the uoroscopic image (g 1). We de-
termined the anteroposterior (AP) contact positions for both the medial and lateral condyles,
axial rotation of the femoral component relative to the tibial component, condylar liftoff, and
weightbearing range of motion. The femorotibial contact positions were determined by nd-
ing the lowest point on the femoral component relative to the tibial component for both the
medial and lateral side. The AP position was measured as the orthogonal distance from these
points to the midline of the tibial component. Positive values indicated the position anterior
to the midline; negative values indicate position posterior to the midline.
141
The insert design of the implant studied has unequal condylar thicknesses on the medial and
lateral sides of the femoral component and the tibial insert in an attempt to recreate a physi-
ological joint line in conjunction with maintaining perpendicular bone cuts. To accurately
determine condylar liftoff, the 3D CAD model of the polyethylene insert was used in the
analysis where it was transparent during the overlay process but reappears in the image at-
tached to the tibial tray after the best t is determined. Then, the algorithm determined the
distance from each condyle to the polyethylene insert. The difference between these two
measurements was calculated to determine the occurrence of condylar liftoff at any exion
increment. We reported condylar liftoff only when the measured difference was greater than
1.0 was and then visually inspected by the operator to verify this occurrence.
We determined differences in femoral condylar positions for the medial and lateral sides
and tibiofemoral axial rotation orientation (angle) at all increments (0, 30, 60, 90, 120,
and Max) of weightbearing exion between the groups of patients, treated by the different
surgeons . We also determined differences in femoral condylar movement for the medial and
lateral sides and relative tibiofemoral axial rotation among all increments of weightbearing
exion between groups. Differences in kinematic data, including orientation at the analysed
increments and the movement between these increments, were tested between the Surgeon
groups using Students t-test and the Tukey-Kramer test using JMP Statistical Discovery
Software (SAS Institute, Cary, NC).
Fig 1 Close-up of a uoroscopy image and uroscopy image with registered CAD model of a BCS TKA from a
random patient.
142
Results
All patients undergoing TKAs in this study experienced posterior femoral rollback (PFR)
of their medial and lateral condyles from full extension to maximum knee exion. At full
extension, the average medial and lateral condyle contact position was 5.2 mm and 7.2 mm,
respectively (Table 2). At maximum knee exion, the average medial condyle contact posi-
tion moved posterior to -8.8 mm and the lateral contact position also moved in the posterior
direction to -15.9 mm. Therefore, from full extension to maximum knee exion, the average
amount of PFR was -14.0 mm medially and -23.0 mm laterally. The most noticeable condylar
movement occurred between 0 and 30 of knee exion, in which 66.3% and 68.1% of the
medial and lateral condylar movement was experienced by the subjects in this study. (Table
3)
Group Medial Lateral
Average contact
position at full
extension (mm)
Average contact
at maximum
exion (mm)
Average contact
position at full
extension (mm)
Average contact
at maximum
exion (mm)
All TKAs 5.2 (range, -4.6
to 14.6; SD,
3.5)
-8.8 (range,
-14.4 to -1.4;
SD, 2.8)
7.2 (range, -4.1
to 18.5; SD,
6.3)
-15.9 (range,
-27.2 to -6.9;
SD, 3.8)
Surgeon 1 5.3 (range, 0.7
to 13.1; SD,
2.7)
-7.8 (range,
-12.8 to -1.9;
SD, 2.0)
5.7 (range, -4.1

to 18.4; SD,
6.2)
-16.3 (range,
-27.2 to -10.4;
SD, 3.4)
Surgeon 2 6.1 (range, -3.4
to 14.6; SD,
3.9)
-9.2 (range,
-13.6 to -1.4;
SD, 2.8)
8.8 (range, -3.6
to 18.5; SD,
6.4)
-16.4 (range,
-24.1 to -6.9;
SD, 3.8)
Surgeon 3 2.1 (range, -4.6
to 7.4; SD, 3.8)
-10.7 (range,
-14.4 to -3.5;
SD, 3.9)
7.3 (range, -2.6
to 16.0; SD,
5.2)
-12.7 (range,
-19.9 to -8.4;
SD, 3.6)
Table 2 Average medial and lateral contact positions

*
for all bicruciate TKAs and groups by surgeon
*
Values in parentheses are the range of contact positions (most posterior and most anterior) and standard
deviation (SD); positive values indicate position anterior to midline; negative values indicate position
posterior to midline;
signicant difference from other surgeon groups (p < 0.05).

143
Medial PFR Lateral PFR Relative AR
Group Average
0 to 30
(mm)
Average
percent
of PFR
to max.
Average
0 to 30
(mm)
Average
percent
of PFR
to max.
Average
0 to 30
(degrees)
Average
degrees of
relative AR
to max.
*
All TKAs 9.2 (range,
1.0 to 23.4;
SD, 4.3)
66.3% 16.1 (range,
3.2
to 31.0;
SD, 7.2)
68.1% 8.2 (range,
-7.4 to 18.0;
SD, 5.4)
77.9%
Surgeon 1 8.2 (range,
2.7 to 18.9;
SD, 3.5)
63.6% 15.1 (range,
4.0 to 30.6;
SD, 6.6)
67.3% 8.1 (range,
-2.4 to 17.1;
SD, 5.1)
79.3%
2.5 to 23.4;
SD, 4.7)
73.1%
18.5 (range,
3.5 to 31.0;
SD, 7.4)
73.6%
9.0 (range,
-1.0 to 18.0;
SD, 5.5)
75.3%
1.0 to 13.0;
SD, 3.8)

54.4%
12.4 (range,
3.2
to 19.8;
SD, 7.0)
53.7%
6.0 (range,
-7.4 to 14.1;
SD, 6.3)
81.1%
Table 3 Average medial and lateral posterior femoral rollback (PFR) and relative axial rotation from full extension
to 30 exion and the average percentage of the movement from full extension to maximum exion it ac-
counts for in each TKA
*
Calculated only for those TKA with relative AR to maximum magnitude greater than 1 (94% of all
TKAs);
anterior movement;
signicant difference (p < 0.05) among indicated surgeons; AR =Axial

Rotation ; SD = standard deviation.
On average, the TKAs analysed in this study experienced a normal-like axial rotation pat-
tern from full extension to maximum exion. The average amount of axial rotation from
full extension to maximum knee exion was 10.8 (range, -4.2-24.7; SD, 6.2) (Table 4).
The average weightbearing range of motion (ROM) for the TKA in this study was 109
(range 60-150;SD, 18.7). Sixty-three (73.3%) of the subjects achieved greater than 100
of weightbearing exion and 25 (29.1%) experienced greater than 120 weightbearing
exion (Table 5).
144
AR groups All TKAs Surgeon 1 Surgeon 2 Surgeon 3
Average AR
*
10.8 (range,
-4.2 to 24.7;
SD, 6.2)
10.6 (range,
-0.3 to 24.7;
SD, 6.1)
11.9 (range,
-1.7 to 24.7;
SD, 6.1)
8.3 (range, -4.2
to 17.7; SD,
6.8)
TKA with nor-
mal AR
82 (95.3%) 38 (95.0%) 34 (97.1%) 10 (90.9%)
TKA AR greater
than 5
71 (82.6%) 33 (82.5%) 30 (85.7%) 8 (72.7%)
TKA AR greater
than 15
22 (25.6%) 10 (25.0%) 10 (28.6%) 2 (18.8%)
TKA with AR
less than 0
4 (4.7%) 2 (5.0%) 1 (2.9%) 1 (9.1%)

Table 4 Average axial tibiofemoral rotations (ARs) and number of TKAs experiencing various amounts of axial
rotation for all bicruciate TKA and groups by surgeon
*
Values in parentheses are the range from most negative (opposite) to most positive (normal) axial rota-
tion (AR) and standard deviation (SD);
opposite axial rotation pattern.

WB groups All TKAs Surgeon 1 Surgeon 2 Surgeon 3
Average WB
ROM (degrees)
109 (range, 60-
150; SD, 18.7)
108 (range, 60-
136; SD, 17.3)
110 (range, 63-
150; SD, 21.5)
109 (range, 84-
134; SD, 15.2)
Less than 90
WB ROM
(number of
TKAs)
15 (17.4%) 5 (12.5%) 7 (20%) 2 (18.2%)
Greater than
100 WB ROM
(number of
TKAs)
63 (73.3%) 31 (77.5%) 26 (74.3%) 9 (81.8%)
Greater than
120 WB ROM
(number of
TKAs)
25 (29.1%) 11 (27.5%) 11 (31.4%) 2 (18.2%)
Greater than
130 WB ROM
(number of
TKAs)
7 (8.4%) 2 (5.0%) 4 (11.4%) 1 (9.1%)
SD = standard deviation.
Table 5 Average weightbearing range of motion (WB ROM) and count of TKA by WB ROM for all bicruciate
TKAs and groups by surgeon
145
Several differences were observed between the patient groups treated by the different sur-
geons. The medial condylar position at full extension for Surgeon 3s group was more poste-
rior compared to Surgeon 1 and 2s group (p=0.007 and p=0.0008). We observed no differ-
ence in lateral condylar position at full extension. At Max exion, the mean medial condylar
position for Surgeon 1 was more anterior than for Surgeon 2 and 3 (p=0.0286 and p=0.002)
and the mean lateral condylar position was more anterior for Surgeon 3 than for Surgeon 1
or 2 (p=0.0042 and p=0.0034)(g 2 & 3). The medial AP movement from 0 to maximum
exion for Surgeon 3 was different from Surgeon 1 (p=0.0139) (g 2). All surgeon groups
were similar (p > 0.05) in AP movement on the lateral side from 0 to maximum exion
(g 3). Other differences were found between surgeon groups in AP condylar position and
movement. The most consistent difference in movement was on the medial side between Sur-
geons 1 and 2, which demonstrated differences in movement from 0 to 30 (p=0.0062), 60
(p=0.005) and 90 (p=0.0213). Other differences occurred between Surgeon 3 and the other
two groups with respect to AP condylar position for both condyles (Table 6). Although the
overall amount of axial rotation was similar, the angles at specic exion increments were
different when compared among surgeon groups. However, when comparing differences in
the ranges, the only difference occurred between from 0 and 90 exion in which Surgeon 2
had signicantly more relative axial rotation than Surgeon 1 group (p=0.0136) and Surgeon
2 group (p=0.0038) (Table 6). Similar to condylar AP movement, the majority of the overall
axial rotation occurred in the rst 30 of knee exion for all surgeon groups (Table 3).
Fig 2 Average medial anteroposterior position for all TKAs and groups by surgeon including error bars indicating
one standard deviation for all surgeon groups showing the differences between these groups. The medial
AP movement from 0 to maximum exion for patients from surgeon 3 (green) was different from patients
from surgeon 1 (blue) (p=0.0139)
146
Fig 3 Average lateral anteroposterior position for all TKAs and groups by surgeon including error bars for the
patient group from Surgeon 3. All surgeon groups were similar (p > 0.05) in AP movement on the lateral
side from 0 to maximum exion. At maximum exion, the lateral condylar position of surgeons 3 group
was more anterior than for surgeons 1 and 2 group (p=0.0042 and p=0.0034, respectively).
Different surgeon
groups
Kinematic orientations and movements
MAP MTRAN LAP LTRAN AR Angle REL AR
Surgeon 1 from
Surgeon 2
30, 60,

120,
Max
*
0-30,
*
0-60,
0-90,
0-Max
0-60
*
0-90
*
Surgeon 1 from
Surgeon 3
0,
*
60,
90,
*

120,

Max
*
30,
*
60,
*
90,
*
Max
*
0, 30,
*
60,
*

90,
*
Max,
*
120
Surgeon 2 from
Surgeon 3
0,
*
60,
90,
*

120
0-30,
*

30-60
30,
*
60,
*

90,
*
Max
*
0-30,
0-60
*
30,
*
60,
*

90,
*
Max
*
0-90
*
Table 6 Differences (p < 0.05) in kinematic orientations and movements among surgeon groups
*
Variable appears in two rows, groups in third row without variable are similar (p > 0.05);
variable
appears in all three rows indicating difference (p < 0.05) across all groups; MAP = medial AP position;
MTRAN = medial AP movement between exion increments; LAP = lateral AP position; LTRAN =
lateral AP movement between exion increments; AR Angle = AR orientation; REL AR = relative AR
between exion increments; Max = maximum; AP = anteroposterior; AR = Axial Rotation.
147
Discussion
The goal of this study was to describe the in vivo kinematics after implantation of a TKA
based on a guided motion principle. Apart from describing the observed kinematic patterns,
the study aimed at studying intra- and intersurgeon differences to evaluate the inuence of a
guided motion prosthesis with dual post-cam constraints on the in vivo kinematics.
The study has some limitations. The methodology allows to detect and describe kinematic
differences between groups of patients treated by different surgeons but it does not relate
these differences to soft tissues conditions. As the operations were performed without the use
of a surgical navigation system, the medial, lateral and anteroposterior laxity at the end of
the procedure was eventually decided by the surgeons subjective assessment of joint stabil-
ity and could not be measured in an accurate and reproducible way. In analogy, uoroscopic
measurements of the weight bearing range of motion and kinematics before the operation
were not available for comparison with the post-operative results.
Numerous knee kinematic analyses of the normal knee have documented greater mean pos-
terior motion of the lateral condyle relative to the medial condyle, leading to a mean internal
rotation of the tibia, with progressive knee exion.
9,15,23,25,27,29
Komistek et al reported the
lateral condyle achieved signicantly more posterior motion than the medial condyle, 19.2
mm and 3.4 mm, respectively, with increasing knee exion during a deep knee bend.
25
They
also reported the occurrence of intersubject variability. From full extension to maximum
knee exion, the medial condyle translation ranged from +3 mm of anterior motion to -4.6
mm of posterior motion. In comparison, the lateral condyle movement was only posterior,
ranging from -5.8 to -24.7 mm. The average tibiofemoral rotation during exion was 16.8
(range, 2.1-27.1). Banks group used computed tomography-derived bone models for mod-
el registration and added MRI-derived articular surfaces for obtaining higher accuracy of the
contact areas. They observed the greatest femoral external rotation during the squat activity
but reported no posterior subluxation of either femoral condyle in maximum knee exion.
In comparing kneel, squat, and stairclimbing motions, they found knee kinematics to vary
signicantly by activity.
29
In the native knee, different methodologies seem to reveal different
kinematic patterns: the rotational patterns are variable and may be inuenced by the bear-
ing surface forces, further inuenced by foot position, body inertia, and muscular activity. A
guided motion prosthetic knee design carries the risk of imposing a motion pattern and ex-
cessively reducing this natural variability. Our data show subjects experienced PFR of their
lateral condyle (mean 23 mm) and a lesser amount of PFR of their medial condyle (mean
14 mm) during a loaded deep knee bend. The results reported for medial condyle PFR were
greater than previously reported for the normal knee
23,25,29
, leading to axial rotation patterns
similar, but less in magnitude, to that of the normal knee (10.8 for the patients in this study
versus 16.8 in the native knee). This greater medial PFR as compared to the normal knee
raises concerns as it can potentially overload the medial structures of the knee.
The variable analysed in this study with the most immediate impact on the patients func-
tion is the weightbearing ROM. The ROM reported in this study would be considered low
when compared with passive ROM. Dennis et al
11
reported weightbearing exion can be 20
less, on average, than passive exion with the same group of patients. In this study, the aver-
age weightbearing exion was 109 with a maximum exion of 150. Sixty-three subjects
(73.3%) achieved greater than 100 of weightbearing exion and 25 (29.1%) experienced
148
greater than 120 weightbearing exion. So far, there is no clinical evidence relating certain
kinematic patterns to better outcomes like improved longevity or better wear performance.
Some authors have even suggested wear is not an issue in TKA as long as prosthetic designs
do not try to reproduce normality in kinematics.
14
In contrast, recent in vivo work comparing
laboratory data to retrieval specimen analysis supported the use of so called high exion
designs as they improve contact conditions and preserve contact area at high exion angles.
7

In addition, there is in vitro evidence that multidirectional sliding is detrimental to the poly-
ethylene, giving a theoretical advantage to guided motion.
5
Fregly and coworkers developed
another argument illustrating the close relation between kinematic behaviour and wear pat-
terns. They wrote a computer model to predict wear patterns based on kinematic in vivo
analysis and validated this model against a retrieval specimen.
17
The kinematic patterns seen in this TKA were consistent for subject-to-subject comparison,
reected by the low SDs in the data. Although the overall motion patterns were similar in
nature, we observed intrasurgeon differences in the in vivo kinematics: the relative axial
rotation between 0 and 90 exion was signicantly greater for the patients of Surgeon
2 compared to the two other surgeon groups (Table 4). This might be the result of a more
externally rotated orientation of the tibial component. Looking at antero-posterior condylar
position, the medial condyle is on average more posterior and the lateral condyle more an-
terior in the patients of Surgeon 3. This may be a phenomenon of the small sample size, but
the differences in midexion were statistically different between Surgeon 3 and the other
two surgeon groups. On the basis of the reported differences in kinematic patterns among
the surgeon groups, the null hypothesis that the implant would act as a constraint mechanical
device is refuted.
In conclusion, consistent kinematic patterns were found from patient to patient. Surgeon to
surgeon comparison revealed some dissimilarities, demonstrating the surgical technique and
soft tissue handling does play a role when using this particular implant. Although we did
not observe normal kinematics in all patients, all patients achieved femoral rollback during
exion and the axial rotation pattern was normal in pattern for 95% of the patients.
149
REFERENCES
1. Andriacchi TP. Functional analysis of pre and post-knee surgery: total knee arthroplasty and ACL reconstruc-
tion. J Biomech Eng. 1993;115:575-581.
2. Banks SA, Markovich GD, Hodge WA. In vivo kinematics of cruciate-retaining and -substituting knee arthro-
plasties. J Arthroplasty. 1997;12:297-304.
3. Bellemans J, Banks S, Victor J, Vandenneucker H, Moemans A. Fluoroscopic analysis of the kinematics of deep
exion in total knee arthroplasty. Inuence of posterior condylar offset. J Bone Joint Surg Br. 2002;84:50-53.
4. Bertin KC, Komistek RD, Dennis DA, Hoff WA, Anderson DT, Langer T. In vivo determination of posterior
femoral rollback for subjects having a NexGen posterior cruciate-retaining total knee arthroplasty. J Arthro-
plasty. 2002;17:1040-1048.
5. Blunn GW, Walker PS, Joshi A, Hardinge K: The dominance of cyclic sliding in producing wear in total knee
replacements. Clin Orthop 273:253-260, 1991
6. Bull AM, Kessler O, Alam M, Amis AA. Changes in knee kinematics reect the articular geometry after arthro-
plasty. Clin Orthop Relat Res. 2008;466:2491-2499.
7. DLima DD, Steklov N, Fergly BJ, Banks SA, Colwell CW. In vivo contact stresses during activities of daily
living after knee arthroplasty. J Orthop Res. 2008; 1549-55
8. Dennis DA, Komistek RD, Colwell CE Jr, Ranawat CS, Scott RD, Thornhill TS, Lapp MA. In vivo antero-
posterior femorotibial translation of total knee arthroplasty: a multicenter analysis. Clin Orthop Relat Res.
1998;356:47-57.
after total knee arthroplasty. Clin Orthop Relat Res. 2003;416:37-57.
10. Dennis DA, Komistek RD, Mahfouz MR, Walker SA, Tucker A. A multicenter analysis of axial femorotibial
rotation after total knee arthroplasty. Clin Orthop Relat Res. 2004;428:180-189.
11. Dennis DA, Komistek RD, Scuderi GR, Zingde SM. Factors affecting exion after total knee arthroplasty. Clin
Orthop Relat Res. 2007;464:53-60.
12. Dennis DA, Komistek RD, Stiehl JB, Walker SA, Dennis KN. Range of motion after total knee arthroplasty: the
effect of implant design and weight-bearing conditions. J Arthroplasty. 1998;13:748-752.
13. Dennis DA, Komistek RD, Walker SA, Cheal EJ, Stiehl JB. Femoral condylar lift-off in vivo in total knee ar-
throplasty. J Bone Joint Surg Br. 2001;83:33-39.
14. Dorr LD: Contrary view: wear is not an issue. Clin Orthop 404:96-99, 2002
15. Draganich LF, Andriacchi TP, Andersson GB. Interaction between intrinsic knee mechanics and the knee exten-
sor mechanism. J Orthop Res. 1987;5:539-547.
16. Feng EL, Stulberg SD, Wixson RL. Progressive subluxation and polyethylene wear in total knee replacements
with at articular surfaces. Clin Orthop Relat Res. 1994;299:60-71.
17. Fregly BJ, Sawyer WG, Harman MK, Banks SA. Computational wear prediction of a total knee replacement
from in vivo kinematics. J Biomech 2005. 38:305-314.
18. Hill PF, Vedi V, Williams A, Iwaki H, Pinskerova V, Freeman MAR. Tibiofemoral movement 2: the loaded and
unloaded living knee studied by MRI. J Bone Joint Surg Br. 2000;82:1196-1198.
19. Hsieh HH, Walker PS. Stabilizing mechanisms of the loaded and unloaded knee joint. J Bone Joint Surg Am.
1976;58:87-93.
20. Insall JN, Hood RW, Flawn LB, Sullivan DJ. The total condylar knee prosthesis in gonarthrosis. A ve to nine-
year follow-up of the rst one hundred consecutive replacements. J Bone Joint Surg Am. 1983;65:619-628.
21. Insall JN, Scuderi GR, Komistek RD, Math K, Dennis DA, Anderson DT. Correlation between condylar lift-off
and femoral component alignment. Clin Orthop Relat Res. 2002;403:143-152.
femur and tibia in the unloaded cadaver knee. J Bone Joint Surg Br. 2000;82:1189-1195.
weight bearing and non-weight bearing knee kinematics using `interventional MRI. J Biomech. 2005;38:269-
276.
24. Komistek RD, Dennis DA. Fluoroscopic Analysis of Total Knee Replacement. Surgery of the Knee. Vol 2, 3
rd

Ed. New York: Churchill Livingstone; 2001:1695.
25. Komistek RD, Dennis DA, Mahfouz M. In vivo uoroscopic analysis of the normal human knee. Clin Orthop
Relat Res. 2003;410:69-81.
26. Komistek RD, Mahfouz MR, Bertin KC, Rosenberg A, Kennedy W. In vivo determination of total knee arthro-
plasty kinematics: a multicenter analysis of an asymmetrical posterior cruciate retaining total knee arthroplasty.
J Arthroplasty. 2008;23:41-50.
150
27. Kurosawa H, Walker PS, Abe S, Garg A, Hunter T. Geometry and motion of the knee for implant and orthotic
design. J Biomech. 1985;18:487-499.
28. Lewis P, Rorabeck CH, Bourne RB, Devane P. Posteromedial tibial polyethylene failure in total knee replace-
ments. Clin Orthop Relat Res. 1994;299:11-17.
29. Lu TW, Tsai TY, Kuo MY, Hsu HC, Chen HL. In vivo three-dimensional kinematics of the normal knee dur-
ing active extension under unloaded and loaded conditions using single-plane uoroscopy. Med Eng Phys.
2008;30:1004-1012.
30. Mahfouz MR, Hoff WA, Komistek RD, Dennis DA. A robust method for registration of three-dimensional knee
implant models to two-dimensional uoroscopy images. IEEE Trans Med Imaging. 2003;22:1561-1574.
31. Markolf KL, Finerman GM, Amstutz HC. In vitro measurements of knee stability after bicondylar replacement.
J Bone Joint Surg Am. 1979;61:547-557.
32. Massin P, Gournay A. Optimization of the posterior condylar offset, tibial slope, and condylar roll-back in total
knee arthroplasty. J Arthroplasty 2006; 21:889-896
33. Noble PC, Gordon MJ, Weiss JM, Reddix RN, Conditt MA, Mathis KB. Does total knee replacement restore
normal knee function? Clin Orthop Relat Res. 2005;431:157-165.
34. Stiehl JB, Dennis DA, Komistek RD, Crane HS. In vivo determination of condylar lift-off and screw-home in a
mobile-bearing total knee arthroplasty. J Arthroplasty. 1999;14:293-299.
35. Stiehl JB, Dennis DA, Komistek RD, Keblish PA. In vivo kinematic analysis of a mobile bearing total knee
prosthesis. Clin Orthop Relat Res. 1997;345:60-66.
36. Stiehl JB, Komistek RD, Dennis DA. Detrimental kinematics of a at on at total condylar knee arthroplasty.
Clin Orthop Relat Res. 1999;365:139-148.
37. Stiehl JB, Komistek RD, Dennis DA, Paxson RD, Hoff WA. Fluoroscopic analysis of kinematics after posterior-
cruciate-retaining knee arthroplasty. J Bone Joint Surg Br. 1995;77:884-889.
38. Victor J, Banks S, Bellemans J. Kinematics of posterior cruciate ligament-retaining and -substituting total knee
arthroplasty: a prospective randomised outcome study. J Bone Joint Surg Br. 2005;87:646-655.
39. Victor J, Bellemans J. Physiologic kinematics as a concept for better exion in TKA. Clin Orthop Relat Res.
2006;452:53-58.
151
CHAPTER 5:
GENERAL DISCUSSION AND CONCLUSION
An important part of this work was dedicated to the development of an experimental model
for studying anatomy, mechanics and kinetics of the knee joint. This ex vivo model was de-
scribed in detail. Strengths of this experimental model include the robust technology allowing
double checks via alternative processing of the data. This was among others described in the
control of the stability of the reference frames with a second CT scan, the calculation of limb
alignment based on optical tracking data, as compared to implant position determined by CT
analysis, and the reliable repeatability of the tests. Also, as the three dimensional data remain
accessible, one can go back and add an additional landmark of interest for further kinematic
analysis. The triple bundle analysis of the MPFL was thus carried out in retrospect, after sug-
gestions and comments by the reviewers of the American Journal of Sports Medicine.
The model unfortunately also carries intrinsic weaknesses. Cadaver specimens are hard to
obtain and are often mechanically weakened due to chronic illness or old age. Working with
cadavers requires signicant surgical expertise and is time consuming. Many studies based
on cadaver specimens include only six specimens for that reason. Degradation of tissue qual-
ity over time occurs, potentially inducing bias in the sequence of tests performed. For that
reason we tested all specimens within 36 hours after thawing. Consequently, we were unable
to perform all tests in every specimen. Some tests were performed on subsets of six speci-
mens, as indicated in the different chapters. The model of motion, a loaded squat, is to be
defended for its simplicity and mechanical control. However, it is only a tiny reproduction of
the variety of motions, actions and loads of daily life activity. With the experience, gained in
this experiment, it seems possible to expand this experimental model to more complex mo-
tions, allowing to study ligaments and kinematics of the bones in a more natural and complex
model. Despite the described drawbacks we have been able to report interesting ndings
regarding functional anatomy and kinematics of the human knee joint and analyse the role of
a prosthetic implant.
Hypothesis testing
1. The ex vivo measured kinematic patterns of the cadaver specimens will correlate
with previously published data.
It might seem trivial trying to conrm existing knowledge on the biomechanics of the knee.
However, given the novelty of our methodological approach, combining real-time motion
recording with three-dimensional morphological analysis, validation of our experimental set-
up was needed.
In addition, after a detailed analysis of the literature and plotting of the published data in a
standard format (Chapter 3, V.A. g. 5), we were surprised to see an impressive variability.
These differences are larger than generally accepted and are amongst others caused by differ-
ences in handling and preparation of the specimens or positioning of the subjects. The human
152
Chapter 5: General discussion and conclusion
knee has an impressive potential for adapting its kinematics depending on those variables. In
that respect, it is interesting to compare the conditions imposed on the subjects when kine-
matics was measured in the active setting, as reported in the existing literature. Some sub-
jects had to perform wall sits against an angled board wall support, others were lying on the
side on a table, or sitting and performing open chain active knee extension. This explains the
wide variability in the reported results and raises the need for repeatable and well-described
conditions when studying knee kinematics. The constraints of our current technology might
prevent perfect standardization but a more systematic approach is the only logical way to
improve our understanding of active kinematics during activities of daily life. The traditional
model as it has been propagated since a long time is only one aspect of the kinematic behav-
iour of the human knee. It is quite interesting to note the quote by Hill et al
1
who described the
knee to ex as would a hinge, without axial rotation when the tibia was externally rotated.
This is exactly the phenomenon we encountered when the hamstrings were loaded.
Another explanation for the wide variability in reported results is related to the mathematical
description of the kinematics. The comprehensive description of kinematics is sometimes
regarded more difcult than its measurement,
2
so we decided to present the data in an intui-
tive format that applies to the surgical setting: relative rotation of the tibia to the femur, and
translation of the medial and lateral condyle in the horizontal plane. As the numerical output
at different levels in our ex vivo experiment did not present as an outlier in the previously
published data, and reacted in a logical and predictable way to variable conditions, we ac-
cepted the experimental framework as valid and thus conrmed the rst hypothesis.
We added new information on the role of the medial and lateral hamstrings and reported a
greater impact of the quadriceps and hamstrings on kinematics, than appears from the exist-
ing literature. This information could help explain the mechanism of noncontact anterior cru-
ciate ligament injury. Despite the recognized role of intrinsic variables like anatomical, hor-
monal, neuromuscular and genetic characteristics, mechanical interaction plays a major role
in the pathogenesis of ACL ruptures. Recent clinical research has suggested the quadriceps
could contract at the point of impact leading to an anterior vector on the proximal tibia, large
enough to rupture the ACL.
3
Our ex-vivo research supports this hypothesis and conrms the
role of the quadriceps as an ACL antagonist between 0 and 70 knee exion. Hamstrings
co-contraction has been shown to be an ACL protagonist, gaining importance with increas-
ing exion. The crucial role of agonist-antagonist relationship in the pathogenesis of ACL
ruptures was recently demonstrated by Boden et al,
4
relating a at foot position and less knee
exion at landing with the occurrence of ACL ruptures in a video analysis. Further clinical
research will be needed to clarify the in vivo role of hamstrings co-contraction, in terms of
timing, strength and body posture. Implementation of this muscular interaction in our experi-
mental model could help explain and quantify the forces involved.
2. The insertion sites of supercial medial collateral ligament, lateral collateral
ligament and medial patellofemoral ligament as anatomically described by La
Prade
5,6
will display a change in overall length of less then 10% during the loaded
squat.
The medial collateral ligament proved to be near isometric with a strain of less than 2%.
The second hypothesis can thus be accepted for the MCL. These data validate the quantita-
153
tive anatomical work by La Prade and have surgical consequences. Graft augmentation of
the chronic MCL injury with symptomatic medial laxity is feasible, provided the insertion
sites of the graft are carefully chosen, based on the above mentioned quantitative anatomic
descriptions.
The second hypothesis was refuted for the medial patellofemoral ligament, which proved
to be anisometric. The description of the different length changes of the cranial and caudal
MPFL bundle is new and previously unpublished information. It opens the door for double
bundle MPFL reconstruction with differential tensioning of the cranial (extension) and cau-
dal (30 exion) bundle. Further clinical studies will need to prove the superiority of this
technique. Potential advantages include the lessened risk of over-constraining the medial
patellofemoral compartment at certain angles in the motion arc. This is important, as chronic
patellofemoral instability often leads to medial patellar cartilage defects, caused by disloca-
tion. The thin line between providing sufcient stability to avoid recurrent dislocation and
increasing the pressure on the medial patellofemoral cartilage beyond the physiologic bound-
aries, requires indeed a more subtle surgical approach than our current techniques allow.
The lateral collateral ligament showed a different pattern between extension and 70 exion,
versus 70 exion to maximum exion. In the rst part of the motion cycle (0-70), the LCL
has less than 2% strain, and can thus be regarded isometric. Beyond 70 exion there is a
gradual trend of slackening of the LCL. The second hypothesis is accepted for the LCL in the
initial phase of the exion cycle only. This has consequences for the tensioned gap tech-
nique of balancing the exion gap and determining the rotation of the femoral component in
total knee arthroplasty. Failure to recognize the natural lateral laxity in exion could lead to
excessive external rotation of the femoral component and varus alignment in exion. As most
current systems use a tensioned gap technique with the patella dislocated, the unphysio-
logic higher tension in the dislocated patellar tendon probably counteracts to a certain extend
the physiologic laxity in the lateral compartment. With the advent of newer techniques, advo-
cating and allowing to perform gap tensioning with the patella in place, the effect of greater
physiologic lateral laxity poses an increased risk of axial plane malalignment.
3. Insertion of a knee prosthesis will alter the kinematics, mechanics and ligament
insertion site distances.
Our third hypothesis, relating to the changes, induced by a prosthesis was conrmed as the
replaced knee demonstrated a higher average internal rotation and more anterior position
of the medial condyle. However, the patterns of the curves and the ranges of rotation and
translation of the bicruciate retaining implant followed closely the native knee patterns. The
bicruciate substituting implant matched the native knee well in the passive setting, but the
posterior condylar translation in the loaded squat was higher than in the native knee. This
nding illustrates the dominance of the kinematic model that ruled when the BCS implant
was designed. The traditional model of axial rotation and asymmetric femoral condylar trans-
lation was indeed observed and generally agreed upon in the passive setting but little attention
was paid to the impact of external forces and muscular action on knee kinematics. As shown
in chapter 3 V.A., the impact of muscular action on knee kinematics is signicant, especially
on the lateral side. The described differences between the native and the replaced (BCS)
knee in the loaded setting also illustrates the intrinsic limitations of a cam-post mechanical
154
interaction for stabilizing the joint and guiding kinematics. The cam-post interaction is a hard
mechanical driver that will impose a motion pattern as soon as both components engage. The
forces acting upon the joint will be transmitted through this interaction, and be transferred to
the bone-cement-prosthesis interface, potentially jeopardizing long-term xation.
Based upon our ndings, the manufacturer of the bicruciate implant has been recommended
to reduce the amount of induced posterior femoral translation in nding a compromise be-
tween the observed active and passive kinematics.
Signicant ligament length changes were detected after insertion of the knee prosthesis. Sur-
gical damage during exposure and making of the bone cuts, different kinematic patterns and
implant geometry account for the observed difference between the native and the replaced
knee. On the lateral side, the observed length change is relatively constant over the exion
arc with a range between 2 and 4 mm. On the medial side, the change occurs mainly in the
mid-exion range. Given the natural isometric nature and stiffness of the MCL, this change
is potentially relevant in the clinical setting. It probably relates to the mid-exion tightness
that is sometimes observed during surgery when using this particular implant. It is related to
the specic geometry of the medial femoral condyle of the implant. The manufacturer was
advised to slightly increase the radius of curvature of the medial condyle in mid-exion to
accommodate this lengthening of the MCL in the mid-exion range.
4. The changes induced by insertion of a prosthesis will be greater when the cruciate
ligaments are sacriced.
Based upon the differences between the kinematics of the cruciate retaining implant versus
the bicruciate substituting implant, the fourth hypothesis can be conrmed. As discussed un-
der hypothesis three, the mechanical substitution for the intricate three-dimensional cruciate
ligament complex can only partially fulll its delicate task of maintaining sufcient stability
without excessively constraining the knee joint. However, retaining both cruciate ligaments
in total knee surgery induces new challenges. A signicant number of patients suffer attrition
of the anterior or posterior cruciate ligament, caused by the biologic changes in the arthritic
process, mechanical impingement with protruding osteophytes or prior ligament trauma. For
them, a bicruciate retaining implant offers no solution. Also, the surgical exposure is signi-
cantly more difcult when both cruciate ligaments are retained and access to the posterior
part of the tibia is more difcult with potential insertion and xation issues for the tibial
component. Design changes (as compared to traditional cruciate sacricing implants) to the
tibial component are needed to retain both tibial cruciate ligament insertion sites and to al-
low insertion and xation of the component. Minor malalignment of the tibial component,
especially in terms of joint line position and posterior slope can lead to kinematic conicts
causing early loosening or ligament failure. For all these reasons, despite delivering good
functional results, bicruciate retaining implants have not always had a succesful survivorship
track record
7,8
. We think the new kinematic insights can give an impetus to improving these
historic designs and increasing accuracy of implantation.
155
5. The ex vivo measured kinematic patterns of the cadaver specimens after insertion
of prosthesis will correlate with in vivo measurements on patients having
undergone knee replacement.
The in vivo part of this work included a large cohort of patients, studied with uoroscopic
techniques. The observed average in vivo tibiofemoral rotation of the patients was 10.8
versus 9.7 in the ex vivo study, conrming the fth hypothesis. Some limitations in the
methodology have to be understood.
The loaded squat, as set-up in the laboratory is only a partial reproduction of the complex
motor task the patients had to perform. In addition, the presence of a raised platform and the
spatial limitations induced by a uoroscopy C-arm, prevented several patients to perform a
full squat.
The only possible comparison was the axial tibiofemoral rotation. Femoral translation could
not reliably be compared as the uoroscopic in vivo technique used tibiofemoral contact
points for describing femoral translation, where the ex vivo experiment used projection of
the centre of the condyles on the horizontal plane of the tibia. The latter technique is more
accurate but could not be performed in the uoroscopy study. Despite those limitations, the
observed similarity in axial rotation validates the methodology used and supports further use
of both the ex vivo experimental model and the in vivo uoroscopy for future research on
knee kinematics.
Where to go from here?
Dealing with degenerative arthritis is more than a technical issue. The higher activity level
and functional demands for a number of middle aged people on one side and the epidemic
proportions of obesity on the other side are provoking a steep increase in the number of pri-
mary knee interventions for degenerative arthritis.
9
In the environment of an ageing popula-
tion and longer life expectancy, this boost in primary interventions will also fuel the need for
more revision procedures.
10
This sets the stage for a difcult and paradoxical task. We need
to provide solutions that offer better functional outcomes, allowing higher activity levels and
at the same time, we have to increase longevity of the procedures.
Prevention
Better patient counseling for the prevention and treatment of obesity will certainly have a
positive effect on the need for surgical interventions. Degenerative arthritis, causing func-
tional incapacity and pain of sufcient impact to warrant knee arthroplasty is mainly a dis-
ease of people suffering from overweight.
Protective surgical measures as meniscal repair have been advocated since a long time. New-
er techniques as meniscal transplantation
11
or autologous cartilage implantation
12
are promis-
ing as conservative surgical techniques to preserve joint function and will have a growing
impact on early interventions.
156
Surgical precision
An important part of the work presented was dedicated to alignment of the knee. The lit-
erature review on rotational alignment of the distal femur revealed an impressive number
of published papers on this subject. This reects in part the ruling disagreement in the use
of the optimal references for determining rotation of the femoral component during surgery.
We have tried to clarify the discussion on this subject in dening a kinematic reference line,
as published in the paper A common reference frame for rotational alignment of the distal
femur (chapter 3, II).
There is little discussion on the critical role of component alignment on kinematics, function-
al result and longevity of knee implants. If we want to improve results in terms of function
and durability, surgical precision will need to improve. The initial enthusiasm for computer
assisted surgical navigation has faded, despite initial positive reports on the post-operative
alignment in the coronal and sagittal plane, as demonstrated in prospective, randomized tri-
als.
13-16
This is in part due to practical concerns including prolonged operation times, tech-
nical difculties with optical tracker visibility and cost of the systems. But the inability to
consistently prove better outcomes in terms of post-operative axial alignment has certainly
played an equally important role in the absence of computer navigation in many operating
rooms. As we demonstrated in our work, the CT scan offers a superior insight in the mor-
phology of the distal femur and allows for a precise determination of the required reference
axes (How precise can bony landmarks be determined on a CT scan of the knee? chapter
3, I). In the past, integration between a pre-operative CT scan and intra-operative real-time
computer assisted surgery proved to be cumbersome, time consuming and difcult to achieve
unless reference frames were xed to the bone prior to the CT scan and the knee surgery.
Unfortunately, this approach required two sterile procedures under anesthesia, not encourag-
ing patients and surgeons in choosing this option. With the advent of patient matched cutting
blocks, this algorithm can be improved. Axial references axes can be accurately found on the
scan and desired rotational alignment can be determined prior to surgery. Cutting blocks are
made with rapid prototyping technology and subsequently offered for single use in surgery.
The main drawback of this pathway is ironically the insufcient CT scan visualization of
remaining cartilage in the joint, leading to an imperfect t of the patient matched cutting
block during surgery. MRI imaging overcomes this problem but is more time consuming,
expensive and less reliable for determining coronal and sagittal alignment based on imaging
of hip and ankle. A concerted effort from the orthopedic industry, the orthopedic surgeons
and health care administration will be needed to stimulate the development and increased use
of these technologies in an attempt to improve surgical precision, functional outcomes and
durability of knee replacements.
Staging the replacement procedure
As stated earlier, protective biological surgical procedures as meniscal transplants and au-
tologous cartilage implantations are carried out in an attempt to postpone the arthritic pro-
cess and avoid joint replacement procedures. The expectations are high but it is unclear at
what pace the biologic revolution will progress. In the mean time, many patients seek a
solution for incapacitating arthritis and need surgical interventions under the form of joint
157
replacement. The results of our work suggest an important role of the cruciate ligaments in
preserving normal-like knee kinematics. The technical hurdles to fully compensate for the
loss of this intricate proprioceptive, stabilizing and motion guiding ligament complex might
be to high for our current technical capabilities. The intriguing three-dimensional interaction
between the anterior and posterior cruciate ligament with its progressive bre recruitment, in
combination with the damping effect of visco-elastic articular cartilage, creates a mechani-
cal environment capable of absorbing impact loads in the presence of sufcient kinematic
freedom.
It seems acceptable to limit resurfacing of the joint to areas with exposed subchondral bone,
retaining the ligaments and the compartments with structurally intact cartilage. From the
standpoint of reproducing normal kinematic features with supposedly better post-operative
functional capabilities, this approach can be defended. The clinical experience with unicom-
partmental replacements supports this reasoning.
It remains questionable whether total knee arthroplasty is compatible with the preservation
of both cruciate ligaments. Potential kinematic conicts caused by changes in alignment and
joint line position have been mentioned above but the concept is too appealing not to explore
again in a well-controlled and supervised trial. There is good hope that the increased qual-
ity of the bearing surfaces, achieved in the last decade, in combination with more precise
surgical techniques, and better understanding of knee kinematics will allow to obtain results,
superior to those reported in the past.
17
Finally, in case of advance arthritis, substitution of the lost ligaments is needed to stabilize
the joint. Since the work of Insall and Burstein, this has traditionally been achieved with a
cam and post mechanical interaction. Despite good clinical results, kinematic abnormalities
and functional impairments have been reported. In addition, polyethylene wear on the post
causes increasing concerns. In the bicruciate prosthesis, used in our experiment, the form of
the cam and post has been adapted to substitute for anterior and posterior cruciate ligament
function. Recent research combining nite element analysis with uoroscopic analysis has
shown the cam-post mechanism to function in vivo as predicted,
18
but long term clinical
follow-up is not available yet. In our study, we have been able to demonstrate reasonable
in vivo kinematic patterns of the bicruciate stabilized implant with few outliers in terms of
axial rotation, but the average load bearing exion was 109, which averages the outcomes
in a multicenter summation analysis on posterior stabilized knee designs
19
(Chapter 4). It is
possible these outcomes suggest the limit has been reached with traditional cam and post
technology in total knee arthroplasty and new ways need to be explored. Copying nature has
often proven to be a good strategy in orthopedic treatments. As such, incorporating two ar-
ticial ligaments in a total knee implant might not be as extreme an idea as initially thought.
Economic considerations
It is beyond discussion that we face an era of shrinking resources in the health care environ-
ment. Knee arthroplasty will not escape this evolution. Efciency of the procedure, cost of
the implant, recovery time and hospital stay will continue to inuence surgical pathways. As
appears from the above, the procedure seems to evolve to a patient specic, technologically
supported operation. We will need to prove clinical superiority of this personalized high-tech
approach to justify the higher cost in comparison to the mass production of cheaper implants
158
and procedures that do not rely on these advanced technologies. Reduction of complications,
better functionality and durability will be the key words.
Conclusion
This works exposes the complexity of knee kinematics. The native knee shows an impressive
individual variability in its morphology. This anatomical variety is reected in the different
kinematic patterns observed. In addition, kinematics of the knee has been shown to change
dramatically under muscular loading. Future research will be directed towards the inuence
of external forces induced by body weight, velocity, muscle strength, foot and limb position.
Better kinematic understanding is needed to improve the outcomes of the treatment of knee
pathology, be it under the form of conservative measures, ligament and cartilage repair or
knee arthroplasty. The experimental model we developed is a potential new tool in the arma-
mentarium of technical aids that are needed for reaching these objectives.
159
REFERENCES
unloaded living knee studied by MRI. J Bone Joint Surg 2000; 82-B:1196-1198.
2. Bull, AM, Amis, AA. 1998. Knee joint motion: description and measurement. Proc Inst Mech Eng [H] 212:
357-372.
3. Hewett TE, Myer GD, Ford KR. Anterior cruciate ligament injuries in female athletes: part 1, mechanisms and
risk factors. Am J Sports Med. 2006;34:299-311.
4. B.P. Boden, J.S. Torg, S.B. Knowles, T.E. Hewett. Video analysis of anterior cruciate ligament injury. Am J
Sports Med. 2009; 37:252-9.
5. LaPrade RF, Engebretsen AH. Ly TV, et al: The anatomy of the medial part of the knee. J Bone Joint Surg Am.
2007; 89:2000-2010.
6. LaPrade RF, Ly TV, Wentorf FA et al: The posterolateral attachments of the knee: a qualitative and quantitative
morphologic analysis of the bular collateral ligament, popliteus tendon, popliteobular ligament, and lateral
gastrocnemius tendon. Am. J. Sports Med. 2003; 31; 854-860.
7. Cloutier JM, Sabouret P, Deghrar A. Total knee arthroplasty with retention of both cruciate ligaments: a nine to
eleven-year follow-up study. J Bone Joint Surg Am. 1999;81:697702.
8. Banks SA, Fregly BJ, Boniforti F, Reinschmidt C, Romagnoli S. Comparing in vivo kinematics of unicondylar
and bi-unicondylar knee replacements. Knee Surg Sports Traumatol Arthrosc. 2005;13: 551556
9. Lavernia C, Lee DJ, Hernandez VH. The increasing nancial burden of knee revision surgery in the United
States. Clin Orthop 2006 446:221-226.
10. Kurtz SM, Lau E, Zhao K, et al. The future burden of hip and knee revisions: U.S. projections from 2005 to
2030. Proceedings of the American Academy of Orthopaedic Surgeons 73rd Annual Meeting Chicago 2006 SE-
53 p597.
11. P.C.M. Verdonk, A. Demurie, K.F. Almqvist, E.M. Veys, G. Verbruggen, R. Verdonk. Transplantation of viable
meniscal allograft. J Bone Joint Surg [Am] 2006; 88:109-118.
12. Saris DB, Vanlauwe J, Victor J et al: Characterized chondrocyte implantation results in better structural repair
when treating symptomatic cartilage defects of the knee in a randomized controlled trial versus microfracture.
Am J Sports Med. 2008; 36:235-46
13. Sparmann M, Wolke B, Czupalla H, Banzer D, Zink A. Positioning of total knee arthroplasty with and without
navigation support. J Bone Joint Surg 2003; 85-B:830-5.
14. Victor J and Hoste D. Computer assisted surgery: Image-Based Computer-Assisted Total Knee Arthroplasty
Leads to Lower Variability in Coronal Alignment. Clin Orthop 2004 Nov;428:131-139
15. Chauhan SK, Scott RG, Breidahl W, Beaver RJ. Computer assisted knee arthroplasty versus a conventional jig-
based technique. J Bone Joint Surg 2004;86-B:372-7.
16. Bthis H, Perlick L, Tingart M, Lring C, Zurakowski D, Grifka J. Alignment in total knee arthroplasty. A com-
parison of computer assisted surgery with the conventional technique. J Bone Joint Surg 2004;86-B:982-87.
17. Robertson O, Knutson K, Lewold S, Lidgren L: The Swedish Arthroplasty Register 1975-1997. An update with
special emphasis on 41.223 knees operated on in 1988-1997. Acta Orthop Scand 2001; 72(5): 503-513.
18. F. Catani, B. Innocenti, C. Belvedere, L. Labey, A. Ensini, A Leardini. The Mark Coventry Award Articular:
Contact estimation in TKA using in vivo kinematics and nite element analysis. Clin Orthop. 2009 in press.
DOI 10,1007/s11999-0941-4.
after total knee arthroplasty. Clin Orthop Relat Res. 2003;416:37-57.
160
ACKNOWLEDGEMENTS
A scientic man ought to have no wishes, no affections, - a mere heart of stone, schreef
Charles Darwin vanop zijn schip The Beagle naar het thuisfront. Het uitvoeren en neer-
schrijven van een wetenschappelijke opdracht vereist een behoorlijke dosis afzondering, re-
ectie en onthechting. Toch moet je tevens onvoorwaardelijk beroep kunnen doen op vol-
doende mensen die je project steunen en genegen zijn. Ik ben dan ook oprechte dank ver-
schuldigd aan nieder die mij in de voorbije jaren heeft geholpen deze doctoraatsthesis tot
een goed einde te brengen.
Samen met Prof. Bellemans heb ik een belangrijk deel van mijn klinische en wetenschap-
pelijke orthopedische loopbaan kunnen volmaken. Het was voor mij dan ook een grote steun
Johan aan mijn zijde te hebben bij de voorbereiding, planning en uitvoering van deze the-
sis. Zijn scherpe geest en aanstekelijke dynamiek hebben mij geholpen de practische moei-
lijkheden te overwinnen en steeds te blijven denken aan de volgende stap die we moesten
nemen. Zijn bijdrage in de talrijke publicaties valt dan ook moeilijk te overschatten. Ik dank
tevens de Katholieke Universiteit Leuven. Ik heb hier een uitstekende opleiding genoten en
apprecieer de kans deze doctoraatsthesis te kunnen voorleggen. De impliciete steun van Prof
Broos en Prof Luyten was voor mij belangrijk en heeft me gesterkt om vol te houden en alle
horden te nemen.
Prof. Somville wens ik te danken voor de aangename samenwerking met zijn team aan de
Universiteit Antwerpen. Zonder de steun van het departement anatomie van Prof. Bortier had
ik de experimenten niet kunnen uitvoeren. Ik bedank Prof. Van Glabbeek, die zijn ervaring
opgedaan in het labo van de Mayo Clinic, ten dienste heeft gesteld. Francis heeft me erg
waardevolle praktische raad gegeven en me daadwerkelijk bijgestaan bij de eerste experi-
menten. Prof. Parizel hielp met de medische beeldvorming en stelde de CT scan ter beschik-
king. Het ligt niet voor de hand anatomische specimens in te scannen tussen de erg drukke
klinische activiteiten door. Michel Geldof heeft zich als dienst-verantwoordelijke zeer in-
schikkelijk opgesteld, waarvoor dank.
De technische omkadering van de laboratoriumexperimenten was indrukwekkend. Het exper-
imenteel model hebben we in nauwe samenwerking met de ingenieurs kunnen ontwikkelen.
Ik dank Prof. Vander Sloten voor zijn overzicht en begeleiding. Luc Labey was mijn baken
voor alle technische vragen. Zijn brede kennis van de biomechanica en handige inventiviteit
hebben veel bijgedragen tot de resultaten die we hebben geboekt. Dank zij Luc kon ik mijn
kennis van de fysica en de wiskunde aanscherpen, en ik denk dat hij nu de basisprincipes van
de chirurgie onder de knie heeft, want hij was mijn naaste assistent bij het prepareren van de
specimens. Pius Wong dank ik voor zijn heldere, analytische aanpak. Pius is betrouwbaar,
scherpzinnig en nauwkeurig. Hij was een onmisbare pion voor de dataverwerking en analyse
van de resultaten. Bernardo Innocenti stelde zich beschikbaar om bijstand te leveren waar
nodig. Ik heb zijn bereidwillig engagement zeer geapprecieerd.
Philippe Danckaert zorgde voor de chirurgische navigatietechnologie en was bij alle experi-
menten een onmisbare schakel. Philippe produceerde ook de erg gewaardeerde wetenschap-
pelijke video voor de scientic exhibit op de meeting van de American Academy of Orthope-
dic Surgeons in 2009. Hij stond steeds klaar om eender welk technisch of practisch probleem
aan te pakken met raad en daad. Ik dank hem dan ook van harte voor zijn loyale houding en
volgehouden steun.
161
Veel collegas hebben mij in de afgelopen jaren gevraagd hoe ik een doctoraatsthesis kon
combineren met de klinische praktijk. Dit werk is mede mogelijk gemaakt door mijn col-
legas staeden van de dienst orthopedie in het St-Lucas ziekenhuis te Brugge. Zij hebben
mij de mogelijkheid gegeven deze thesis te maken. In het bijzonder dank ik Dr Geert Van
Damme, die in mijn afwezigheid op donderdag de drukke raadplegingen verzorgde en de
geopereerde patinten opvolgde.
Voor mijn gezin was deze onderneming geen makkelijke periode. Gedurende meerdere jaren
heb ik thuis de meeste tijd doorgebracht in mijn bureel en veel vrije tijd is in dit project opge-
gaan. Dit was alleen mogelijk omdat Marleen beschikbaar was om onze kinderen Arnout,
Klaas en Emilie op te vangen en te begeleiden in hun tienerjaren. Het is dan ook aan haar te
danken dat zij met een goede basis op weg kunnen in het leven. Ten slotte is dit het mooiste
en belangrijkste project om tot een goed eind te brengen, Charles Darwin ten spijt.
162
Appendix 1. List of abbreviations
2D Two-Dimensional
3D Three-Dimensional
ACL Anterior Cruciate ligament
AL Ankle Load
AlCor Coronal Alignment
AP Antero-Posterior
BCR Bi-Cruciate Retaining prosthesis
BCS Bi-Cruciate Substituting prosthesis
CR Coefcient of repeatability
CT Computer Axial Tomography
FA Flexion Angle
FHC Femoral Hip Centre
FKC Femoral Knee Centre
FLCC Femoral Lateral Condyle Centre
FLE Femoral Lateral Epicondyle
FMAx Femoral Mechanical Axis
FMCC Femoral Medial Condyle Centre
FME Femoral Medial Epicondyle
FMS Femoral Medial Sulcus
FTAx Femoral Transverse Axis
ICC Intra-Class Correlation Coefcient
Lat Lateral
LCL Lateral Collateral Ligament
LFT Lateral Femoral Condyle Translation
Max Maximum
Med Medial
MFT Medial Femoral Condyle Translation
MPFL Medial Patellofemoral Ligament
MRI Magnetic Resonance Imaging
MSE Mean Square Error
OA Osteo-Arthritis
PCL Posterior Condylar Line
PID Proportional Integral Derivative
QL Quadriceps Load
ROM Range of Motion
SD Standard Deviation
sMCL Supercial Medial Collateral Ligament
TAC Tibial Ankle Centre
TEA Trans-epicondylar axis
TKA Total Knee Arthroplasty
TKC Tibial Knee Centre
TLCC Tibial Lateral Condyle Centre
TMAx Tibial Mechanical Axis
TMCC Tibial Medial Condyle Centre
TR Tibial Rotation
TRAx Trochlear Anteroposterior Axis
TTA Tibial Tubercle Anterior
163
Appendix 2: Anatomic denitions

!
"#!$%!&'()%*!#+!,(%-#,%('"./%!012!
"#!$%!&'()%*!#+!,#3"-#,%('"./%!012!
"#!$%!4'5465'"%*! !
REFERENCE ABBREVIATION DEFINITION
REFERENCE POINTS FEMUR !
0%+"(%!#7!"8%!8.,! 8.,!4"(! 0%+"(%!#7!$%3"!7."!3,8%(%!'(#6+*!7%&#('5!8%'*!
0%+"(%!#7!"8%!)+%%! )+%%!4"(! 0'(".5'9%!+%'(%3"!"#!'+"%(.#(!,#.+"!#7!7%&#('5!+#"48!
:'"%('5!%,.4#+*;5%! 5'"!%,.! <%%!(%7!:'=('*%!1><?!
?%*.'5!%,.4#+*;5%! &%*!%,.! <%%!(%7!:'=('*%!>@><!
1**64"#(!"6$%(45%! '**!"6$! <%%!(%7!:'=('*%!>@><!
A'3"(#4+%&.63!"6$%(45%! 9'3"!"6$! <%%!(%7!:'=('*%!>@><!
2(#485%'(!9(##/%!,(#B.&'5! "(#48!,(#B! ?#3"!,(#B.&'5!,#.+"!#7!"(#485%'(!9(##/%!
0%+"(%!&%*.'5!4#+*;5%! &%*!4#+*!4"(! 0%+"(%!#7!$%3"!7."!3,8%(%!'(#6+*!&%*.'5!4#+*;5%!
0%+"(%!5'"%('5!4#+*;5%! 5'"!4#+*!4"(! 0%+"(%!#7!$%3"!7."!3,8%(%!'(#6+*!5'"%('5!4#+*;5%!
REFERENCE POINTS TIBIA !
0%+"(%!#7!,5'"%'6! ".$!4"(! ?.*,#.+"!$%"C%%+!".$.'5!3,.+%3!,(#D%4"%*!#+!"8%!$#+%!
0%+"(%!#7!'+)5%! '+)5!4"(! 0%+"(%!#7!$%3"!7."!4.(45%!'(#6+*!".$.'5!,5'7#+*!#7!'+)5%E!
%B456*.+9!&%*.'5!&'55%#563!
?%*.'5!4#+*;5'(!4%+"(%! &44! 0%+"(%!#7!$%3"!7."!4.(45%!'(#6+*!"8%!%*9%!#7!"8%!4#("%B!#7!"8%!
&%*.'5!".$.'5!,5'"%'6!
:'"%('5!4#+*;5'(!4%+"(%! 544! 0%+"(%!#7!$%3"!7."!4.(45%!'(#6+*!"8%!%*9%!#7!"8%!4#("%B!#7!"8%!
5'"%('5!".$.'5!,5'"%'6!
2.$.'5!"6$%(45%! 2.$!"6$! 0%+"(%!#7!$%3"!7."!4.(45%!'(#6+*!"8%!".$.'5!"6$%(#3.";!
2.,!F.$65'! 2.,!7.$! ?#3"!,(#B.&'5!,#.+"!#7!7.$65'!
REFERENCE POINTS PATELLA !
1,%B!,'"%55'%! ,'"!',%B! &#3"!*.3"'5!,#.+"!#7!',%B!,'"%55'%!
=(#B.&'5!4'(".5'9%!(.*9%! ,'"!,4(! &#3"!,(#B.&'5!,#.+"!#7!4%+"('5!4'(".5'9%!(.*9%!#7!,'"%55'!
G.3"'5!4'(".5'9%!(.*9%! ,'"!*4(! &#3"!*.3"'5!,#.+"!#7!4%+"('5!4'(".5'9%!(.*9%!#7!,'"%55'!
:'"%('5!,#.+"! ,'"!5'"! &#3"!5'"%('5!,#.+"!#7!,'"%55'!
LIGAMENT INSERTIONS FEMUR !
<6,%(7.4.'5!&%*.'5!4#55'"%('5!
5.9'&%+"!
.+3!3?0:!F! HEI!&&!,(#B.&'5E!JEK!&&!,#3"%(.#(!"#!&%*!%,.!
?%*.'5!,'"%55#7%&#('5!5.9'&%+"! .+3!?=F:!F! LEM!&&!'+"%(.#(!'+*!HEK!*.3"'5!"#!'**!"6$!
:'"%('5!4#55'"%('5!5.9'&%+"! .+3!:0:!F! LEJ!&&!,(#B.&'5!'+*!HEL!&&!,#3"%(.#(!"#!5'"!%,.!
LIGAMENT INSERTIONS TIBIA !
<6,%(7.4.'5!&%*.'5!4#55'"%('5!
5.9'&%+"!,(#B.&'5!
.+3!3?0:!,(#B! N+!365463!#7!'+"%(.#(!'(&!#7!3%&.&%&$('+#363E!3'9.""'5!#+!
&.*5.+%!".$.'E!'56&.+6&!&'()%(!
<6,%(7.4.'5!&%*.'5!4#55'"%('5!
5.9'&%+"!*.3"'5!
.+3!3?0:!*.3"! G.3"'5!.+3%(".#+!#+!".$.'E!'56&.+6&!&'()%(!
LIGAMENT INSERTIONS FIBULA
:'"%('5!4#55'"%('5!5.9'&%+"! .+3!:0:!F.$! KEI!&&!,#3"%(.#(!"#!'+"%(.#(!&'(9.+!#7!7.$65'(!8%'*!'+*!IKEJ!
*.3"'5!7(#&!".,!#7!7.$65'(!3";5#.*!,(#4%33!
164
?%*.'5!,'"%55#7%&#('5!5.9'&%+"! .+3!?=F:!=! JLEJO!#7!"8%!"#"'5!,'"%55'(!5%+9"8!7(#&!,(#B.&'5!".,!#7!,'"%55'E!
3'9.""'5!&.**5%!#7!&%*.'5!$#+;!$#(*%(!
FEMORAL COMPONENT !
2.,!&%*.'5!,#3"%(.#(!4#+*;5%! F0!&%*!
2.,!5'"%('5!,#3"%(.#(!4#+*;5%! F0!5'"!
2.,!75'+9%! F0!75'+9%!
TIBIAL COMPONENT !
?%*.'5!7.+-,5'"%!&%(9%! 20!&%*!
:'"%('5!7.+-,5'"%!&%(9%! 20!5'"!
2.,!3"%&! 20!3"%&!
POST-OP CONTROL REFERENCES
! :'"!%,.!
! 2.,!F.$!
AXES FEMUR !
F%&#('5!&%48'+.4'5!'B.3! F?1B! 8.,!4"(!-!)+%%!4"(!
F%&#('5!"('+3/%(3%!'B.3! F21B! &%*!4#+*!4"(!-!5'"!4#+*!4"(!
P,.4#+*;5'(!'B.3! P,.!1B! &.*,#.+"!&%*!%,.Q'**!"6$!-!5'"!%,.!!
2(#485%'(!'B.3! 2R1B! "(#48!,(#B!-!)+%%!4"(!
=#3"%(.#(!"'+9%+"!5.+%!7%&6(! =0:! &%*!4#+*!,#3"!-!5'"!4#+*!,#3"!
AXES TIBIA !
2.$.'5!&%48'+.4'5!'B.3! 2?1B! ".$!4"(!-!'+)5!4"(!
2.$.'5!"('+3/%(3%!'B.3! 221B! &44!-!544!
=#3"%(.#(!"'+9%+"!5.+%!".$.'! =2:".$! "'+9%+"!"#!"C#!&#3"!,#3"%(.#(!,#.+"3!'"!5%/%5!#7!".$.'5!,5'"%'6!
2.$.'5!"6$%(45%!'B.3! 2.$!"6$!1B! ".$!4"(!-!".$!"6$!
LIGAMENT INSERTIONS PATELLA
165
Appendix 3. Professional career
STUDIES
Katholieke Universiteit Leuven,
Faculty of Medicine, Capucijnevoer, 3000 Leuven
Candidate in medical sciences:
1978-81 cum laude
Doctor of Medicine, Surgery and Obstetrics:
1981-82 rst probation cum laude
1982-83 second probation cum laude
1983-84 third probation magna cum laude
1984-85 fourth probation magna cum laude
Degree of Doctor of Medicine, Surgery and Obstetrics obtained on June 28, 1985.
Orthopaedic Training
Approved training program as a full time resident:
01/08/85 - 31/01/87: J.C. MULIER, University Hospital, Pellen be rg, Belgium
01/02/87 - 31/01/88: R.S.M. LING, Princess Eli sabeth Ortho paedic Hospital, Exeter, UK
01/02/88 - 31/07/88: G. FABRY, University Hospital, Pellenberg, Belgium
01/08/88 - 31/07/89: L. BECKERS, Imeldaziekenhuis, Bon hei den, Belgium
01/08/89 - 31/07/90: F. MULIER, H. Hartziekenhuis, Leuven, Belgium
01/08/90 - 31/07/91: G. FABRY, University Hospital, Pellen b erg, Belgium
ECFMG
Clinical science component and English test performed and passed on july 19-20, 1988.
Basic science component per formed and passed on January 22-23, 1991.
Fellowship
in Knee Surgery and Arthroscopy
L. PAULOS and T. ROSENBERG, Salt Lake City, Utah, USA
Certications
Certied as Orthopaedic Surgeon on august 6, 1991. Number of ministry of health
1/35019/05/480
Certied as Geneesheer specialist in de verzekeringsgeneeskunde en de medische expertise
10/03/09
166
Professional Appointments
1/8/91-31/7/95: Staff member Orthopaedic department UZ Pellenberg, University Hospitals,
Herestraat 1, 3000 Leuven.
1/8/95-present time: Orthopaedic Surgeon, Orthopaedic department, A.Z. St-lucas, St-Lucas-
laan, 8310 Brugge, Belgium.
Titels - Memberships
President of the Belgian Knee Society 1999-2001
President of the Belgian Orthopaedic Association (BVOT) 2002-2003
Board Member of the Belgian Orthopaedic Association (BVOT)
Member of the American Knee Society
Member of the European Society for Sports Medicine and Knee Surgery (ESSKA)
Member and Scientic Advisor of the European Federation of National Associations of Or-
thopaedics and Traumatology (EFORT)
Honorary member of the Socit Franaise de Chirurgie Orthopdique et Traumatologique
(SOFCOT)
International Member of the American Association of Orthopaedic Surgeons (AAOS)
Member of the Belgian Arthroscopy Association (ABA)
Board Member of the National Council for Quality Control in Medicine (NRKP)
Member of Erkenningscommissie Orthopedie
Recent International Key-Note Lectures
9/10-13/10/06: AUSTRALIAN AND NEW ZEALAND ORTHOPAEDIC ASSOCIATIONS,
Canberra, Australia.
Presidential Guest of G. Sikorsky, President of the Australian Orthopaedic Association
Kinematics and TKA, Whats the problem?
17/02/07: AMERICAN ACADEMY OF ORTHOPAEDIC SURGEONS, San Diego, USA
Presidential Guest of G. Engh, President of the American Knee Society
What will TKA look like in 2020: The Impact of Technology, Economics and Demographic
Changes
11/11/08: SOCIT FRANAISE DE CHIRURGIE ORTHOPDIQUE ET TRAUMA-
TOLOGIQUE, Paris, France
Honorary guest of D. Huten, president of SOFCOT
Alignement et cinmatique du genou
02/04/09: BRITISH ASSOCIATION FOR SURGERY OF THE KNEE (BASK), Edinburgh,
UK
Presidential Guest of C. Dodd, Oxford, president of BASK
The role of the cruciate ligaments in the native and the replaced knee
167
Recent peer-reviewed publications
Victor J, Labey L, Wong P, Innocenti B, Bellemans J. The inuence of muscle action on
tibiofemoral knee kinematics. J Orthop Res. 2009; JOR-09-0208 accepted for publication.
Victor J, Van Glabbeek F, Vander Sloten J, Parizel PM, Somville J, Bellemans J. An Ex-
perimental model for kinematic analysis of the knee. J Bone Joint Surg. [Am] 2009; in press
JBJS-D-09-00498
Victor J. Rotational alignment of the distal femur: A literature review. Orthop Traumatol Surg
Res. 2009; in press doi 10.1016/j.otsr.2009.04.011
Victor J, Van Doninck D, Labey L, Innocenti B, Parizel PM, Bellemans J. How precise can
bony landmarks be determined on a CT scan of the knee? Knee. 2009; in press doi 10.1016/j.
knee.200-9.01.001
Victor J, Wong P, Witvrouw E, Vander Sloten J, Bellemans J. How isometric are the me-
dial patellofemoral, the supercial medial collateral and the lateral collateral ligament of the
knee? Am J Sports Med. 2009; in press doi 10.1177/0363546509337407
Victor J, Mueller JKP, Sharma A, Komistek RD, Nadaud MC, Bellemans J. In vivo kine-
matics after a cruciate substituting total knee arthroplasty. A comparative kinematic analysis
study. Clin Orthop. 2009; in press CORR-D-09-00103
Victor J, Van Doninck D, Labey L, Van Glabbeek F, Parizel P, Bellemans J. A common refer-
ence frame for describing rotation of the distal femur. J Bone Joint Surg. [Br] 2009; 91-B:
683-690
Arnout N, Victor J, Cleppe H, Soenen M, Van Damme G, Bellemans J. Avoidance of patel-
lar eversion improves range of motion after total knee replacement: a prospective random-
ized study. Knee Surg Sports Traumatol Arthrosc. 2009; in press doi 10.1007/s00167-009-
0863-4
Bellemans J, Carpentier K, Vandenneucker H. Vanlauwe J, Victor J. The John Insall Award:
Both morphotype and gender inuence the shape of the knee in patients undergoing TKA.
Clin Orthop. 2009; in press doi 10.1007/s11999-009-1016-2
Innocenti B, Truyens E, Labey L, Wong P, Victor J, Bellemans J. Can medio-lateral baseplate
position and load sharing induce asymptomatic local bone resorption of the proximal tibia?
A nite element study. J Orthop Surg Res. 2009; in press
Harato K, Bourne RB, Victor J, Snyder M, Hart J, Ries MD: Midterm comparison of poste-
rior cruciate-retaining versus substituting total knee arthroplasty using the Genesis II pros-
thesis. A multicenter prospective randomized clinical trial. Knee. 2008; 15:217-21
Saris DB, Vanlauwe J, Victor J et al: Characterized chondrocyte implantation results in bet-
ter structural repair when treating symptomatic cartilage defects of the knee in a randomized
controlled trial versus microfracture. Am J Sports Med. 2008; 36:235-46
Vanlauwe J, Almqvist F, Bellemans J, Huskin JP, Verdonk R, Victor J. Repair of symptomatic
cartilage lesions of the knee: the place of autologous chondrocyte implantation. Acta Orthop
Belg. 2007; 73:145-58
168
Victor J, Ries M, Bellemans J, Robb WM and Van Hellemondt G: High-exion, motion
guided total knee arthroplasty: who benets the most? Orthopedics 2007; 30:77-79
Cool S, Victor J, De Baets T: Does a minimally invasive approach affect positioning of com-
ponents in unicompartmental knee arthroplasty? Early results with survivorship analysis.
Acta Orthop Belg. 2006, 72:709-715
Victor J. Which implant do I pick? A glossary of promises. Orthopedics 2006; 29:839-841
Bellemans J, Dhooghe P, Vandenneucker H, Van Damme G, Victor J: Soft tissue balance
in total knee arthroplasty. Does stress relaxation occur perioperatively? Clin Orthop. 2006;
452:49-52
Bellemans J, Vandenneucker H, Victor J, Vanlauwe J: Flexion contracture in total knee ar-
throplasty. Clin Orthop. 2006; 452: 78-82
Victor J, Bellemans J. Physiologic kinematics as a concept for better exion in TKA. Clin
Orthop. 2006; 452:53-58
Victor J. Do we need a national register? Acta Orthop. Belg. 2006; 72:521-523
Vandamme G, Defoort K, Ducoulombier Y, Van Glabbeek F, Bellemans J, Victor J: What
should the surgeon aim for when performing computer-assisted knee arthroplasty? J Bone
Joint Surg. [Am] 2005. 87A: 52-58
Victor J, Banks S and Bellemans J: Kinematics of posterior cruciate ligament-retaining and
substituting total knee arthroplasty. J Bone Joint Surg [Br] 2005; 87B: 646-655
Verborgt O and Victor J. Post impingement in posterior stabilised total knee arthroplasty.
Acta Orthop Belg. 2004; 70: 46-50
Victor J and Hoste D. Image-based computer-assisted total knee arthroplasty leads to lower
variability in coronal alignment. Clin Orthop. 2004; 428: 131-139
Bellemans J, Banks S, Victor J, Vandenneucker H, Moermans A. Fluoroscopic analysis of the
kinematics of deep exion in total knee arthroplasty. Inuence of posterior condylar offset. J
Bone Joint Surg. [Br] 2002; 84:50-53
Witvrouw E, Victor J, Bellemans J., Rock B, Van Lummel R, Vanderslikke R, Verdonk R.
A correlation study of objective fucntionality and WOMAC in total knee arthroplasty. Knee
Surg Sports Traumatol Arthrosc. 2002; 10: 347-351

Comparative Study On The Biomechanics of Knees

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Comparative Study On The Biomechanics of Knees

Caricato da

Copyright:

Formati disponibili

Katholieke Universiteit Leuven

Group Biomedical Sciences

Where the total variance of measurements by different observers is

. ICC values range from 0 to 1, indicating better

. ICC values range from 0 to 1, indicating better

5.7 (range, -4.1

Table 2 Average medial and lateral contact positions

signicant difference from other surgeon groups (p < 0.05).

signicant difference (p < 0.05) among indicated surgeons; AR =Axial

4 (4.7%) 2 (5.0%) 1 (2.9%) 1 (9.1%)

opposite axial rotation pattern.

Potrebbero piacerti anche