 

054
. ..._~\»..w..
2 a. .
.3. _ . i ..
1 . b J o w.
I a:

.. »

 

.1' I.. :0.
1.1.11

db...

2.! L

3 . mg
la y .
1.5!}.3,
5;. .
1!. ‘o

7

. .1.

“fun...

155.“:
~

.3 . . .A c .
:.....v1...(! A.
5.32.... 11...: 5|

1. .4. :i 1 1 .0:
. n » 1.x:

.3 l

 

 

. a}: l . 0:
1. Av: . .
no- .1 .

ﬁﬁnwgm Wm“... . . pug??? _. is... . .. gun,

 

‘ THESIS

,) r15,

(55¢)? {.2 3

This is to certify that the
thesis entitled

BIOMECHANICAL RESPONSE OF TISSUES TO VARIOUS
BLUNT IMPACT ORIENTATIONS ON THE KNEE

presented by

ERIC G. MEYER

has been accepted towards fuiﬁllment
of the requirements for the

 

 

MS degree in ENGINEERING MECHANICS

@24ch

’ ﬂ Major Professor’s Signature

ﬂf/QZ /o 5/
i /
Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRCIDateDuepBS-p. 1 5

BIOMECHANICAL RESPONSE OF TISSUES TO VARIOUS BLUNT
IMPACT ORIENTATIONS ON THE KNEE

Eric G. Meyer

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Mechanical Engineering

2004

ABSTRACT

BIOMECHANICAL RESPONSE OF TISSUES TO VARIOUS BLUNT INEPACT
ORIENTATIONS ON THE KNEE

By: Eric G. Meyer

Osteoarthritis (0A) is a chronic degenerative joint disease affecting a large
percentage of older people. Post-traumatic CA has been demonstrated to occur following
a ligament injury or a single blunt impact to a diarthrodial joint. Lower extremity injuries
are a frequent outcome from automotive accidents. The automotive knee injury criterion
is based on data for bone fracture in cadaver experiments. The current study combines
human cadaver data and a chronic post-traumatic animal model to investigate blunt
impacts on the patello-femoral (PF) and tibio-femoral (TF) joints. In Chapter 2, the role
of the impact direction on PF joint response was investigated. Many automotive
occupants sit with their legs slightly abducted and this orientation can signiﬁcantly
reduce fracture tolerance and change the orientation of patella fractures. Chapter 3
documents that compressive load axially in the tibia on an unconstrained knee will cause
the tibia to displace anteriorly with respect to the femur and produce ACL rupture. These
data may demonstrate one mechanism for non-contact ACL injuries to occur. Chapter 4
investigates this effect further by documenting the effect an axial tibia load has on the
stiffness response of the knee to an anterior knee impact. This combined loading scenario
reduces the amount of shear displacement between the tibia and femur. Finally, Chapter 5
documents accelerated subchondral bone changes in rabbits at 12 weeks, following a
single blunt impact of approximately 50% of the fracture force. The data presented in this
thesis may be applicable to injury prediction and the development of a new knee for the

anthropomorphic dummy used in automotive crashes.

ACKNOWLEDGEMENTS

I would like to thank my mentor Dr. Roger Haut for his expertise, leadership,
support and dedication throughout my research at the Orthopeadic Biomechanics
Laboratories (OBL). I would also like to acknowledge Dr. G. Thomas Mase and Dr.
Thomas J. Pence for their excellence in teaching and for serving on my committee. I
would also like to thank Clifford Becket, Jane Walsh and Jean Atkinson for their hard
work, knowledge and willingness to help me. Finally, I would like to acknowledge
everyone who worked along side me at OBL; Dan Phillips, Eric Clack, Steve Rundell,
Micheal Sinnott, Eugene Kepich, Joanne Ewen, and Anthony Meram and to everyone
involved in MSU’s Chapter of the Biomedical Engineering Society for their help and
friendship.

Most of all I would like to thank my wife, Susan Meyer and my parents, Kevin
and Marcia Meyer for their unwavering love and support of me throughout my academic

career and life.

iii

TABLE OF CONTENTS

RESEARCH PUBLICATIONS ................................................................... v
LIST OF FIGURES ............................................................................... vii
LIST OF TABLES .................................................................................. ix
INTRODUCTION ................................................................................... 1
CHAPI'ER 2: The effect of impact angle on knee tolerance to rigid impacts
Abstract .................................................................................... 23
Introduction ................................................................................. 25
Methods .................................................................................... 29
Results ...................................................................................... 30
Discussion ................................................................................. 45
References ................................................................................. 57

CHAPTER 3: Excessive compression of the human tibio-femoral joint causes ACL
rupture before bone fracture.

Abstract ..................................................................................... 60
Introduction ................................................................................. 61
Methods ..................................................................................... 62
Results ...................................................................................... 65
Discussion .................................................................................. 67
References .................................................................................. 71

CHAPTER 4: The effect of axial load in the tibia on the response of the 90° ﬂexed knee
to blunt impacts with a deformable interface.

Abstract ..................................................................................... 73
Introduction ................................................................................ 75
Methods .................................................................................... 80
Results ....................................................................................... 88
Discussion ................................................................................ 102
References ................................................................................ 107

CHAPTER 5: The chronic post-traumatic effect of a single blunt impact in tissues of the
rabbit knee.

Abstract ................................................................................... 1 16
Introduction ............................................................................... 1 17
Methods ................................................................................... l 19
Results .................................................................................... 125
Discussion ................................................................................ 130
References ................................................................................ 132
CONCLUSIONS ................................................................................. 134

iv

RESEARCH PUBLICATIONS

PEER REVIEWED MANUSCRIPT S

1.

Meyer EG, Haut RC. The Effect of Knee Impact Angle on Tolerance to Rigid
Impacts. Stapp Car Crash Conference (2003)
Meyer EG, Haut RC. Excessive Compression of the Human Tibio-Femoral Joint

Causes ACL Rupture before Bone Fracture. Journal of Biomechanics (In Review)

. Meyer EG, Sinnott MT, Jayaraman , Haut RC. The Effect of Biaxial Impact on

the 90° Flexed Human Knee Joint Stability and Injury Tolerance. Stapp Car Crash
Conference (In Preparation)

Kirsch J, Dejardin L, Meyer E, Decamp C, Haut R. Effect of Intramedullary Pin-
plate Combination on the Mechanical Properties of Pantarsal Arthrodesis; A
Comparitive in Vitro Analysis in Dogs. American Journal of Veterinary Research
(In Press)

Von Pfeil D, Dejardin L, Meyer E, Weerts R, Decamp C, Haut R. Biomechanical
Comparison of an Interlocking Nail and a Plate-Rod Combination; An in vitro
analysis in a canine fracture model. American Journal of Veterinary Research (In
Preparation)

Wertheimer S, Hansen M, Le L, Meyer E, Haut R. Comparison of 3.5 mm and 4.0

mm Cortical Syndesmotic Screw Pull Out Strength. (In Preparation)

PEER REVIEWED ABSTRACTS

1. Haut R, J ayaraman V, Sevensma E, Meyer E. Anterior Cruciate Ligament
Rupture Due to Compression of the Human Tibiofemoral Joint. Orthopaedic
Research Society (2003)

2. Kirsch J, Dejardin L, Meyer E, Decamp C, Haut R. Mechanical evaluation of
canine pantarsal arthrodesis. American College of Veterinary Surgeons (2003)

3. Von Pfeil D, Dejardin L, Meyer E, Weerts R, Decamp C, Haut R. Biomechanical
Comparison of an Interlocking Nail and a Plate-Rod Combination; An in vitro
analysis in a canine fracture model. Veterinary Orthopedic Society (2004)

4. Meyer E, Meram A, Haut R. Single Mechanical Impact Produces Post-traumatic
Bone Injury in Tissues of the Rabbit Knee. Orthopaedic Research Society (In

Preparation)

vi

LIST OF FIGURES

INTRODUCTION
1.1: Anatomical features of diarthrodial joints such as the knee ...................... 2
1.2: Radiographic diagnosis of osteoarthritis by loss of joint space .................. 3
1.3: Osteoarthritis disease progression from normal to severe ........................ 5
1.4: Anatomical structures of the knee ................................................... 6
1.5: Knee muscle groups and their motion effect ....................................... 7
1.6: Cadaver sled test producing blunt impact to the knee ........................... 10
1.7: Isolated knee impact of the patello-femoral joint ................................. 11
1.8: Frontal impact to a human knee joint ﬂexed 90° ................................. 14
1.9: Inherent tilt of the tibial plateau and anterior displacement of the tibia ....... 15
1.10: Bone bruise in the lateral femoral condyle ....................................... 17
1.11: Surface ﬁssure of the AC and an occult microfracture at the tidemark ...... 18
1.12: Indentation test to measure the mechanical properties of cartilage ........... 19
CHAPTER 2
2.1: Impact loading directions of the knee-thigh-hip complex ...................... 27
2.2: Isolated human knee joint ﬂexed 90° with impact loading on the patella... 30
2.3: Oblique orientation of the knee joint with respect to the femur ............... 30
2.4: Representative load-time curves for sequential impacts on a specimen ....... 33
2.5: Representative load-displacement curves for sequential impacts ............... 35
2.6: Representative load-displacement plots and linear regression .................. 37
2.7: Bar graph of linear regression slopes for axial and oblique tests 1, 2 and 3...38
2.8: Representative pressure distributions of a direct impact ........................ 39
2.9: Bar graph of retropatellar surface average pressure and contact area ......... 40
2.10: Representative photograph of patella fractures on paired human knees. ....41
2.11: Schematic representations of the patellar fracture patterns on the ............ 42
2.12: Bar graph of peak loads generated in axial versus oblique impacts .......... 42
2.13: Specimen showing atom medial retinaculum and AC surface lesion ....... 43
2.14: Fracture load-age plot for axial and oblique directions ........................ 44

2.15: Abduction angle-stature plot for drivers in a two-hour automobile trip. ....54

CHAPTER 3
3.1: Anterior neutral-position shift from the inherent tilt of the tibial plateau. ....62
3.2: Femur ﬁxture allowing motion in the X-Y plane and rotation of the tibia. ...64

3.3: A typical ACL injury in an unconstrained knee preparation .................... 65
3.4: Relative joint motions caused by TF compressive loads ........................ 66
CHAPTER 4

4.1: Orientation of anterior knee impact on the patella and tibial tuberosity ...... 77
4.2: Combined loading scenario caused by IP contact and ﬂoor pan intrusion. . ..78
4.3: Schematic of the 90° ﬂexed knee after dissection and potting in epoxy ...... 81

4.4: Biaxial impact experimental set-up for the 90° ﬂexed knee ..................... 82
4.5: Lateral CT image of the 90° ﬂexed knee with tibial plateau Slope ............ 86
4.6: Knee joint BMD measurement region in a DEXA scan ......................... 87

vii

4.7: Hexcel deformation from a representative specimen (31382R6) ............... 88

4.8: Load-time curves for tests 1-3 from a representative specimen ................ 89
4.9: Load-displacement curves from test 3 ............................................. 89
4.10: Bar graph of anterior knee stiffness with increasing ATL ..................... 90
4.11: Tibia drawer displacement with increasing axial tibia load ................... 93

4.12: Bar graphs of percentage of load carried by the tibial tuberosity ............. 95
4.13: Pressure Film from a representative specimen .................................. 96
4.14: PF force from pressure ﬁlm in tests with increasing axial tibia load. . . . . ....97

4.15: Representative injuries to paired human knees .................................. 98
4.16: Pressure ﬁlm for a fractured specimen (31382L) .............................. 100
4.17: CT scan of specimen 31390R prior to and following impacts ............... 101

4.18: Photograph of specimen 31382R during dissection ........................... 101

4.19: CT scans of specimen 31382R after progressive impacting .................. 102
4.20: Orientation of applied and resultant forces for a 90° ﬂexed human knee...107

CHAPTER 5

5.1: Positioning of the rabbit’s 90° ﬂexed knee directly beneath the actuator... 120
5.2: Custom epoxy interface in position on the knee ................................. 120
5.3: Rabbit knee secured with straps beneath a deformable interface ............. 122
5.4: Anesthetised rabbit positioned for a gravity-accelerated mass impact ....... 122
5.5: Indentation material testing machine ............................................. 124
5.6: Schematic of the rabbit tibial plateau histology slide ........................... 126
5.7: Gross dissection photographs of cartilage ﬁssures and meniscal tear ........ 127
5.8: Results from mechanical indentation testing of the 12 week animals ........ 128
5.9: Results of histomorphometric scoring of 12 week rabbits ..................... 129

5.10: Increased subchondral bone thickness and splitting in the AC .............. 129

5.11: Trabecular bone porocity of 12 week knees .................................... 130

viii

LIST OF TABLES

INTRODUCTION
1.1: Results for rigid impacts to the PF joint in a 90° ﬂexed knee .................. 12
1.2: Rigid and deformable results from PF impacts at various ﬂexion angles. ....13
CHAPTER 2
2.1: Biomechanical data for axial impacts .............................................. 35
2.2: Biomechanical data for oblique impacts ........................................... 36

2.3: Pressure ﬁlm data from the retropatellar and anterior surface of the patella..39

CHAPTER 3
3.1: Experimental failure data for 60° and 120° unconstrained knee joints ........ 66
3.2: Experimental failure data for 90° unconstrained knee joints ................... 66
CHAPTER 4
4.1: Impact sequence for biaxial knee experiments .................................... 83
4.2: Biomechanical data for tests 1-3, 3 kN AKL with varying ATL ............... 91
4.3: Biomechanical data for tests 4-6, 6 kN AKL with varying ATL ............... 92
4.4: Multiple linear regression variables for tibia drawer ............................. 93
4.5: Biomechanical data for test 7, 9 kN AKL impacts ............................... 94
4.6: Multiple linear regression variables for PF force ................................. 97
4.6: Biomechanical data and injuries for failure impacts ............................. 98
CHAPTER 5
5.1: Histomorphometric scoring table .................................................. 125

ix

INTRODUCTION

Treatments for medical conditions have become increasingly sophisticated and
powerful in the last century. These breakthroughs have increased life expectancy, quality
of life and allow people to do things at increased age that never would have been possible
a century ago. However, with increased age our population also becomes more
susceptible to other health problems, especially arthritis, diabetes and cancer. The
common thread among these diseases is that there currently is no cure. In fact, in many
ways we do not even understand the basic science behind what causes these diseases or
their progression.

Nearly 50% of Americans over the age of 65 have some form of arthritis (CDC
Fact Sheets, 1997) with total costs of the disease approximately $80 billion per year.
Osteoarthritis (0A), or degenerative joint disease, is the most common musculoskeletal
disease and the most common form of arthritis affecting 20.7 million people in the USA
alone (1999). From the patient’s perspective this disease is characterized by diarthrodial
joint pain and tenderness, loss of range of motion and localized inﬂammation around the
affected joint. Since the main function of diarthroidial joints is to allow body movement
and locomotion, this disease has grave consequences for a patient’s quality of life. The
most commonly affected diarthrodial joints are the knee (Figure 1.1), hip, shoulder and
ﬁngers. The knee and hip are both necessary for locomotion so OA often renders people

unable to stand or walk without intense pain.

 
   

Joint Cavity Articular cartilage

Cancellous Bone
and Marrow

   

Articular
Cartilage

Figure 1.1. Anatomical features of diarthrodial joints such as the knee.

Diarthrodial joints allow movement by transferring forces between muscles and
bones with very little friction, while also providing cushioning and distributing the forces
over larger areas. Articular cartilage (AC) is a soft, near frictionless material that covers
the ends of bones in diarthrodial joints and accomplishes these functions. This material is
a form of connective tissue and is primarily composed of cells (chondrocytes), ﬁberous
matrix (collagen), a ground substance (proteoglycans) and interstitial ﬂuid (mostly
water). The solid phase (chondrocytes, collagen and proteoglycans) accounts for 15-30 %
of the wet weight of AC. The remaining 70-85% of the weight is water that pressurizes
the cartilage. Proteoglycans have a negative charge that attracts electrolyte ions, this
creates an osmotic gradient between the intercellular and the extracellular ﬂuid. Collage
ﬁbers provide the structural support for the surface tension that is developed by the
pressurized cartilage. As the cartilage is loaded and compressed during normal activities,
ﬂuid is squeezed out, similar to squeezing a sponge. There is a frictional component to

this ﬂuid ﬂow that helps the cartilage respond appropriately depending on the level of

compression. Articular cartilage is nonvascular (no blood supply), so the ﬂuid ﬂow is
also important for transporting nutrients and waste products into and out of the cartilage.
Osteoarthritis is a degenerative disease that affects the cartilage and subchondral
bone of diarthrodial joints. It is characterized by irregular loss of cartilage in areas of high
load, sclerosis of subchondral bone (SB), subchondral cysts and osteophytes.
Biomechanically, the cartilage material properties, such as the tensile, compressive and
shear moduli change. The hydraulic permeability of the cartilage also changes due to
degradation of the collagen, causing increased water content and excessive swelling.
Additionally, the SB thickness and stiffness change as it undergoes remodeling due to
changing stress levels. The progression of this disease involves chronic fragmentation of
the cartilage surfaces and remodeling of SB. Clinical diagnosis of osteoarthritis comes
only when a signiﬁcant reduction of the joint space is seen radiographically (Figure 1.2)
(Hamerman, 1989), although MRI is proving to be a useful tool in diagnosing 0A at an

earlier point (Hodgson et a1 1992).

Loss ofJoint
Space _ . Osicophylcs

 

Figure 1.2. Radiographic diagnosis of 0A by loss of joint space. Loss of joint space is

due to wearing away of the AC and typically associated with osteopyhtes of the bone.

The initiation and progression of OA are not fully understood. In many patients
the disease is due to a lifetime of high stresses in a particular joint from an occupation or
recreational activity (Dieppe et al 1992). A signiﬁcantly higher percentage of patients
with ligament tears or sprains go on to develop CA as a result of the change in the way
forces are transferred through the joint after an injury. There is also the possibility of a
single mechanical insult initiating the disease process, especially if there is bone fracture
or soft tissue damage near the articulating surfaces (Chapchal et al. 1978, Davis et al.
1989, Nagel et al. 1976, Volpin et al. 1990). Association between end stage OA from one
particular cause have been difﬁcult due to the long incubation time before chronic
changes occur and radiographic evidence appears, typically at least 10-20 years (Wright
1990).

There are two theories for how OA initiation and progression occur (Figure 1.3).
The ﬁrst is that ﬁssures and damage of the AC occur due to a mechanical insult thus
changing the material properties and its ability to adequately absorb and transmit loads.
This changes the stresses seen by the SB and initiates bone remodeling. The remodeled
SB has increased stiffness that damages the overlying cartilage and the cycle continues.
The other theory involves a similar cycle, except the mechanical insult causes trauma
initially to the SB in the form of occult rrricrocracks or “bone bruises”. These cracks
initiate remodeling of the SB, which in turn damages the AC by increasing the stresses

seen in the overlying tissue, and the chronic degradation cycle begins (Vellet et al. 1991).

Thickening of the

Healthy articular
subchondral bone

cartilage surface

 

Figure 1.3. Osteoarthritis disease progression from healthy to end stage.

Since this process is still not clearly understood there is a lot of research currently
under way in this area. Many investigators are focusing on new ways of diagnosing
degenerative joint disease earlier, while others are looking for possible treatments or
interventions for patients that are already in the degenerative cycle. There is also a group
of researchers investigating causes, such as a single blunt impact, and ways of preventing
these injuries in the ﬁrst place.
Knee Anatomy

The knee is a synovial joint in the leg where three bones, the femur, tibia and
patella, meet (Figure 1.4). The femur and tibia are two long bones in constant contact that
rotate relative to each other (similar to a hinge) in order to produce knee ﬂexion. The
femur has two articulating surfaces, the medial and lateral condyles that are in contact
with the medial and lateral tibial plateaus. The patella, or kneecap, is a sesamoid bone
that rests on the anterior articulating surface of the femur to protect the knee and act as a
lever for transferring muscle forces between the upper and lower leg.

The knee has two ﬁbrocartilage pads, the medial and lateral menisci between the
femur and tibia, in addition to the AC on each bone’s articulating surface. There is a

ﬁbrous capsule of many ligaments and tendons that hold the joint together, as well as

keep the synovial ﬂuid in a “closed system”. The knee joint is made up of medial and
lateral collateral ligaments and the patellar ligament and quadriceps tendon that support
the patella. The anterior cruciate ligament (ACL) and posterior cruciate ligament (PCL)
are two ligaments on the interior of the knee joint that prevent anterior and posterior

motion of the tibia relative to the femur.

Articular
Carilage

Posterior
Cruciate

Anterior Li gamenr

Cruciate
Ligament

Lateral . ,‘é‘ew’r L

.4-F y

; Medial
, Meniscus

I"

Meniscus ’ t .
”h

M

 

Figure 1.4. Anatomical structures of the knee.

The major muscle groups that produce knee ﬂexion and extension are the
hamstrings and quadriceps, respectively (Figure 1.5). The hamstrings are really four
individual muscles that originate along the edges of the pelvis and insert at the
tibia/ﬁbula. The quadriceps also encompasses four muscles that originate on the pelvic
girdle and insert into the patella and reach the tibial tuberocity via the patellar ligament.
Muscle force plays an important role in the stability of the knee by assisting the knee

ligaments in constraining the joint.

 
 
     

quadricep
muscles
(contract)

Figure 1.5. Knee muscle groups and their motion effect.

Knee Injuries

In addition to 0A, the knee is susceptible to many other acute injuries caused by
blunt impact from a fall, an automotive accident, sports or work related circumstances, or
any number of other possibilities. High severity injuries include gross bone fracture and
ligament trauma such as sprains, dislocation, rupture or avulsion. Two types of acute
injury scenarios will be investigated in the current study; automotive accidents primarily
producing bone fracture and sports related ligament injuries. The most common injuries
from frontal automotive crashes are patellar or femoral fractures, while the most common
sports injury is an ACL tear.

Automotive Trauma Literature Review

Automobile accident injuries are the cause of a signiﬁcant percentage of societal
cost from medical spending, lost workdays and reduced quality of life. In the United

States alone 3.2% of annual medical costs are for injuries that result from motor vehicle

collisions, second only to cancer (Miller et al. 1998). Another study estimates these
annual costs at $21.5 billion (Miller et al. 1995).

Lower extremity injuries can be the cause of permanent disability and
impairment (States 1986), and are a frequent outcome of automotive accidents (Fildes et
al. 1997). These injuries have been shown to comprise a signiﬁcant percentage of the
costs associated with motor vehicle trauma. MacKenzie et al. (1988) document that 40%
of the automotive trauma costs in Maryland are from lower extremity injuries. Luchter
and Walz (1995) ﬁnd that 27.8% of the injuries in the 1993 National Accident Sampling
System (NASS) database are for the lower extremity, making it the most frequently
injured body region.

Many studies show the most commonly injured lower extremity region is the knee
(Hartemann et al. 1977 and Bourrett et al. 1977, Fildes et al. 1997 Atkinson and Atkinson
2000). Atkinson and Atkinson (2000) document that for the years 1979-1995
approximately 25% of all injuries recorded in the NASS database are to the lower
extremity and 10% are knee injuries. The knee injury rates remain relatively constant
over this period even with the addition of mandatory seat belt laws and airbags. Most of
these knee injuries are rated Abbreviated Injury Scale (AIS) l (Fildes et al. 1997 and
Atkinson and Atkinson 2000), but even these “mild” injuries have signiﬁcant costs
associated with them, between approximately $1,400 and $2,500 per injury (I-Iendrie et
al. 1994 and Miller et al. 1995). For more serious knee injuries (AIS 2-4), fracture is the
most common, however sprains, ligament avulsions and ruptures, contusions and
dislocations are also documented. These more serious knee injuries result in 40 or more

lost workdays, 40-50% of the time, however may not have been the only injury sustained.

Patella fractures are shown to be the most common knee injury rated AIS 2-3 with
femur fractures occurring slightly less frequently (Atkinson and Atkinson 2000). In the
years 1993-1995 these fractures account for 2.4 and 2.2, out of every 1000 injuries
respectively. Together with tibial plateau fractures they account for between 25-50% of
all knee injuries. Tendon and ligament injuries combined accounted for approximately
25% of all knee injuries.

Knee fractures occur in accidents with a change of velocity of less than 45 kmph
90% of the time and with zero intrusion 61% of the time (Fildes 1997). Fractures were
due to contact with the steering column or instrument panel in most cases. Nagel and
States (1977) suggest major injury to the knee does not occur when the knee contacts
smoothly contoured sheet metal, only rigid structures. Deformable knee bolsters have
been commonly used in recent model cars, but are not standardized or regulated.
Additionally, it has been shown that even with a deformable interface that protects
against gross bone fracture there is still the possibility of causing microscopic damage to
the subchondral bone (Atkinson and Haut 2001). These “bone bruises” have been shown
to occur at much lower loads than knee fracture. Since most knee injuries occur to people
11-40 years old there is plenty of time for a chronic disease to progress as a result of

these nricrofractures.
Fracture Experiments

The earliest biomechanical impact studies on the knee were for seated fresh and
embalmed cadavers with either sled (Figure 1.6) or pendulum drop impactors (Patrick et
al. 1965, Powell et al. 1975 and Cooke and Nagel 1991) or a pneumatic cylinder (Melvin

et al. 1975). These studies used rigid and padded interfaces on 90° ﬂexed knees and

documented fractures for a wide range of loads between approximately 7.3 kN and 21.0
kN. Rigid impactors result in transverse fractures of the patella, split fractures of the
femoral condyles, and mid-femoral shaft fractures in 84%, 44%, and 25% of cases,

respectively. Padding reduces fracture to the patella by 65%, to the femoral condyles

 

Figure 1.6. Cadaver sled test producing blunt impact to the knee (Patrick et al. 1975).
by 14%, and to the femoral shaft by 6% (Viano et al., 1980). Additional data was
collected in the Melvin et al. and Powell et al. studies from strain gages attached to the
femoral shaft. The bending moments in the femur were recorded during axial and off-axis
impacts to the knee. These studies serve as the basis for the current automotive knee
injury criterion of 10 kN in the femur during a 30 mph full frontal crash into a rigid
barrier. However, in the time since these ﬁrst studies were reported there has been more
information collected about knee fracture mechanisms with a wide range of variables.

Recent studies have used isolated fresh human cadaver knees to investigate many
of the factors important in predicting knee fracture (Figure 1.7). These experiments
document many of the details that were overlooked by the preliminary knee impact

studies.

10

Impact Force

5

 

 

 

 

 

 

 

l—___l

3— Pressure Film

  

Figure 1.7. Isolated knee impact of the patello-femoral joint.

Impacts have typically been applied in one of two directions, either aligned with the axis
of the femur or aligned with the axis of the tibia. Impacts along the axis of the femur
mainly affect the patello-femoral (PF) joint, but depending on the contact area may also
cause shear loading of the tibia with respect to the femur. These impacts simulate the
knee impacting the dashboard or knee bolster in an automotive accident. Tibia axis
impacts affect the tibio-femoral (TF) joint and are meant to simulate intrusion of the
ﬂoor-pan applying a force through the ankle up into the knee.

Recent rigid impacts studies for the PF joint at 90° ﬂexion have shown the force
required for fracture to be between 4.5 and 8.5 kN (Table 1.1). This research is mostly on
aged cadavers, due to availability constraints, but is comparable to many studies that are
also used aged cadavers. A study by Atkinson et al. (1997) documents a signiﬁcant
relationship between the geometry of the cartilage surface and the peak load at fracture.
As the area of cartilage decreased relative to the area of bone there was a signiﬁcant

decrease in the peak load at fracture. They did not ﬁnd any signiﬁcant trends between

11

peak load and age, sex, height or weight. Other studies, on the other hand, document that

the strength of cortical bone decreases with specimen age (Burstein et al. 1976, Yamada

 

 

 

 

 

 

 

 

 

 

1970).
Study Peak Load Number of Injuries Age Time to Peak
(kN) Specimens Pat Fx Fem Fx Occult (yrs) (ms)
Atkinson and Haut, 2001a 4.7 (1.6) 6 6 0 5 69 (12) 4.7 (1.6
Atkinson and Haut, 2001b 5.5 (1.8) 6 6 0 5 - 3.4 (1.1
Ewers et al., 2000 4.6 (1.0) 6 3 0 4 75 (13) 4.9 (0.9)
l 4.5 (1.2) 6 1 0 3 7&3) 54 (3)
IAtkinson at L1. 1997 5.0 (1.5) 6 3 1 4 65 (12) 4.1 (0.5)
Haut and Atkinson, 1995 6.7 (2.1) 10 8 1 3 77 (7) 5
6.7 (1 .3 10 7 2 5 48 (9) 5
IHaut, 1989 8.5 (3Q 8 4 4 - 72 (11) 9.4 (1 .2)
Powell et al., 1975 11.2 (2.4) 4 4 o - 58 (14) -
LMelvin et al., 1975 19ﬂ3§L 3 3" - - 65 (14) 4.7 (1.9)

 

 

Table 1.1. Previous results for rigid impacts to the PF joint in a 90° ﬂexed knee.

Atkinson and Haut (2001a) investigated the effect of knee ﬂexion angle on
patello-femoral impacts (Table 1.2). They documented an increase in fracture tolerance
with increasing ﬂexion angle. Impacts at knee ﬂexion of 60° resulted in patella fracture at
a peak load of 2.9 kN, while 120° ﬂexion impacts resulted in femur and patella fracture at
a signiﬁcantly different load of 6.4 kN (Table 1.2). This study also found that the location
of the injuries changed depending on ﬂexion angle. Impacts at 60° ﬂexion caused injuries
located near the distal edge of the patella, while 90° ﬂexion injuries were located
centrally on the patella and 120° injuries were near the proximal edge of the patella.

Another study by Atkinson and Haut (2001b) investigated the role of impact
interface on fracture load with a range of ﬂexion angles. Rigid and deformable (3.3 MPa
crush strength honeycomb material) interfaces were used on paired specimens. Generally,
the deformable interfaces increased the peak load for fracture. However, this effect was

only statistically signiﬁcant at 120° ﬂexion. This study veriﬁed the results of the previous

12

ﬂexion angle study by also documenting a signiﬁcant increase in fracture tolerance with
increased ﬂexion angle. Other studies have also investigated the effect of deformable
interfaces on fracture load (Atkinson et al., 1997, Hayashi et al., 1996). They also show
that signiﬁcant increases in fracture load can be achieved with a deformable interface and
document a relationship between the stiffness of the interfaces and peak load at fracture.
To protect the knee against fracture the stiffness interface was suggested by Hayashi et al.

(1996) to have ultimate crush strength of approximately 90 psi.

Flexion Load Fx. Fx. Fx.

120
Sub Fx.
Sub Fx.
120 Sub Fx.

1
120 1.4
Deform b
Delorm b 1.4 1 1
120 Deform b 1
Table 1.2. Previous injury results from PF impacts at various knee ﬂexion angles with

 

rigid and deformable interfaces.

Many recent studies on the PF joint during blunt knee impact have used pressure
ﬁlm to document the contact and load distributions (Atkinson and Haut, 2001; Atkinson
and Haut, 2001; Ewers et al., 2000; Atkinson et al., 1997; Atkinson and Haut, 1995;
Haut, 1989). Atkinson and Haut (1995) document that the average pressures and contact
areas on the interior surfaces of the knee are slightly higher on the lateral versus the
medial patellar facet. The contact area is highest at 90° knee ﬂexion (Atkinson and Haut,

2001 and Atkinson and Haut, 2001). Other studies also document that the patello-femoral

l3

contact pressures and areas are reduced with addition of a padded interface (Haut, 1989;
Atkinson et 51., 1997; Atkinson and Haut 2001).

Viano et al. (1978) investigated the effect of loading the knee through different
anatomical regions. Frontal impacts were applied to the knee (including the patella and
tibial tuberocity) and the tibia (near and below the tibial tuberocity) (Figure 1.8). The
authors document PCL rupture and avulsion for knee impacts at approximately 7 kN of
load. Tibial impacts produce fractures of the tibia and ﬁbula at approximately 5 kN.
Additional experiments on isolated knee joints in the Viano et al. study showed that PCL
rupture occurs at approximately 2.25 cm of posterior tibial displacement and

approximately 2.5 kN of load.

PATELLAR
DISPLACEMENT
AND WEDGING

 

  
 

 

tieiAL/FIBULAH
DISPLACEMENT

 

Figure 1.8. Frontal impact to a human knee joint ﬂexed 90° (Viano 1978).

The effect of an axial tibia load has recently been studied using two impact
protocols. Constrained experiments on the TF joint do not allow anterior-posterior (AP)
or medial-lateral (ML) motion of the tibia with respect to the femur. These experiments

result in fracture to the medial and lateral tibial plateau, medial femoral condyle and

14

femoral notch at 8.0 :1.8 kN (Banglmaier et al. 1999). Other previous studies (Hirsch
and Sullivan, 1965 and Kennedy and Baily, 1968) also constrained the TF joint and
document a similar mean fracture load of approximately 8 kN. Recently a study by
Jayaraman et al. (2001) documents the effect of TF joint constraint on knees at various
ﬂexion angles (60-120°). An interesting result occurred for unconstrained knees at all
ﬂexion angles, instead of bone fracture 14 of 16 knees failed by ACL ruptures. The
authors hypothesize that this result was due to an inherent tilt of 8-15° in the tibial plateau
(Figure 1.9). ACL rupture occurs at a load of 4.9 i 2.1 kN which is statistically less than
the load to cause fracture of bone in the constrained knees. The study also documents that
during unconstrained TF compression the femur moves medially with respect to the tibia
and the tibia rotates internally. Additionally for constrained joints the load to prevent
motion of the femur relative to the tibia is 1.2 :05 kN. This may be the tensile load in the

PCL prior to rupture in unconstrained experiments.

    
   

QUADRICEPS

15 DEGREES

ANTERIOR
SHIFT

   

Figure 1.9. Tilt of the tibial plateau and resultant anterior displacement of the tibia.

15

These investigations have led to the following generalizations to predict traumatic
injury to the knee. Axial femoral loading through the patella will most often produce
patella fracture at approximately 6 kN for rigid impacts. However for padded impacts the
load tolerance increases and the most commonly injured site is the femoral condyle
(Atkinson 1997). Axial loading of the tibia will produce fracture in the medial and lateral
tibial plateau, femoral condyle and femoral notch at approximately 8 kN, when the femur
is constrained from translating relative to the tibia. For unconstrained tibio-femoral
loading the most common injury is mid-substance tear of the ACL at approximately 6
kN. Shear loading of 2.5 kN between the tibia and femur results in PCL tears or avulsions
at approximately 2.25 cm of relative displacement.

Knee fracture is easily documented during experiments and by clinicians after real
world automotive accidents. Therefore a knee injury criterion based on fracture is the
most obvious choice. However, many of these fracture studies also document other
injuries that are not as obvious as gross fracture. Additionally, lower extremity trauma
causes acute pain followed by a chronic disease process that can lead to an end-stage
disease such as osteoarthritis (States 1986). Subtle damage to cartilage and subchondral
bone can occur without radiographic evidence of bone fracture (Pritsch et al. 1984).
Recent studies have focused on identifying occult bone trauma (Figure 1.10) and relating
it to clinical ﬁndings. These radiographically occult injuries to the bone, otherwise
referred to as occult fractures or bone bruises, may account for patient pain (Kapelov
1993). However, a direct association between mechanical insult and disease has been
difﬁcult because visible evidence of the disease may not show up for years (Wright

1990).

16

 

Figure 1.10. Bone bruise in the lateral femoral condyle.
Subfracture Experiments

Biomechanical impact studies, attempting to document subfracture injuries, have
not had the luxury of being able to rely on human cadaver experiments. Since the nature
of many of these injuries is their chronic progression, it is necessary to use an in vivo
animal model that can document these long-term changes. This animal data must then be
combined with human cadaver experiments to create useful information about automotive
tolerances to chronic disease. Additional problems arise because of the controversy over
the mechanisms of the chronic disease process. Recent studies have shown that impact
can cause damage to both the cartilage and bone without gross fracture. Long-term
studies with a rabbit model have further shown decreasing mechanical properties of
cartilage and increased thickness of the subchondral bone over time (Newberry et al.
1996, Ewers et al. 2000). These studies have gone on to explore ways that the rate of
disease progression is increased or decreased, by exercise or pharmaceutical treatment, so

that early diagnosis of the initial injuries could prevent or delay an end-stage disease.

Histological methods are a common way of documenting subfracture injuries such
as occult bone microcracking and cartilage ﬁssuring (Figure 1.11). These studies use
semi—quantitative scoring to analyze the condition of tissue as a result of impact and
chronic degradation. Occult injuries such as subchondral bone nricrocracks are
documented in a number of impact studies with and without gross fracture of bone
(Ewers et al., 2000; Newberry and Haut, 1996; Haut and Atkinson, 1995). High rate
impacts cause more microcracks than low rate impacts (Ewers et al., 2000). This
microcracking is hypothesized to cause chronic subchondral bone thickening and
remodeling post impact.

Cartilage fissures have been documented as a result from blunt impact. The
number of ﬁssures and their depth are related to the applied impact load. They have also
been shown to increase over time post impact in a rabbit model (Haut 1989). Haut
documents that at fracture up to 60% of the patello-femoral contact area exceeds 25 MPa.

This level of pressure was previously shown to cause cartilage ﬁssures in an in vitro

explant model (Repo and Finlay 1977).

 

Figure 1.11. Surface ﬁssure of the AC (A) and an occult rnicrocrack at the tideka (B).
The mechanical properties of cartilage have also been documented through the

use of indentation testing with a cylindrical indenter (Figure 1.12). The instantaneous

18

(Gu) and relaxed (Gr) shear moduli are obtained from this method. Recent studies have
shown that within 12 months post impact cartilage undergoes signiﬁcant mechanical

softening (Newberry et al. 1997, Ewers et al. 2000).

+ P“)

    

CARTILAGE

‘ ...ns, “MW «Av-x M~Wb\.ﬁﬁ—A¢~L

Figure 1.12. Indentation test to measure the mechanical properties of cartilage.

These investigations have led to the following generalizations about subfracture
injuries. Impacts on human knees can result in occult microcracks of the subchondral
bone before gross bone fracture. These microcracks may be precursors to gross bone
fracture since they occur in the same regions where fractures typically occur. Deformable
interfaces protect against gross fracture but microfractures still occur for stiffer interfaces.
In animal experiments both ﬁssures of the cartilage and microcracks of the underlying
bone have been shown initially with a subfracture impact. At one year the stiffness and
mechanical properties of the cartilage decrease and the subchondral bone thickness

increases.

References

Atkinson PJ, Haut RC. (1995) Subfracture insult to the human cadaver patellofemoral
joint produces occult injury. J Orthop Res 13: 936-944.

Atkinson PJ, Garcia JJ, Altiero NJ, Haut RC. (1997) The Inﬂuence of Impact Interface on
Human Knee Injury: Implications for Instrument Panel Design and the Lower
Extremity Injury Criterion, Stapp Conf Proc 41: 167-180.

Atkinson T, Atkinson P. (2000) Knee injuries in motor vehicle collisions: A study of the
National Accident Sampling System database for the years 1979-1995. Accid Anal
Prev 32 (6): 786.

Atkinson P, Haut R.C. (2001a) Impact responses of the flexed human knee using a
deformable impact interface. J Biomech Egr 123 (3): 205-211.

Atkinson PJ, Haut RC. (2001b) Injuries produced by blunt trauma to the human
patellofemoral joint vary with flexion angle of the knee. J Orthop Res 19(5): 827-833.

Banglmaier R, Dvoracek-Driksna D, Oniang’o T, Haut R. Axial compressive load
response of the 90° ﬂexed human tibiofemoral joint. Stapp Conf Proc 42: 127-140.

Bourrett P, Corbelli S, Cavallero C. (1977) Injury Agents and Impact Mechanisms in
Frontal Crashes in about 350 Field Accidents. Stapp Conf Proc 21: 215-258.

Chapchal G. (1978) Posttraumatic Osteoartritis After Injury of the Knee and Hip Joint.
Reconstr Surg Trauma 16: 87-94.

Cooke FW, Nagel DA. (1991) Biomechanical Analysis of Knee Impact. Stapp Conf Proc
13: 117-133.

Davis M, Ettinger W, Neuhaus J, Cho S, Hauck W. (1989) The Association of Knee
Injury and Obesity with Unilateral and Bilateral Osteoarthritis of the Knee. Am J
Epidemiology 130: 278-288.

Dieppe P, Cushnaghan J, McAlindon T. (1992) Epidemiology, Clinical Course and
Outcome of Knee Osteoarthritis. Articular Cart and Osteoarthritis Raven Press Ltd,
NY: 617-627.

Ewers BJ, Jayaraman V, Banglmaier R, Haut R. (2000) The effect of loading rate on the
degree of acute injury and chronic conditions in the knee after blunt impact. Stapp
Conf Proc 44: 299-314.

Fildes B, Lane L, Vulcan P, Seyer K. (1997) Lower Limb Injuries to Passenger Car
Occupants. Accid Anal Prev 29: 789-795.

20

Hartemann F, Thomas C, Henry C, Foret-Bruno JY, Faverjon G, Tarriere C, Got C, Patel
A. (1977) Belted or Not Belted: The Only Difference Between Two Matched
Samples of 200 Car Occupants. Stapp Conf Proc 21: 97-150.

Haut RC. (1989) Contact Pressures in the Patello-Femoral Joint During Impact Loading
on the Human Flexed Knee. J Orthop Res 7: 272-280.

Haut RC, Atkinson PJ. (1995) Insult to the Human Cadaver Patello-Femoral Joint:
Effects of Age on Fracture Tolerance and Occult Injury. Stapp Car Crash J 39: 281-
294.

Hayashi S, Choi HY, Levin RS, Yang KH, King A1. (1996) Experimental and Analytical
Study of Knee Fracture Mechanisms in a Frontal Knee Impact. Stapp Car Crash J 40:
160-171.

Hendrie D, Rosman D, Harris A. (1994) Hospital inpatient costs resulting from road
crashes in Western Australia. Aust J Public Health 18: 380—388.

Hirsch G, Sullivan L. (1965) Experimental Knee Joint Fractures. Acta Orthop Scand 36:
391-399.

Hodgson RJ, Carpenter TA, Hall ID. (1992) Magnetic Resonance Imaging of
Osteoarthritis. Articular Cart and Osteoarthritis Raven Press Ltd, NY: 629-667.

Jayaraman VM, Sevensma ET, Kitagawa M, Haut RC. (2001) Effects of Anterior-
Posterior Constraint on Injury Patterns in the Human Knee During Tibial-Femoral
Joint Loading from Axial Forces Through the Tibia. Stapp Car Crash J 45: 449-468.

Kapelov R, Teresi L, Bradley WG, Bucciarelli NR, Murakami DM, Mullin WJ, Jordan
JE. (1993) Bone Contusions of the Knee; Increased Lesion Detection with Fast Spin-
echo MR Imaging with Spectroscope Fat Saturation. Radiology 189: 901-904.

Kennedy, J .C., Bailey, W.H., 1968. Experimental tibial-plateau fractures: studies of the
mechanism and a classiﬁcation. J Bone Jt Surg 50(A): 1522-1534.

Luchter S, Walz MC. (1995) Long-term Consequences of Head Injury. J Neurotrauma
12: 281-298.

MacKenzie E, Siegel J, Shapiro S. (1988) Functional Recovery and Medical Costs of
Trauma: An Analysis by Type and Severity of Injury. J Trauma 28: 281-298.

Melvin J, Stalnaker R, Alem N, Benson J, Mohan D. (1975) Impact response and
tolerance of the lower extremities. Stapp Conf Proc 19: 543-559.

Miller TR, Martin PG, Crandell JR. (1995) Cost of Lower Limb Injuries in Highway
Crashes. Proc ICPLEI: 47-57.

21

Miller M, Osborne J, Gordon W, Hinkin D, Brinker M. (1998) The Natural History of
Bone Bruises: A Prospective Study of Magnetic Resonance Imaging: Detected
Trabecular Microfractures in Patients with Isolated Medial Collateral Ligament
Injuries. Am J Sports Med 26: 15-19.

Nagel D, States J. (1977) Dashboard and bumper knee-will arthritis develop? AAAM 21:
272-278.

Newberry WN, Haut RC. (1996) The Effects of Subfracture Impact Loading on the
Patello-Femoral Joint in a Rabbit Model. Stapp Car Crash J 40: 149-159.

Patrick L, Kroell C, Mertz H. (1965) Forces on the human body in simulated crashes.
Stapp Conf Proc 9: 237-259.

Powell W, Ojala S, Advani S, Martin R. (1975) Cadaver femur responses to longitudinal
impacts. Stapp Conf Proc 19: 561-579.

Pritsch M, Comba D, Frank G, Horoszowski H. (1984) Articular Cartilage Fractures of
the Knee. J Sports Med 24: 299-302.

Repo R, Finlay J, (1977) Survival of Articular Cartilage After Controlled Impact. J Bone
Jt Surg 59A: 1068-1076.

States JD. (1986) Adult Occupant Injuries of the Lower Limb. Proc Symp Biomech: 97-
107.

Viano DC, Culver CC, Haut RC, Melvin JW, Bender M, Culver RH, Levine RS. (1978)
Bolster Impacts to the Knee and Tibia of Human Cadavers and an Anthropomorphic
Dummy. Stapp Conf Proc 22: 403-428.

Viano DC, Stalnaker RL. (1980) Mechanisms of Femoral Fracture. J Biomech 13: 701-
715.

Vellet AD, Marks PH, Fowler PJ, Munro TG. (1991) Occult Post-Traumatic
Osteochondral Lesions of the Knee: Prevalence, Classiﬁcation and Short-term
Sequelae Evaluated with MR Imaging. Radiology 178: 271-276.

Volpin G, Dowd G, Stein H, Bently G. (1990) Degenerative Arthritis After Intra-articular
Fractures of the Knee: Long Term Results. J Bone Jt Surg 72: 634-638.

Wright V. (1990) Post-traumatic osteoarthritis- A medico-legal mineﬁeld. J Rheum 29:
474-478.

22

CHAPTER TWO

THE EFFECT or IMPACT ANGLE ON KNEE TOLERENCE To

RIGID IMPACTS

ABSTRACT

While the number of deaths from vehicle accidents is declining, probably because
of mandatory seat belt laws and air bags, a high frequency of lower extremity injuries
from frontal crashes still occurs. For the years 1979-1995 the National Accident
Sampling System (NASS) indicates that knee injuries (AIS 1—4) occur in approximately
10% of cases. Patella and femur fractures are the most frequent knee injuries. Current
literature suggests that knee fractures occur in seated cadavers for knee impact forces of
7.3 to 21.0 kN. Experimental data shown in a study by Melvin et al. (1975) further
suggests that fracture tolerance of the knee may be reduced for an impact directed
obliquely to the axis of the femur. The current study hypothesized that the patella is more
vulnerable to fracture from an oblique versus an axial impact (directed along the femoral
axis), and that the fracture pattern would vary with impact direction. Isolated, 90° ﬂexed,
paired human knee joints (73 i 16.9 years) were impacted at sequentially higher loads
either axially or obliquely from the medial aspect with a rigid interface on the patella.
The peak load at fracture for each case was recorded, and a detailed description of the
fracture pattern was documented. For axial impacts all nine knees failed by linear and
comminuted patella fracture with an average peak load of 5.9 i 1.4 kN. Seven of nine
obliquely impacted knees also failed by linear and comrninuted patella fracture with a
signiﬁcantly lower peak load of 3.5 i 1.4 kN. The peak load data from fracture

experiments for all knees showed a strong correlation with age and direction of the

23

impact. Additionally, the fracture pattern for the axial impacts was generally oriented
along a horizontal plane on the patella, while the fractures for oblique impacts were
generally oriented along a vertical plane. In two oblique experiments patella fracture did
not occur, as the patella dislocated at a load of 3.4 i 0.2 kN. In one of these cases the
medial aspect of the patello—femoral joint capsule was visibly torn, and in the other case
surface damage was noted on the articular cartilage covering the lateral femoral condyle.
In addition to the acute injuries described in this study, these data may suggest a potential
for chronic diseases of the knee in cases where similar injuries are produced. Clinical
studies have shown an increased risk of osteoarthritis in patients suffering patella
fractures and damage to joint cartilage. Also, acute dislocation of the patella may cause
soft tissue injury of the knee leading to chronic mal-tracking of the patella. These data
may be particularly relevant in cases where occupants sit with abducted lower limbs prior

to a frontal crash.

24

INTRODUCTION

Traumatic injury is the third leading cause of death in the US. Only heart disease
and cancer exceed the frequency of traumatic death each year [Rice et al., 1989]. Motor
vehicle injuries rank second to cancer in total societal cost, accounting for 3.2% of annual
medical costs in the US. [Miller et al., 1998]. Overall, while the number of deaths from
vehicle accidents is declining, probably because of mandatory seat belt laws and air bags,
3 high frequency of lower extremity injuries still occurs [Dischinger et al., 1992]. A
recent study of the 1979-1995 National Accident Sampling System (NASS) indicates that
knee injuries constitute approximately 10% of all injuries (AIS, abbreviated injury scale,
1-4) recorded each year in frontal crashes [Atkinson & Atkinson, 2000]. This study
further suggests that the knee impact scenario for the years 1979-1995 remained
relatively constant, as the knee injury rates showed little variation from year to year. For
the years 1993-1995 approximately 26% of the total knee injuries were categorized as
AIS 2 or 3 type injuries (bone fracture and severe soft tissue trauma). Additionally these
more descriptive knee injury codes show that patella and femur fractures are the most
frequent injuries.

The automotive trauma literature suggests that femur force does a good job
separating human cadaver injury from non-injury of the knee-thigh-hip for axial (directed
along the axis of the femur) impacts onto the knee [Morgan et al., 1990]. Early
experiments using seated unembalmed human cadavers document the blunt impact
response of the knee-thigh-hip complex using rigid and lightly padded interfaces. These

studies show that the fracture force ranges from approximately 7.3 kN to approximately

25

21.0 kN [Melvin et al., 1975; Patrick et al., 1965; Powell et al., 1975]. A large
percentage of the fractures represented in this database occur in the supracondylar region
of the femur or in the patella. Recent experiments, using the isolated knee joint
preparation, also document supracondylar femur fractures and patella fractures following
repetitive, axial impact loads on the anterior surface of the patella with rigid impact
interfaces [Atkinson & Haut, 2001b]. In these studies loads producing bone fracture
range from 4.3 to 5.8 kN for the isolated human knee joint ﬂexed 60, 90 or 120 degrees
(angle between femur and tibia, where 180° refers to a straight, fully extended leg). For
each ﬂexion angle of the knee the study documents horizontal (a medial-lateral oriented
fracture plane) fractures of the patella. In a few cases comminuted patella fractures are
also noted with 90 and 120 degrees of knee ﬂexion. As was typical in the 1970's studies
that established impact tolerance forces for the human knee-thigh-hip during axial blunt
impact to the knee, the Atkinson et al. study was conducted with aged cadaver specimens
[70.4114 years]. While the driving population is likely less aged than the above group of
specimens, another study using the isolated human knee preparation and a rigid impact
interface documents that fracture loads do not signiﬁcantly vary with specimen age
[Atkinson et al., 19983].

Chronic experiments using a small animal model, the Flemish Giant rabbit, also
show that axial (directed along the femoral axis) blunt impacts on the knee are less
damaging to retropatellar cartilage than impacts directed obliquely (from the medial
aspect with respect to the femoral axis) on the patella [Ewers et al., 2002]. The study
using this animal model motivated a question about the role of impact direction on

fracture tolerance of the human knee. One early study using the seated human cadaver

26

also investigated medial oblique impacts on the knee versus axial impacts [Powell et al.,
1975]. In this study, while the oblique impacts generated strains on the femur suggesting
a curved beam effect and the axial impacts suggested a beam-column effect, the authors
did not speciﬁcally address how impact direction affects the femur or patella fracture
load. On the other hand, Melvin et al., 1975 investigated the impact fracture responses of
the human cadaver knee-thigh-hip complex for axial and oblique (medial-directed)
impacts on the knee with a rigid impact interface covered with a 1" thick foam pad
(Figure 2.1). Fractures were noted at 18.4 i 2.0 kN (n=5) for axial impacts on the knee.
In contrast, patella fractures were noted in 2 cases at 8.1 and 10.5 kN for oblique impacts

(11° of hip abduction).

    

 

 

 

Femoral condyle
Patella

Oblique Imp”Ct Axial Impact

Figure 2.1. Axial and oblique impact loading direction of the knee-thigh-hip
complex from Melvin et al., 1975.

27

These data may suggest a reduced fracture tolerance of the human patella for oblique
versus axial impacts on the knee. Melvin et al., 1975 furthermore show a "typical"
supracondylar femur fracture and a fractured patella. The patella fracture plane is
interestingly vertical (inferior to superior oriented) rather than horizontal (medial to
lateral oriented), as has been previously shown in studies using the isolated knee
preparation for axial impacts on the patella [Atkinson et al., 1997, Atkinson & Haut,
2001b].

Therefore the following questions were asked: 1) Is the patella more vulnerable to
fracture from an oblique (medial directed with respect to the femoral axis) versus an axial
(directed along the femoral axis) impact? 2) Does the fracture pattern depend on
direction of the blunt impact on the knee? The objectives of the current study were
therefore to document the peak loads for fracture and fracture patterns on the human
patella for a rigid interface impacting axially and obliquely (from the medial direction)

with respect to the axis of the femur.

METHODS

Blunt impact was delivered to pairs of knee joints from nine human cadavers aged
73 :t 16.9 years. The limbs were procured from university sources (see Acknowledgment)
and stored at -20°C until testing. The joints were selected from donors with no known
knee injuries or signs of surgical intervention during a postmortem evaluation. Twenty-
four hours prior to testing, the joints were thawed to room temperature. The preparations
were sectioned 15 cm superiorly and inferiorly to the knee joint. Superﬁcial muscle

tissues were excised from each preparation leaving the articular joint capsule and the

28

quadriceps tendon-patella-patellar ligament complex intact (Figure 2.2). The femur was
cleaned with alcohol and potted in a 6.3 cm diameter, 10 cm deep cylindrical aluminum
sleeve with room temperature curing epoxy (Fibre Strand #6371, Martin Senour Co.,
Cleveland, OH). This allowed approximately 5 cm of the femur to be exposed beyond the
potting material. The test specimens were mounted in the ﬁxture with 90° of knee
ﬂexion. The angle was established by aligning the femur and tibia visually with a 90°
square tool. A special clamp was designed to attach the quadriceps tendon to a pneumatic
cylinder (Model #D-48349-A-6, Bimba Manufacturing Co., Monee, IL). The pneumatic
cylinder was pressurized prior to each experiment to generate 1.3 kN of force in the
quadriceps tendon. The quadriceps tendon force was applied in both the axial and oblique
impacts to simulate an in vivo muscle force, and help keep the patella in the femoral
groove for oblique impacts. One knee from each pair was randomly selected for axial
(directed along the femoral axis) impact, while the contralateral knee from each pair was
impacted at an oblique angle (15° or 30°) medially with respect to the femoral axis
(Figure 2.3). A servo-controlled hydraulic testing machine (Model 1331 retroﬁtted with
8500 plus digital electronics, Instron Corp., Canton, MA) was programmed to load the
anterior surface of the patella with a haversine waveform that generated a peak load in 50
ms. The 50 ms time to peak approximated the time to reach peak load (20 to 60 ms)
documented for a Hybrid III midsize male dummy during a typical automotive crash
simulation [Rupp et al., 2002]. A repetitive, increasing load protocol was followed where
test 1 corresponded to an input load of 1 kN and tests 2, 3, 4, 5 and 6 corresponded to
programmed input loads of 3, 5, 7, 9, and 11 kN, respectively. Specimen # 03-328L test 1

was carried out with an input load of 7 kN, due to a mechanical problem. Test 1 on the

29

 

<——Load Cell ‘ 4— Load Cell

[j—‘lmpactor Joint capsul '- a Irnpactor
a-fir

 

 

 

 

   

Patella A l ,
. I FILE ng e setting Q.
(21:33:11.5 / Tibia ﬁxtu - ¥5 ~ Patella

 
   

clamp Adjustable /] I : /-Femur
Pulley to height clamp Quadn'ceps - \ I Q dril
' for tibia tendon clam &"ﬁi~: ua ceps
direct load p tendon
from a Femur potted in a 71%
p23: 23:? cylindrical sleeve 0

 

 

Figure 2.2. Isolated human knee joint ﬂexed
90° with impact loading on the patella. A
tension of 1.3 kN was applied to the
quadriceps tendon immediately prior to the
impact load.

axis.

contra-lateral limb, #03-328R, was performed with an input of 5 kN and test 2 at 7 kN.
All impacts were delivered with a rigid interface (aluminum, 5 cm diameter) that was
attached to the hydraulic actuator of the testing machine. A 2500 lb. (11.1 kN) load
transducer (Model #10101a-2500, Instron Corp. Canton, MA) was attached behind the
interface. After each impact in the repetitive series of experiments, the specimen was
examined for gross fracture of bone or dislocation of the patella by visual inspection and
palpation of the retropatellar surface under the quadriceps tendon. The anterior surface of
the patella was also visually inspected, and the location of the patella at the end of the
loading was documented. Experimental data (load and displacement) from the testing
machine were collected at 1000 Hz and recorded on a personal computer with a 16—bit
analog/digital board (model DAS 1600; Computer Boards, Mansﬁeld, MA, U.S.A.).

Prior to each sequential test, pressure sensitive ﬁlm (Prescale, Fuji Film Ltd., Tokyo,
Japan) was inserted into the patello-femoral joint to measure the magnitude and

distribution of contact pressures generated in the joint during impact. Low (0-10 MPa)

30

Figure 2.3. Oblique orientation of the knee
joint preparation with respect to the femoral

and medium (10-50 MPa) range pressure ﬁlms were stacked together and sealed between
two sheets of polyethylene (0.04 mm thick) to prevent exposure of the ﬁlm to body ﬂuids
[Atkinson et al., 1998b]. The ﬁlm packets were inserted into the patello-femoral joint
under the quadriceps tendon, without disturbing the medial and lateral aspects of the
articular capsule. A new pressure ﬁlm packet was placed in the joint immediately prior to
each test in the sequence. A ﬁlm packet was also placed on the anterior surface of the
patella to record the magnitude and distribution of contact pressures between the rigid
interface and the knee. The ﬁlm was digitally scanned (ScanMaker E6, Microtek
International Inc., Redondo Beach, CA), and the pressure distributions were quantiﬁed
using commercial software (PhotoStyler, Aldus Corporation, Seattle, WA). The image
resolution was set at 150 dpi, and the ﬁlm data was converted to a gray scale (NIH
Image, version 1.6). The gray scale values were converted to pressures using a previously
established methodology [Atkinson et al., 1998b]. Brieﬂy, calibration tests were
performed on the low and medium stacked ﬁlm packets while sandwiched between two
polished stainless steel plates. A haversine, displacement-controlled, waveform was used
to generate calibration loads in approximately 50 ms, using the servo-hydraulic testing
machine. This provided a dynamic calibration for the ﬁlm packets. As described by
Atkinson et al. (1998b), the stacking order of the ﬁlm and the rate of loading are
important factors in the measurement of contact pressures using this ﬁlm.

Each joint was grossly examined following the failure experiment. The medial and lateral
capsules were carefully probed in an attempt to view any gross ruptures of the capsule,
especially adjacent to the patella. Injuries were photographed. Thereafter, the medial and

lateral capsules were carefully cut to fully expose the retropatellar surface and femoral

31

condyles. These articular surfaces were wiped with India ink in order to help visualize
surface lesions on the articular cartilage. Photographs recorded any surface defects and
damage.

The means and standard deviations for peak load, the corresponding displacement
and time to peak load were documented for each experiment, and the pressure ﬁlm
contact area and average pressure were determined for each impact direction in the
fracture experiments. Statistical comparisons of these values for the axial and oblique
impact directions were performed with paired Student’s t-tests. Statistical signiﬁcance
was determined for p < 0.05 in these experiments. Additionally, in the sub-fracture
experiments joint stiffness was computed as the slope of a linear regression line through
the load-displacement data for three load regions in tests 1, 2 and 3 on each specimen.
The stiffness of the test ﬁxture was also documented by loading it directly in the testing
machine. The ﬁxture stiffness was documented to be 5 tol4 times higher than that of the
joint for oblique and axial impacts, respectively. In post-processing of the load deﬂection
data, the ﬁxture deﬂection at each load was subtracted from the data to determine the
stiffness of the knee joint itself. The means and standard deviations of the 3 load-range
stiffnesses were determined for each experiment. These data were statistically compared
with a one-way ANOVA and Student-Newman-Keuls post-hoe tests. Regression analyses
(univariate and multivariate) were used to determine the relationships, if any, between
peak loads for bone fracture and specimen age.

RESULTS
The output load-time response curves for the axial and oblique experiments were

skewed haversines with a longer unloading than loading time (Figure 2.4). The time to

32

peak was 54 t 8 ms for all experiments. While the time to peak load was typically less
for fracture experiments (47 i 8 ms for axial, 48 i 9 ms for oblique) versus non-fracture
experiments (53 1- 4 ms for axial, 60 i 7 ms for oblique), there were no statistically
signiﬁcant differences in the time to peak load between axial and oblique experiments for

the fracture or the non-fracture experiments (Tables 2.1 and 2.2).

8- 3-

 

 

 

 

 

 

 

 

 

 

A B

; ( ) ( ) —Test1
E ‘ :N Test 2 E 6 - Test 3
:6 ;.' '. T9313 ‘5 —-Test4
W ,AI“: T6514 U
3 4 ' ll 34‘

l . - - - -Test5
0" r. CI.
0 l . 0
«I 1 cu \
n. j ._ a. .
E 2 - I. E2 - 1 \

' —

o I I I I I I I I I o A T T I l l I I I
0 50 100150 200 250300350400450 0 50 IUU 13315::

Time (ms) Time (ms)

Figure 2.4. Representative load-time curves for sequential impacts on a specimen (03-
017) for axial (A) and oblique (B) directed loadings.

The impact responses of the isolated knee in the oblique experiments generally
indicated more specimen compliance than for the axial impacts. The overall shapes of the
load-displacement responses for axial and oblique experiments were nonlinear (Figure
2.5). The impact responses did not vary significantly between sequential tests on the
same specimen, except for an increase in peak load. The repeatability of sequential tests
on the same limb was documented by a comparison of the average slopes of these curves
for various load ranges (Figure 2.6). The response curves were divided into three load
ranges that corresponded to the loads developed in tests 1, 2 and 3 on each specimen. In

the lowest load range (0 — 630 N) the average slope (patello-femoral joint stiffness) for

test 1 (433 i 164 N/mm) was not different from that generated in test 2 (490 t 198
N/mm) or test 3 (418 i 176 N/mm) for the axial impacts on all specimens (Figure 2.7).
Similarly, the slope of the response curves in the second load range (630 N - 2440 N)

computed from test 2 (893 i 275 N/mm) was not different from that generated in test 3

(1015 1 224 N/mm) for the specimens.

 

 

 

 

8 - 8 7
. (A) (B)
—-Test1 —Test1
92.56 - —Teet2 g6 - —Test2
—-Test4
34 -——Test4 3 4 -
‘6 ----- TestS ‘5 M
G a
Q a
£2 ' g 2 J /
0 I I *1 0 I I 1
o 10 20 30 o 10 20 30
Actuator Displment (mu) Actuator Dlsplaeement (mm)

Figure 2.5: Representative load-actuator displacement curves for sequential impacts on a
specimen (03-017) for axial (A) and oblique (B) directed experiments. In this experiment
patellar fracture occurred for the axial loading while on the contralateral limb a medial
oblique impact dislocated the patella.

34

Table 2.1: Biomechanical data for axial impacts. The fracture experiment is the last entry for each specimen.

 

 

 

 

 

 

 

 

 

 

 

 

Specimen Sex lnpact Peak Load Displmment Time to .
ID (Age) Nurrber (kN) (mm) Peak (ms) "W W‘s

(2-687Fl Male (94) 1 0.74 224 56 Femoral Cornyle’Fracture
2 244 523 53 Linear Patella Fracture
3 4.01 8.76 56
4 4.41 7.87 40

02-698L Female (88) 1 0.74 224 57 Corrm'nuted Patella Fracture
2 259 5.18 51
3 3.94 7.02 48
4 5.49 11.22 51

03-035R Female (85) 1 0.76 211 56 Corminuted Patella Fracture
2 269 4.71 51
3 4.38 7.57 61
4 4.25 5.31 38

03-048L Female (82) 1 0.85 2.84 54 Linear Patella Fracture
2 259 383 51
3 4.33 7.11 53
4 5.05 10.44 52

02-686L Female (81) 1 0.85 1.53 54 Femoral Condyle Fracture
2 270 3.50 52 Corrm'nuted Patella Frazture
3 4.33 6.75 49
4 5.04 5.64 36

03-328L Female (64) 1 5.71 17.89 54 Linear Patella Fracture

03-017L Male (60) 1 0.63 284 64 Corrm’nuted Patella Fracture
2 269 5.52 51
3 4.90 7.04 53
4 696 8.22 50
5 7.29 1334 43

(Xi-383L Female (60) 1 0.76 293 57 Linear Patella Fractu'e
2 280 5.45 53
3 4.95 7.47 50
4 5.79 10.31 53
5 7.76 14.14 56

(Xi-3698 Male (43) 1 0.82 1.79 57 Linear Patella Fracture
2 276 343 53
3 4.89 4.43 47
4 6.24 5.19 54
5 7.26 5.90 46
6 7.99 8.10 53

Fracture Average 5.89' 10.44 47
[Fracture Standard Deviation 1.43 4.18 8

 

 

* Signiﬁcantly different than oblique.

35

Table 2.2: Biomechanical data for oblique impacts. The fracture or dislocation experiment is the last entry for each

 

 

 

 

 

 

 

 

 

 

 

specimen.
Specimen Sex Impact Peak Load Displacement Time to
ID (Age) Number (kN) (mm) 2:: "My C°mmems
02-687L Male (94) 1 0.77 2.17 57 Linear Patella Fracture
2 2.34 5.67 57
3 2.80 6.14 37
02-698R Female (88) 1 0.55 2.82 64 Linear Patella Fracture
2 1.97 8.61 57
3 2.68 11.14 48
03-035L Female (85) 1 0.62 2.66 63 Cornminuted Patella Fracture
2 2.46 6.27 54
3 3.75 8.68 45
4 3.75 9.88 39 j
03-048R Female (82) 1 0.54 2.84 64 Linear Patella Fracture
2 1.90 8.74 60
3 2.02 8.14 38
02-686R Female (81) 1 0.56 2.95 66 Linear Patella Fracture
2 1.87 8.96 57
3 2.99 15.40 62
4 3.24 18.30 52
03.3233 Female (64) 1 1.77 10.16 44 Gross Dislocation of Patella
2 3.59 30.43 54 Tom Medial Capsule
03-017R Male (60) 1 0.55 3.09 66 Gross Dislocation of Patella
2 2.13 8.48 61 Cartilage Frssures on Laeral
3 3.11 12.00 51 Femoral Condyle
4 3.30 21.80 57
(KB-3838 Female (60) 1 0.44 3.10 68 Cornminuted Patella Frmture
(Angle at 15 degrees) 2 1.50 12.98 67
3 3.51 17.27 61
03-369L Male (43) 1 0.19 1.29 70 Cornminuted Patella Fracture
(Angle at 15 degrees) 2 2.24 5.81 57
3 4.15 7.68 57
4 5.73 10.13 56
5 6.26 20.48 49
"Feature Average 347' 13.05 48
Fracture Standard Deviation 1.36 5.57 9

 

 

 

* Signiﬁcantly different than axial.

36

6000- 3500-

 

 

 

 

 

 

 

 

 

 

 

 

A B
5000 4 ( ) 3000 . ( )
24000 1 y = 213.75x - 6.272 5 2500 4 Y = 16577" * “9295
V 2 Test 1 ; 2000 - R2 = 0.9911
g 3000 d R = 0.9884 8 15m .1
320007 .1 1000 .
1000 - 500 .
0 I 1 0 ' ' I
0 5 10 0 5 10 15
Displacement (mm) Displacement (mm)
6000 ‘ 3500 1
5000 - y = 721.9x - 1312.9 3000 «y = 287.37x - 287.29
24000 - R2 = 0.9831 2 2500 1 R2 = 0.9958
:3000 1 Test 2 z 2000 .
10 3 1500 «
32000 « _. 1000 ~
1000 - 500 -
o f “ 0 I I I
0 5 10 0 5 10 15
Displacement (m) Displacement (mm)
6000] 3500 -
A 5000 A 3000 «
a 4000 . T t3 5 2500 .
g 2000 . a 1500 -
..r y =1133-9x - 3131-5 3 1000 . y= 250.14x+211.83
1°00 ‘ R2 = 0.9832 500 - R2 = 0.9718
0 ' “1 0 f . .
0 5 1° 0 5 10 15
Displacement (mm) Displacement (mm)

Figure 2.6. Representative load versus displacement plots of a specimen (03-017) for
axial (A) and oblique (B) tests 1, 2 and 3 in their respective load ranges

The data for the ﬁrst and second load ranges (tests 1 and 2) helped validate, on average,
the repeatability of sequential load response data on the limbs. The oblique impacts also
had slopes that were similar for each load range in these repeated tests. For the lowest
range of loading (0 —440 N) the slopes were 223 i 56 N/mm, 219 i 27 N/mm and 199 i

51 N/mm in tests 1, 2 and 3 on these specimens, respectively. These data veriﬁed that the

37

2000-

 

 

 

 

 

 

 

I Test] (A) 2000“ I T5” (B)
1500— @ Test2 E Test2
E D “M3 E El Tests
\ * E
% 1000- E. 1000.
1o" 9
m _
5001 m
0.1 I 1
l 2 3
Load Range Load Range

Figure 2.7. Bar graph of linear regression slopes for axial (A) and oblique (B) tests 1, 2
and 3. * Signiﬁcantly different than oblique.

responses of the knee preparation did not vary, on average, between repeated tests on the
limbs in the low range of loading. For the second range of loading (440 N — 1840 N) the
slopes were 400 i 162 N/mm and 386 :1: 142 Mm for tests 2 and 3, respectively. A
comparison of the data for the axial and oblique impacts showed a signiﬁcant difference
in the average slopes of the response curves for load range 1 (p = 0.007) and load range 2
(p <0.001), with the axial responses having a higher slope (stiffer) than that for oblique
impacts. The slope of the highest load range from test 3 on the specimens was also
different (greater) for axial (above 2440 N) (1055 i 537 N/mm) versus oblique (above
1840 N) (415 :t 301 N/mm) impacts (p = 0.003). These data also showed that while the
load-displacement responses of the knee for axial and oblique impacts were nonlinear,
the slopes of these curves for the various load ranges indicated more dramatic changes at
in stiffness for the sequentially higher load experiments in the axial than for the oblique
impacts. The contact pressures within the patello—femoral joint were documented using

Fuji ﬁlm in this study for tests 3 and 4 on each specimen (Table 2.3).

38

Table 2.3: Pressure ﬁlm data from media] and lateral facets of the patella and the anterior
surface of the patella. * Signiﬁcantly different than medial.

 

 

 

 

 

 

Joint Orientation Test Number Average Pressure (MPa) Contact Area (mm42)
Medial Lateral Anterior Medial Lateral Anterior
Axial 3 16.1 (3.2) 18.1 (4.8) 16.4 (1.9) 115.6 (55.2) 209.8 (49.7) 182.3 (72.0)
Oblique 3 14.2 (6.3) 22.0 (3.7)* 18.2 (3.6) 19.2 (14.8) 219.5 (73.8)' 156.6 (68.7)
Axial 4 18.1(4.9) 21.3 (5.2) 18.5 (3.2) 145.7 (84.3) 242.1 (52.8) 250.3 (92.2)
Oblique 4 12.8 (7.6) 24.2 (30)" 19.1 (2.9) 14.6 (14.7) 302.4 (66.9)” 315.1 (164.9)I

 

 

 

 

 

 

 

 

 

Generally the pressure distributions were smooth with little or no apparent creasing of the
ﬁlm, which would indicate artiﬁcially high areas of contact pressure (Figure 2.8). The
average pressure on the lateral facet signiﬁcantly exceeded that on the medial facet for

the oblique impacts (test 3 p = 0.007 and test 4 p = 0.016), but no differences in average

Proximal
5" ' _ Proximal
‘ .. ,3, .
2 '7' 3; DJ 2 '4
Distal
(A) Distal (B)
- Proximal
Proxrmal
1r...
v— a" I
8 8 a '5
O H '0 0.)
2 ,3 0 ‘5
D' t 1 2 ..1
rs a
(C) (D)

 

Figure 2.8. Representative pressure distributions of a direct impact (03-048L, test 4) on the
anterior surface (A) and retropatellar surface (B) and an oblique impact (03-048R, test 3) on the
anterior surface (C) and retropatellar surface (D).

39

pressure were measured on the medial versus the lateral facet for the axial impacts
(Figure 2.9). Additionally, the lateral patello-femoral joint contact area was signiﬁcantly
higher than the media] contact area in test 3 (p < 0.001) and test 4 (p = 0.001) of the
oblique impacts and test 3 (p = 0.033) of the axial impacts, but was not signiﬁcant in test
4 (p = 0.051) for the axial experiment. There was also a signiﬁcant difference in the
medial contact areas for the axial and oblique impacts (test 3 p = 0.004 and test 4 p =

0.010). The peak loads carried by the lateral patellar facet, based on the pressure

30 - (A)
25 -
20 ' ITest 3 Medial
Test 3 Lateral
El Test 4 Medial
Test 4 Lateral

Average Pressure (MPa)
6‘1

 

 

Axial Oblique

(B)

I Test 3 Medial
I Test 3 Lateral
DTest 4 Medial
DTest 4 Lateral

Contact Area (mmz)

 

 

 

 

 

Axial Oblique
Figure 2.9. Bar graph of retropatellar surface average pressure (A) and contact area (B)
for tests 3 and 4.
* Signiﬁcantly different than medial side. # Signiﬁcantly different than oblique.

40

ﬁlm data, were signiﬁcantly higher than those on the medial facet for oblique (test 3 p =
0.011 and test 4 p=0.022) experiments, and axial (test 3 p< 0.001 and test 4 p=0.002)
experiments. The average pressure and contact area recorded from the anterior surface of
the patella were not different for axial versus oblique impacts.

As mentioned previously, gross injury of the knee joint occurred on or after test 4
for most specimens (Table 2.1 and 2.2). Fracture of the patella for axial impacts was
evident in each case (Figure 2.10A). In two cases the injury was associated with split
fracture of the femoral condyles. Linear fractures were oriented horizontally across the
patella in ﬁve axial directed experiments. Fracture of the patella resulted in signiﬁcant
comminution in four cases of axial impact (Figure 2.11A). The peak load and
corresponding displacement of the hydraulic actuator for axial impacts on the patella
were 5.9 :1: 1.4 kN and 10.4 t 4.2 mm, respectively (Table 2.1). Seven of nine obliquely

impacted knees exhibited fracture of the patella (Figure 2.11B). In contrast to the axially

' .tella

   

.._g
Proxrmal'

Proximal

rm- I
'8 d 'l)istal

  

(A) Femoral groove (3)

Figure 2.10. Representative photograph of patella fractures on paired human knees (02-
698). Horizontal oriented and comminuted fracture for axial impact (A), and vertical
oriented fracture for oblique impact (B).

41

 

 

 

 

 

 

 

 

 

Figure 2.11. Schematic representations of the fracture patterns on the retropatellar
surface for each axial (A) and oblique (B) directed impact. Fractures consisted of linear
and comminuted patterns.

Peak Load (RN)
0 -* [0 OD -b 01 O) \1 00

W

 

 

 

 

 

 

Axial Oblique

Figure 2.12. Bar graph of peak loads generated in
axial versus oblique impacts.
* Signiﬁcantly different than oblique.

directed impacts on the patella,
the fracture patterns for the
oblique impacts generally
appeared to be oriented more
vertically on the patella (Figure
2.10b). In three cases patella
fractures were comminuted. In
oblique experiments for cases of
patella fracture the peak load and
actuator displacement were 3.5 i
1.4 kN and 13.1 1- 5. 6 mm,
respectively (Table 2.2). The peak

loads generated in axial and

oblique fracture experiments were signiﬁcantly different (p = 0.008) (Figure 2.12).

Patella fracture did not occur in two oblique experiments. Rather, the patella dislocated

A

2

at a load of 3.4 :1.- 0.2 kN and an actuator displacement of 26.1 i 6.1 mm. In one
specimen (03-328R) the retinaculum (articular capsule) was separated at the medial
border of the patella (Figure 2.13A). During this traumatic event the hydraulic actuator
displaced approximately 30 mm, compared to 13.1 :t 5.6 mm in fracture experiments, as
the patella was dislocated laterally out of the femoral groove between the condyles. The
dislocation was visually evident during the experiment. In a second specimen (03-017)
the patella was again dislocated laterally without a visible fracture. The actuator
displacement at peak load in this second case was approximately 22 mm (Figure 2.5B), as
the patella was dislocated laterally from the femoral groove. In this case of patellar
dislocation a surface lesion was noted on the articular cartilage covering the lateral
femoral condyle (Figure 2.13B). This articular surface could not be entirely viewed prior
to the experiment because of the intact nature of the joint capsule. Because of the pristine
appearance of this cartilaginous surface, the injury was suspected to be due to abnormal
loading on the cartilage during acute dislocation of the patella.

Patell - (A)

    
  

Fissures 03)

Torn medial

 

Figure 2.13. Specimen 03-328 showing atom medial retinaculum from an oblique
impact (A). Specimen 03-017 showing an articular cartilage surface lesion from an
oblique impact (B).

43

The current experiments were conducted on specimens 43 to 94 years of age. Because of
the large variation in specimen age, the peak load data were examined for a dependence
on specimen age (Figure 2.14). Univariate regression analyses of peak load for fracture
cases versus specimen age indicated a signiﬁcant correlation for the axial and oblique
(without 15° impact data) experimental data. The correlation coefﬁcient (R2) was lower
for oblique (0.436) than for the axial impacts (0.823). A multivariate linear regression
analysis of these data, using age (43-94 years) and direction (0 = axial and l = 30°
oblique) generated the following expression (where age and direction were signiﬁcant
predictors, p < 0.001 and p = 0.002, respectively):
Peak load = 11536.6 — 77.4*Age — 1520.7*Direction (2.1)

This more general expression also ﬁt the experimental data well (Figure 2.14). The
univaritate and multivariate predictions were based on only the 30° oblique data, but the
equations appeared to predict the 15° data for the youngest specimen within less than

10% of the peak load for fracture.

 

 

 

9000 ' y = -76.936x + 11504

8000‘ ' ‘ . ,_ - R2=0.8228 ' AxialData
2 7000 _ “3... . 0 Oblique Data
3 O Oblique (15 deg)
g Axial (nultivariate model)
'3 —— Oblique (multivariate model)
a ----- Axial (univariate model)

2000 - y = -86.565x + 10806 ----- Oblique (univariate model)

1000 l R2 = 0.436

0 T I , , 1 l
40 0 90 100

60 70 8
A96 (YearS)

Figure 2.14. Fracture load-age plot for axial and oblique directions with univariate and
multivariate linear regression lines.

44

DISCUSSION

The objective of this study was to document peak loads generated on the human
patello-femoral joint during fracture experiments and the fracture patterns on the patella
for rigid interface impacts with a 90° ﬂexed knee. The experiments involved impact of
one specimen from a pair of isolated cadaver knees with an axial (directed along the axis
of the femur) patellar impact. The opposite knee was impacted with the same interface,
but the impact was directed obliquely (15° and 30° medial) to the femoral axis. While
most previous studies have focused on axial loading of the knee-thigh-hip complex,
experimental data documented in Melvin et al., 1975 suggest that the peak load in a bone
fracture experiment may be reduced and the patellar fracture pattern may be altered for an
oblique versus an axial directed impact on the knee. In the current study, the peak load
generated during fracture of the patella was signiﬁcantly reduced from 5.9 1 1.4 kN for
axial impacts to 3.5 i 1.3 kN for oblique impacts. Furthermore, in cases of patella
fracture these surfaces were generally oriented vertical (inferior to superior) for oblique
impacts and horizontal (medial to lateral) for axial directed impacts.

The failure loads in the current study were considered conservative measures for
tolerance of the human knee to blunt impact because the study involved only rigid impact
interfaces, the use of isolated joints, aged specimens with unknown tissue properties and
a repeated impact protocol. These factors may inﬂuence the failure loads. The
experimental results showed that the load-actuator displacement curves did not
signiﬁcantly vary between repeated tests on the same specimen, suggesting that joint
tissue integrity may have remained high prior to gross failure. The experimental protocol

was also to perform oblique experiments at 30° with respect to the axis of the femur.

45

However, after experiments on two of the younger specimens resulted in acute
dislocations of the patella, the angle of impact was reduced to 15° for the two youngest
specimens. The purpose of this change in protocol was to increase the probability of bone
fracture versus patellar dislocation.

The experimental data generated in the current investigation using axial directed
impacts on the patella compared well with previous experiments from this laboratory for
axial impacts on the isolated human knee. Previous studies show that the load to produce
linear horizontal or comminuted fracture of the patella with a rigid interface for the 90°
flexed joint is 5.5 i 1.8 kN [Atkinson and Haut, 2001a] and 4.7 i 1.6 kN [Atkinson and
Haut, 2001b] for specimens ages 70.4 2 14 years and 69 :1: 16 years, respectively, versus
5.9 :1: 1.4 in the current study using specimens aged 73 :1: 16.9 years.

The data in the current study, for an age range of 43 to 94 years, showed a
correlation of peak loads in axial and oblique fracture experiments with specimen age. An
earlier study indicates that for specimens ranging in age from 34-85 years, peak loads in
fracture experiments were not dependent on specimen age [Atkinson and Haut, 1998a].
While the number of specimens in each study was different (15 in the 19988 study and 9
in the current study) and the range of specimen ages was different, neither factor was
believed to fully explain discrepancies in the two studies. Another difference in the
studies was the rate of loading. The earlier study by Atkinson and Haut (1998a) used a
4.8 kg free flight impact missile with a rigid interface that delivered repetitive impacts at
sequentially higher energy until visible fracture of bone. The peak load was reached in
5.0 1 2.6 ms, which contrasts with 48 i 9 ms in the current study. Another previous study

by this laboratory suggests that the number of occult microcracks in the subchondral bone

46

underlying retropatellar cartilage increases with the rate of loading on the knee [Ewers et
al., 2000]. Since the same study shows fractured patellae for the high rate of loading
versus none at the same load for a low rate of loading, the possibility may exist that peak
load in patella fracture experiments could vary with specimen age more signiﬁcantly in
experiments using low versus high rates of loading on the knee. In fact, previous studies
by others typically use low rates of loading and document that the strength of cortical
bone decreases with specimen age [Burstein et al., 1976; Yamada, 1970]. These literature
data would support data generated in the current study, which also suggested a decrease
in peak load generated in patella fracture experiments with increased specimen age. The
suggestion that a fracture load-age correlation for rigid impacts on the human patella may
depend on the rate of impact loading would require validation in future studies. This type
of experimental data may be particularly relevant in the future as the driving population
tends to become more aged.

Another limitation of the current study was the use of only a rigid impact
interface. The primary reason for using this interface was to establish basic impact data
for the knee that could be compared with previous studies by this laboratory [Atkinson et
al., 1995; 1997; 1998a; 1998b; 1999; 2001a; 2001b; Ewers et al., 2000] and early studies
by others using the seated human cadaver [Melvin et al., 1975; Patrick et al., 1965;
Powell et al., 197 5]. The fracture load data are not to be interpreted as directly relevant to
impact interactions of the knee with deformable interfaces. A recent study, for example,
using a deformable interface (3.3 MPa crush strength aluminum material, Hexcel)
documents signiﬁcantly greater peak loads in patella fracture experiments with a

deformable versus a rigid interface condition on the 90° ﬂexed, isolated human joint

47

preparation described in the current study (Atkinson and Haut, 2001a). While for the
same impact energy more gross fractures and occult microcracks underlying retropatellar
cartilage are evident for rigid than for deformable interfaces, similar patella fracture
patterns are evident for these axial directed impacts on the knee. The occult microcracks
are found in areas of high patello-femoral contact pressure, and they are currently
hypothesized to be precursors of gross fracture of bone in the patella (Atkinson and Haut,
2001b).

Another previous study from this laboratory also shows that the location of the
horizontal linear fractures from axial directed impacts on the knee coincides with the
orientation and location of the largest principal tensile stresses developed in the
subchondral plate for a 3-D non-contact, impact model of the human patella [Atkinson
and Haut, 1999]. The 3-D model applied patello-femoral contact pressures on the
retropatellar surface and ﬁxed the anterior surface of the patella in the area of interface
contact. The study suggests that occult microcracks in the subchondral bone and the
subsequent gross fracture of the patella results from excessive tensile stresses developed
in the bone underlying retropatellar cartilage by the impact loading. It was also
interesting to note that horizontal fractures of the patella from axial impacts in the current
study were typically located more distally than described in previous studies [Atkinson
and Haut, 1995; Atkinson and Haut, 1999]. The effect may be due to localized tensile
stresses developed in the patella near the insertion of the patellar tendon. These stresses
did not exist in previous studies because those experiments avoided using load in the
quadriceps tendon during impact experiments and model simulations. In the current study

a 1.3 kN load was applied to the quadriceps tendon immediately prior to impacting the

48

knee and during impact on the knee. The major reason for applying a quadriceps force in
the current study was to help keep the patella in the femoral groove for the oblique
impacts. Therefore, the protocol also had to be used in the axial directed impacts on
contralateral limbs of each specimen. The 1.3 kN level of quadriceps force applied during
blunt impact to the knee was not intended to represent the maximum force that could be
generated by this muscle. Rather, this level was chosen to be relevant to a normal
physiological condition. In fact, the maximum in vivo quadriceps force is documented to
be approximately 3.9 kN for a group of 12 healthy male subjects from an early hallmark
study [Lindahl et al., 1969]. The quadriceps muscle force, on the other hand, developed
during extension of the knee from 90° to 60° of ﬂexion is approximately 30% of the
maximum force [Lieb and Perry, 1968]. Hence, a quadriceps force of 1.3 kN was chosen
for the current study. Microdamages have also previously been documented near the
insertion of the patellar tendon for axial patello-femoral impact studies with a canine
model [Thompson et al., 1991]. In the canine study "step-off" fractures of the
subchondral bone plate underlying retropatellar cartilage are evident, even without
superficial damage of retropatellar cartilage. Such microfractures may be precursors of
the gross bone fractures observed in the patella for the current study using the human
cadaver. These data may also suggest that high levels of in vivo quadriceps muscle
tension, that could approach maximal prior to or during a crash, may put the patella at
greater risk of fracture.

Another limitation of the current study was the implementation of a repetitive
impact protocol to increase the force input levels on the knee until gross fracture of bone.

One reason for employing this methodology was to compare the current results with

49

those from previous studies of this laboratory [Atkinson et al., 1995; 1997; 19988; 1998b;
1999; 20018; 2001b; Ewers et al., 2000]. While single versus multiple impacts were not
conducted previously on the human patello-femoral joint, 8 study was performed on the
isolated human tibio—femoral joint at 90° ﬂexion [Banglmaier et al., 1999]. The study
showed that repetitive impact testing results in tibio-femoral bone fracture for peak loads
of 8.0 1 1.8 kN. In contrast, single impact experiments on contralateral limbs results in 8
33% frequency of fracture for peak loads of 5.8 11.5 kN. As expected, single impact
subfracture experiments also show occult microcracks in subchondral bone underlying
tibial and femoral cartilage surfaces. Hence, the repeated impact test data of the current
study were very likely conservative measures of the fracture tolerance of the human
patella under a single impact scenario. On the other hand, occult microcracks may have
been equally produced in the oblique and the axial impacts, so the objective of the current
study was likely not compromised by this experimental protocol.

An axial versus oblique knee impact study has also been conducted using a small
animal model, the Flemish Giant rabbit [Ewers et al., 2002]. The study shows that one-
year following an oblique directed impact to the medial aspect of the patella, retropatellar
cartilage is mechanically softened and has numerous superﬁcial lesions (ﬁssures). In
contrast, following axial impact experiments with the same applied load on the patella,
few superﬁcial lesions are noted and the retropatellar cartilage is not softened. The effect
is hypothesized due to a reduction in the level of shear stresses developed in the
superﬁcial layer of retropatellar cartilage in axial versus oblique experiments [Li et al.,
1995]. These lower shear stresses in the retropatellar cartilage for axial versus medial

impacted knees are suggested to be the result of a more uniform distribution of contact

50

pressure across the medial and lateral patellar facets for the axial experiments, which
contrasts with a signiﬁcantly higher level of contact pressure on the lateral versus the
medial facet in oblique (medial-directed) impacts. The contact pressure distributions
shown for oblique and axial directed impacts on the patella for the rabbit model compare
well with data from the current study using the human cadaver knee. Since the current
study was conducted on aged cadavers, that typically have signiﬁcantly degraded
retropatellar cartilage compared to the rabbit model, it was difﬁcult to assess any
differences in the extent of impact-induced ﬁssuring of cartilage in axial versus oblique
experiments. In one experiment, however, a distinct surface ﬁssure was noted on the
lateral femoral condyle after acute dislocation of the patella during an oblique impact. It
was unknown, however, which impact generated this defect. Similarities of these data
with the animal model data may suggest a signiﬁcant chronic problem could exist
following oblique versus the axial directed impacts on the knee following even a
subfracture load on the patella. Acute subfracture injuries, however, would have to be
investigated in future experiments.

The current study was also limited by the use of human cadaver specimens, as the
potential effects of acute trauma could not be studied in a chronic setting. In fact, the
study of Thompson et al. (1991) using the canine model indicates that "step-off" fractures
of the subchondral bone plate heal after l-year, but the overlying cartilage repairs with
ﬁbrocartilage versus normal hyaline cartilage. While these data may suggest occult
microfractures themselves are not clinically relevant, the long-terrn durability of the
ﬁbrocartilaginous repair tissue in the joint is questionable. Correspondingly, clinical

studies also suggest a high percentage of patients suffering gross patella fractures tend

51

towards development of a chronic osteoarthritis in the patello-femoral joint [Robinson
and States, 1978; Nummi, 1971]. This potential chronic problem may put the oblique
impacted trauma patient at a greater risk of disease. Another concern for this trauma
patient is the potential for chronic disease of the knee following an acute dislocation of
the patella via an oblique impact, such as that shown in 2 of 9 specimens from the current
study. Numerous clinical investigations have described acute damage to the medial
retinaculum (capsule) following dislocations of the patella caused by excessive twisting,
valgus stress or direct blows to the knee [Burks et al., 1998]. This type of injury was
grossly visualized in one current experiment, 03-328. Clinical studies have associated
acute dislocation of the patella with damage to the medial patello-femoral ligament in
approximately 60-90% of cases [Sallay et al., 1996; Ahmed and Duncan, 2000; Nomura
et al., 2002]. While injury to the medial retinaculum could lead to chronic mal-tracking of
the patella and a subsequent chronic disease, early diagnosis of this soft tissue injury and
its surgical repair could restore normal tracking of the patella and minimize the potential
for chronic joint disease [Sandmeier et al., 2000]. On the other hand, clinical dislocations
of the patella are also associated with chondral and osteochondral lesions of the patella
and the lateral femoral condyle, as was observed in one specimen from the current study
[Sallay et al., 1996; Ahmed and Duncan, 2000]. These injuries have also been associated
with articular cartilage degeneration reminiscent of early-stage osteoarthritis [Johnson et
al., 1998]. While the current study suggested that oblique loading of the patella in an
automobile crash might increase the potential for patella fractures or dislocations, the
frequency of oblique versus axial loading in ﬁeld accidents was not investigated in the

current study. The current study involved experiments at 15° and 30° oblique to the

52

femoral axis. Melvin et al., 1975 performed experiments at 25° oblique to the axis of the
femur and had difficulty producing patella fracture except in two cases. Interestingly in
the NASS database, accidents that occur between 11 o’clock and 1 o’clock are all
categorized as frontal. This range of vehicle impact directions coincides with 1 30° for
the extremes of a frontal crash classiﬁcation. Thus, the occupant may impact the
automotive interior obliquely in a “NASS frontal” crash. On the other hand, in 12 o’clock
frontal crashes the occupant may also impact the instrument panel, or other surfaces,
obliquely. An early report by Schneider et al., 1983 indicates an abduction angle of the
occupant lower extremity for a “standardized normal driving posture” of mid-size males
at approximately 8°. This “splay angle” was measured in a simulated, vehicle-seating
package for a range of vehicles from that period. A standardized posture had to be
developed because they noted “occupant seating is very atypical”. In contrast, a more
recent unpublished report suggests automobile drivers can sit with knees signiﬁcantly
abducted under normal driving conditions [Reynolds, 1996]. In this study 39 drivers
were ﬁlmed during 2-hour highway drives in a mid-size sedan. Surface markers were
placed on the lateral epicondyle of the right knee and the right hip at the greater
trochanter. Locations of these markers were recorded with four video cameras during
the drives. The abduction angle was calculated in the horizontal plane, from a line
between these two points and a reference line perpendicular to the seat. The drivers
abducted (rotated the lower extremity laterally) 10-30° (Figure 2.15), with the degree of
abduction greater for taller males and greater for males than females. These data may
suggest that even in the event of a 12 o’clock frontal crash, automobile drivers may

contact the instrument panel or steering column with abducted lower

53

Stature (mm)

 

1 0 H I l T T I l
1 450 1 500 1 550 1 600 1 650 1 700 1 750 1 800 1 850 1 900

l T

4°“ 1 E If; 3.1 1 i

if 1'! l
gab I. 1 1i {II f ; E
:Hr l

 

-50 -
Figure 2.15. An abduction angle-stature plot for male and female drivers in a two-
hour automobile trip (H. Reynolds, ERL LLC). Data collected with each change in

position. The plot shows the mean and range of leg “splay” angles of the lower
extremity during the drives.

extremities. So, the oblique impact direction may be a more typical scenario in ﬁeld

accidents than an axial impact on the knee. This could predispose the occupant to patella

fracture or even acute patellar dislocation. Currently, this soft tissue injury is not
speciﬁcally documented in the NASS database. A documentation of patella fracture
orientation from ﬁeld accidents may be helpful for reconstruction studies, since oblique
versus axial experiments showed a difference in the patterns of patella fracture in the
current study. These data may help in the analysis of injury causations in ﬁeld accidents.
This study examined the peak loads on the human knee generated during fracture
of the patella with a rigid interface. As impact interfaces in current automobiles are
typically deformable, the current data is not directly applicable to this setting. However,

these data do expand on the current database that is used to establish fracture tolerance

criteria for the lower extremity.

54

CONCLUSION

The study showed that the patella was more vulnerable to fracture with a rigid
interface when the knee was impacted oblique versus axially with respect to the femoral
axis. Furthermore, while axial impacts resulted in horizontal linear or comminuted patella
fractures, oblique impacts (15° - 30° media] to the femoral axis) resulted in fractures
whose surfaces were oriented vertically across the patella. In two cases of oblique impact
on the knee the patella dislocated laterally prior to fracture. Soft tissue injuries were
noted in both cases. In the oblique and axial experiments, peak loads were also
correlated inversely with specimen age in the current study.

Since clinical studies often show an association of patella fractures and
dislocations with the onset of chronic disease in the knee, these trauma patients appear to
be more at risk from oblique than axial impacts on the knee. Yet, the current study only
involved rigid impact interfaces that are not similar to current automobile interiors. The
direct relevance of oblique versus axial impact loads on the knee to the automobile
occupant will ultimately need to be determined using more deformable impact interfaces

in future studies.

ACKNOWLEDGMENTS

This study was supported by a grant from the Centers for Disease Control and
Prevention (R49/CCR503607). Its contents are solely the responsibility of the authors and
do not necessarily represent the ofﬁcial views of the Centers for Disease Control and

Prevention. We also thank Mr. R.S. Wade, Director of the Anatomical Services Division /

55

State Anatomy Board, University of Maryland for help in procuring human test

specimens for this research project.

56

REFERENCES

Ahmed AM, Duncan NA. (2000) Correlation of patellar tracking pattern with trochlear
and retropatellar surface topographies. J Biomech Egr 122 (6): 652-660.

Atkinson PJ, Haut RC. (1995) Subfracture insult to the human cadaver patellofemoral
joint produces occult injury. J Orthop Res 13: 936-944.

Atkinson PJ, Garcia JJ, Altiero NJ, Haut RC. (1997) The Inﬂuence of Impact Interface on
Human Knee Injury: Implications for Instrument Panel Design and the Lower
Extremity Injury Criterion, Stapp Conf Proc 41: 167-180.

Atkinson PJ, Mackenzie CM, Haut RC. (1998a) Patellofemoral joint fracture load
prediction using physical and pathological parameters. SAE 67-73.

Atkinson PJ, Newberry WN, Atkinson TS, Haut RC. (1998b) A method to increase the
sensitive range of pressure sensitive ﬁlm. J Biomech 31: 855-859.

Atkinson PJ, Haut RC. (1999) Occult injuries in the subchondral plate are precursors to
gross fracture in an experimenal model of impact injury in the human knee. 45th
Orthopaedic Research Society Meeting.

Atkinson P, Haut R.C. (20018) Impact responses of the ﬂexed human knee using a
deformable impact interface. Journal of Biomechanical Engineering 123 (3): 205-211.

Atkinson PJ, Haut RC. (2001b) Injuries produced by blunt trauma to the human
patellofemoral joint vary with ﬂexion angle of the knee. J Orthop Res 19(5): 827-833.

Atkinson T, Atkinson P. (2000) Knee injuries in motor vehicle collisions: A study of the
National Accident Sampling System database for the years 1979-1995. Accid Anal
Prev 32 (6): 786.

Banglmaier R, Dvoracek-Driksna D, Oniang’o T, Haut R. Axial compressive load
response of the 90° ﬂexed human tibiofemoral joint. Stapp Conf Proc 1999; 42 127-
140.

Burks RT, Desio SM, Bachus KN, Tyson L, Springer K. (1998) Biomechanical
evaluation of lateral patellar dislocations. Am J Knee Surg 11 (1): 24—31.

Burstein AH, Reilly DT, Martens M. (1976) Aging of bone tissue: Mechanical properties.
J Bone Jt Surg 58-A (1): 82-86.

Dischinger P, Cushing B, Kems T (1992) Lower extremity fractures in motor vehicle

collisions: Inﬂuence of direction, impact and seatbelt use. 36th AAAM Conf: 319-
326.

57

Ewers BJ, Jayaraman V, Banglmaier R, Haut R. (2000) The effect of loading rate on the
degree of acute injury and chronic conditions in the knee after blunt impact. Stapp
Conf Proc 44: 299-314.

Ewers BJ, Weaver BT, Haut RC. (2002) Impact orientation can signiﬁcantly affect the

outcome of a blunt impact to the rabbit patellofemoral joint. J Biomech 35: 1591-
1598.

Johnson DL, Urban WP, Caborn DNM, Vanarthos WJ, Carlson CS. (1998) Articular
cartilage changes seen with magnetic resonance imaging-detected bone bruises

associated with acute anterior cruciate ligament rupture. Am J Sports Med 26: 409-
414.

Li X, Haut RC, Altiero NJ. (1995) An analytic model to study blunt impact response of
the rabbit P-F joint. J Biomech Egr 117: 485-491.

Lieb F, Perry J. (1968) Quadriceps function. An anatomical and mechanical study using
amputated limbs. J Bone Jt Surg 50A: 1535

Lindahl O, Movin A, Ringqvist I. (1969) Knee extension. Measurement of the isometric
force in different positions of the knee-joint. Acta Orthop Scand 40 (1): 79-85

Melvin J, Stalnaker R, Alem N, Benson J, Mohan D. (1975) Impact response and
tolerance of the lower extremities. Stapp Conf Proc 19: 543—559.

Miller T, Martin P, Crandell JR (1995) Cost of lower limb injuries in highway crashes.
Proc Inter Conf Pelvic Lower Extremity Inj: 47-57.

Morgan RM, Eppinger RH, Marcus 1H, Nichols H. (1990) Human cadaver and hybrid III
responses. Inter Conf Biomech Impacts.

Nomura E, Horiuchi Y, Inoue M. (2002) Correlation of MR imaging ﬁndings and open
exploration of medial patellofemoral ligament injuries in acute patellar dislocations.
Knee 9 (2): 139-143.

Nummi J. (1971) Fractures of the patella: A clinical study of 707 patellar fractures.
Annales Chirurgiae et Gynaecologiae Fenniae Sup 179: 1-85.

Patrick L, Kroell C, Mertz H. (1965) Forces on the human body in simulated crashes.
Stapp Conf Proc 9: 237-259.

Powell W, Ojala S, Advani S, Martin R. (1975) Cadaver femur responses to longitudinal
impacts. Stapp Conf Proc 19: 561-579.

Rice D, MacKenzie E, et al. (1989) Cost of Injury in the United States: A Report to
Congress, Atlanta, GA.

58

Reynolds, H. (1996) Personal Communication. Ergonomics Research Laboratory, LLC.
(See www.crlllc.com for details of the study.)

Rupp JD, Reed MP, Van Ee CA, Kuppa S, Wang SC, Goulet JA, Schneider LW. (2002)
The tolerance of the human hip to dynamic knee loading. Stapp Car Crash Conf J
46:211-228.

Robinson SC, States JD. (1978) Epidemiology, treatment, and prevention of patellar
fractures. In: Symposium on Reconstructive Surgery of the Knee: 11-120. St. Louis,
C. V. Mosby Company.

Sallay P, Poggi J, Speer K, Garrett W. (1996) Acute dislocation of the patella: A
correlative pathoanatorrric study. Am J Sports Med 24(1): 52-60.

Sandmeier RH, Burks RT, Bachus KN, Billings A. (2000) The effect of reconstruction of
the medial patellofemoral ligament on patellar tracking. Am J Sports Med 28 (3):
345-349.

Schneider L, Robbins D, Pﬂug M, Snyder R. (1983) Anthropometry of motor vehicle
occupants. USDOT NHTSA Publication (1): 33.

Thompson R, Oegema T, Lewis J, Wallace L. (1991) Osteoarthritic changes after acute
transarticular load. J Bone Jt Surg 73-A (7): 990-1001.

Yamada H. (1970) Mechanical properties of locomotor organs and tissues. In: Strength of

Biological Materials: 19-105. Ed by PG Evans. Baltimore, The Williams & Wilkins
Company.

59

CHAPTER THREE

EXCESSIVE COMPRESSION OF THE HUMAN TIBIO-FEMORAL

JOINT CAUSES ACL RUPTURE BEFORE BONE FRACTURE

ABSTRACT

The knee is one of the most frequently injured joints in the human body. A recent
study suggests that axial compressive loads on the knee may play a role in injury to the
anterior cruciate ligament (ACL) for the ﬂexed knee, because of an approximate 10°
posterior tilt in the tibial plateau (Li et al., 1998). The hypothesis of the current study was
that excessive axial compressive loads in the human tibio-femoral (TF) joint would cause
relative displacement and rotation of the tibia with respect to the femur, and result in
isolated injury to the ACL when the knee is ﬂexed to 60°, 90° or 120°. Sixteen isolated
knees from eleven fresh cadaver donors (74.31105 years) were exposed to repetitive TF
compressive loads increasing in intensity until catastrophic injury. ACL rupture was
documented in 14/16 cases. The maximum TF joint compressive force for ACL failure
was 4.9 d: 2.1 kN for all ﬂexion angles combined. For the 90° ﬂexed knee, the injury
occurred with a relative anterior displacement of 5.413.8mm, a lateral displacement of

4.111.4mm, and a 7.817.0° internal rotation of the tibia with respect to the femur.

60

INTRODUCTION

The knee is one of the most frequently injured joints in the human body.
Epidemiological studies have shown there are approximately 80,000 anterior cruciate
ligament (ACL) tears in the United States each year with a total cost near one billion dollars
(Grifﬁn et al., 2000). This study also reports that 70% of ACL tears are due to “non-contact”
types of injury. In a study of high school soccer, volleyball and basketball players for one
season, 6 of 8 serious ACL injuries were “non—contact” (Hewett et al., 1999). Many studies
have investigated the loading mechanisms that cause injury to the ACL. Boden et al. (2000)
suggests that ACL injuries frequently occur in landing from a jump on one or both legs. In
jump landings the knee may be ﬂexed 60-80° (Hewett et al., 1996). ACL injury is also
common in skiing with 25% to 30% of all ski related knee injuries involving the ACL
(Speer et al., 1995). These injuries are mainly associated with twisting or a hard landing
from a jump with a ﬂexed knee (Ettlinger et al., 1995).

The ACL functions as the primary restraint to limit anterior tibial displacements for
both 30° and 90° of knee ﬂexion GButler et al., 1980, Fukubayashi et al., 1982, Torzilli et al.,
1994). It provides approximately 85% of the total ligamentous restraining force during
anterior tibial displacement. Fleming et al. (2001) conﬁrm this notion by showing that
normal, in vivo weight bearing in the knee induces tensile strain in the ACL during anterior
neutral-position shift of the joint. This study also supports earlier investigations showing that
tensile forces develop in the ACL under physiological levels of tibio-femoral (TF) joint
compressive loading (Li et al., 1998, Markolf et al., 1981). Torzilli et al. (1994) show that
physiological levels of compressive loading on the human TF joint generates anterior tibial

translation with respect to the femur for all ﬂexion angles greater than 15°. This response is

61

primarily thought due to an inherent posterior tilt of the tibial plateau of 10°-15° (Li et al.,
1998) (Figure 1). This study further suggests “that excessive compressive loads caused by
impact loads along the tibial shaft (e.g., load from a jump landing) may contribute to injury

of the anterior cruciate ligament, especially when the knee is ﬂexed.”

     

QUADRICEPS

=IO'

15 DEGREES

Figure 3.1. Anterior neutral-position shift, which is attributed to the inherent tilt of the tibial
plateau, may cause an anterior translation of the tibia during TF joint loading.

The hypothesis of the current study was that the injury mechanism for a ﬂexed,
isolated human knee joint under excessive TF compression would be ACL rupture. The
study will document the amount of relative joint displacement and rotation, and the
compressive loads required to cause this injury in joints ﬂexed 90°. Additional data showing
the same injury in knees ﬂexed to 60° and 120° will also be documented in the study.
METHODS

Experiments were conducted on 16 knees from 11 pairs of human TF joints (74.3 1
10.5 years of age). The joints were procured through university sources (See
Acknowledgement), stored at -20° C, and thawed to room temperature for 24 hours prior to

testing. The joints were selected from donors with no known knee injuries. Five joint pairs

62

had been previously thawed and refrozen after another study. One knee from each of these
ﬁve pairs was randomly selected for sequential TF joint loading with 60° of joint ﬂexion (0°
ﬂexion for a straight leg), and the opposite joint was loaded with 120° of ﬂexion. Six other
knee specimens were loaded with 90° of joint ﬂexion. In these 6 specimens motions of the
femur were measured during axial loading of the tibia. Each joint preparation was sectioned
approximately 15 cm proximal and distal to the center of the knee. The femur and tibia
shafts were cleaned with 70% alcohol and potted in cylindrical aluminum sleeves with room
temperature curing epoxy (Fibre Strand, Martin Senior Corp. Cleveland, OH), using a
previously established protocol (Banglimaier et al., 1999).

A load transducer (Model #101018-2500, Instron Corp. Canton, OH; load accuracy
0.5% of 18896 N and resolution 0.004 N) was ﬁxed to the actuator of a servo-hydraulic
materials testing machine (Model 1331, Instron Corp. Canton, OH) (Figure 2). A rotary
encoder (Model #RCH25D-6000, Renco Encoders Inc. Goleta, CA; accuracy 10.01°) and
stainless steel shaft were attached to the offset bar. The stainless steel shaft was connected to
the potted tibia after passing through a sleeve bearing on a horizontal stabilizer bar. The
rotary encoder allowed the intemal-external rotations of the potted tibia to be measured with
respect to the femur. The sleeve bearing and horizontal stabilizer bar allowed axial forces in
the tibia to compress the TF joint with a minimal bending moment applied to the load
transducer. The potted femur was secured to a ﬁxture that allowed knee joint ﬂexion angles
of 60°, 90° and 120°. A bed of epoxy was added to the anterior surface of the femur to help
distribute compressive loads over the proximal end of the femur. The ﬁxture was attached to

an X-Y translational plate that had linear encoders (Model #XOOZOIA, Renishaw, Hoffman

63

Estates, IL; accuracy 11 um) attached to record anterior-posterior and medial-lateral

motions of the femur relative to the tibia during tests

 

ACTUATOR

 

 

 

[ ILOAD CELL

 

 

 

 

 

   
  
 

 

 

 

Adjustable
Angle

Potting Material ——9 ‘
I c‘.’ \

ADJUSTABLE}
51mm}: l

 

 

1 l I

 

 

 

 

 

 

 

 

 

X - YPlate

 

 

 

 

I

Figure 3.2. An adjustable ﬁxture that recorded the motion of the femur in the X(anterior-

posterior), Y(medial-lateral) plane and rotation of the tibia about its axis during TF joint

loading.

on the 90° ﬂexed joints. All joints were repeatedly loaded using a protocol in which a

preload of 6-8 N was applied, followed by a single 10 Hz haversine load waveform that

simulated the time to peak ground reaction loads in a typical jump landing (Richards et al.,

1996). The load was increased by increments of 500 N in successive tests until catastrophic

injury of the joint. The injury type and location were documented photographically for each

specimen.

TF compression load and time to peak load were recorded in these experiments. In
the 90° ﬂexion experiments three additional variables; anterior-posterior femur displacement
relative to the tibia, medial-lateral femur displacement relative to the tibia, and intemal-
external rotation of the tibia relative to the femur were also recorded. All data are given as
mean 1 one standard deviation.

RESULTS

Fourteen of sixteen knee joints at ﬂexion angles of 60°, 90° and 120° suffered ACL
rupture at a combined peak load of 4.9 1 2.1 kN. ACL ruptures were nrid-substance and
occurred near its femoral insertion (Figure 3). Five of six tests with 90° ﬂexion resulted in a
torn ACL at a peak load of 5.6 1 3.0 kN. Four of ﬁve 60° ﬂexion tests and all ﬁve 120°
ﬂexion tests resulted in a torn ACL at peak loads of 4.9 1 1.5 kN and 4.4 1 1.0 kN,
respectively (Table 1).

In the 90° test series the peak TF loads resulted in a posterior displacement of 5.4 1
3.8 mm for the femur with respect to the tibia/ﬁbula (Figure 4 A&B). There was also 4.1 1

1.4 mm of medial motion of the femur with respect to the tibia/ﬁbula, and 7.8 1 7.0 degrees

 

Figure 3.3. A typical injury in an unconstrained knee was rupture of the ACL.

65

of internal rotation of the tibia in these experiments (Table 2). Similar joint motions were

observed in knees at 60° and 120° of ﬂexion, but were not measured.

 

Table 3.1. Experimental data for 60° and 120° ﬂexed knee joint preparations in failure tests.

TIBIA

PROXINIAL
MOTION

K .
POSTERIOR ‘0"

MOTION <

 

Figure 3.4. The application of TF loads resulted in proximal translation and internal rotation

of the tibia, whereas the femur moved medial and posterior relative to the tibia.

Ade 550111501

 

Table 3.2. Experimental data for 90° ﬂexed knee joint preparations in failure tests.

66

DISCUSSION

The current study showed that rupture of the ACL occurred in 14/ 16 cases at 4.9 1
2.1 kN of TF joint compressive loading. During compression the femur displaced posteriorly
and medially with respect to the tibia, and the tibia rotated internally with respect to the
femur. These motions would be similar to anterior and lateral motion of the tibia with
respect to a ﬁxed femur. Since the goal was to apply axial loads in the tibia, the femur was
allowed to move in the current study.

Impulsive axial compressive loading of the knee occurs in vivo during landing from a
jump with a ﬂexed knee (Hewett et al., 1996). During landing ground forces are transmitted
through the tibia to the TF joint. The ground reaction forces during a landing are
approximately 4.2 and 6.1 times body weight for females and males, respectively (Hewett et
al., 1996). These force levels are high enough to suggest that a typical individual may
rupture an ACL, based on the current data from the aged isolated cadaver knee preparation.
Additionally, the forces of contracting muscles, such as the quadriceps, could also produce
signiﬁcant compressive loads in the TF joint (Torzilli et al., 1994).

Although the line of action of the quadriceps muscle on the tibia is directed slightly
posteriorly for ﬂexion angles above 65° (Draganich et al., 1987), the effect of quadriceps
loading on anterior-posterior tibial translation at 90° of knee ﬂexion has been inconclusive
(Li et al., 1999, Li et al., 2004, Torzilli et 81., 1994). These studies document between 0.2
mm of posterior translation and 1 mm of anterior translation of the tibia for quadriceps loads
between 133 N and 400 N. This may be due to the competing effects of the slight posterior
vector of the quadriceps and the stronger proximal vector producing joint compression. A

case study of elite skiers by McConkey (1986) suggests that extreme levels of quadriceps

67

force might be sufﬁcient to produce ACL rupture. In the documented cases skiers
strenuously attempting to recover from a falling-back position damaged their ACL.
However, in at least 5 of 13 cases of ACL injury from that study a jump landing caused the
large contractions of the quadriceps muscles. In light of data from the current study the
mechanism of ACL injury in these cases may have actually been associated more with the
combination of TF compression loads generated during the landing, and the joint
compression due to quadriceps contraction.

In contrast, hamstrings muscle contraction produces a posterior directed force on the
tibia (Pandy and Shelbume, 1997 and Li et al., 1999). This joint constraint force was
documented in a previous study at 90° ﬂexion for isolated joint specimens (J ayaraman et al.,
2001). A mean load of 1.2 1 0.5 kN is sufﬁcient to prevent anterior motion of the tibia at
fracture-causing load levels of TF joint compression (9.2 1 2.6 kN, n=6) .

While previous studies have documented ACL injury loads from anterior shear
loading (Aune et al., 1997) and quadriceps tension for the slightly ﬂexed knee (DeMorat et
al., 2004), the authors are not currently aware of previous studies that document the TF
compressive loads that cause rupture of the ACL. Several knee joint compression studies
have used a constrained knee joint model in which the tibia is only allowed to move in the
proximal-distal direction relative to the femur (Hirsch and Sullivan, 1965, Kennedy and
Bailey, 1968 and Banglmaier et al., 1999). The most recent study, Banglmaier et al. (1999),
examines TF impact responses using 90° ﬂexed, isolated human knee joints. The authors
document fracture of the medial and/or lateral tibial plateau, or medial femoral condyle or
notch at a maximum compressive load of 8.0 1 1.8 kN. This load is signiﬁcantly higher than

the failure load documented in the current study. In fact, the lower extremity injury criterion

68

for TF joint loading used in the automotive industry is currently based on data from these
constrained knee joint studies (Kuppa et al., 2002). The current study would suggest that the
current injury criterion should be re-examined using unconstrained knee joint preparations.
A study by Woo et al. (1991) documents that approximately 6 mm of tensile
deformation and 658 1 129 N of load is required to rupture the 60-97 year old ACL mid-

substance in its anatomic orientation. Since the ACL is oriented approximately 7° with

respect to the tibial plateau in the 90° ﬂexed knee (Herzog and Reed, 1993), it seems
reasonable that tensile failure of the ACL occurred at 5.4 1 3.8 mm of posterior translation
of the femur in the current study. Yet, this does not take into account medial translation of
the femur and internal rotation of the tibia that also occurred during TF joint compression.
Because of differences in test methodologies between Woo et al., 1991 and the current study
(tension versus compression) failure loads could not be directly compared.

There were a number of limitations in the current study that need to be addressed in
the future. Since isolated knee preparations were used, the potential effects of muscle forces
acting across the knee were not included. Since documented loads in jump landings may
approach those causing ACL rupture in the current study with aged cadaver specimens,
these data suggest muscle action is essential in preventing ACL injuries. Because of their
ages these specimens only represent a small portion of today’s population, and not that of
the population active in sports. Additional studies are required on a younger population of
human knees to address the potential for ACL injuries in sports related jump landings.
Additionally, the repetitive nature of the tests in the current study may be a limitation. The

potential consequence of accumulated microdamages in soft tissue structures, such as the

69

ACL, must be considered in future studies. Microdamages are known to occur prior to gross
rupture in other ACL models (Yahia et al., 1990).

In summary, this study showed that ACL rupture occurred in the human knee via
excessive TF compressive loading. Since the peak compressive loads were in the range that
could regularly be generated during 8 jump landing, the role of muscle forces in preventing
anterior translation, lateral displacement and internal rotation of the tibia are essential
constraints and therefore must be investigated in future studies using younger human

cadaver specimens that represent the population involved in sports today.
ACKNOWLEDGEMENT

This study was supported by a grant from the Centers for Disease Control and
Prevention, National Center for Injury Prevention and Control (R49/CCR503607). Its
contents are the sole responsibility of the authors and do not necessarily represent the
ofﬁcial views of the Centers for Disease Control and Prevention. The authors wish to
gratefully acknowledge Cliff Beckett, Vijay Jayaraman, Eric Sevensma, Masaya Kitagawa
and Chris O’Neill for technical assistance during this study. We also thank Mr. R.S. Wade,
Director of the Anatomical Services Division / State Anatomy Board, University of

Maryland for help procuring these test specimens.

70

REFERENCES

Aune, A.K., Cawley, P.W., Ekeland, A., 1997. Quadriceps muscle contraction protects the
anterior cruciate ligament during anterior tibial translation. Am J Sports Med. 25 (2): 187-
190.

Banglmaier, R., Dvoracek-Driksna, D., Oniang'o, T., Haut, RC, 1999. Axial compressive
load response of the 90° ﬂexed human tibio-femoral joint. 43“1 Stapp Car Crash Conf.,
127-140.

Boden, B.P., Dean, G.S., Feagen, J .A Jr., 2000. Mechanisms of anterior cruciate ligament
injury. Orthopedics. 23 (6): 573-8.

Butler, D.L., Noyes, F.R., Grood, ES, 1980. Ligamentous restraints to anterior-posterior
drawer in the human knee. J Bone Jt Surg. 62A: 259-270.

DeMorat, G., Weinhold, P., Blackburn, T., Chudik, 8., Garrett, W., 2004. Aggressive
quadriceps loading can induce noncontact anterior cruciate ligament injury. Am J Sports
Med. 32 (2): 477-483.

Draganich, L.F., Andriacchi, T.P., Anderson, G.B.J., 1987. Interaction between intrinsic
knee mechanics and the knee extensor mechanism. J Orthop Res. 5: 539-547.

Ettlinger C.F., Johnson, R.J., Shealy, JE., 1995. A method to help reduce the risk of serious
knee sprains incurred in alpine skiing. Am J Sports Med. 23(5): 531-537.

Fleming, B., Renstrom, P., Beynnon, B., 2001. The effect of weight-bearing and external
loading on anterior cruciate ligament strain. J Biomech 34: 163-170.

Fukubayashi, T., Torzilli, P.A., Sherman, M.F., Warren, RR, 1982. An in vitro
biomechanical evaluation of anterior-posterior motion of the knee, tibial displacement,
rotation and torque. J Bone Jt Surg. 64A: 258-264

Grifﬁn, L.Y., Agel, J ., Albohm, M.J., et al. 2000. Noncontact anterior cruciate ligament
injuries: Risk factors and prevention strategies. J Am Acad Orthop Surg. 8 (3): 141-150.

Hewett, T.E., Stroupe, A.L., Nance, T.A., Noyes, FR, 1996. Plyometric training in female
athletes: Decreased impact forces and increased hamstrings torques. Am J Orthop Soc
Sports Med. 24 (6): 765-773.

Hewett, T.E., Lindenfeld, T.N., Riccobene, J.V., Noyes, FR, 1999. The effect of
neuromuscular training on the incidence of knee injury in female athletes: A prospective
study. Am J Orthop Soc Sports Med. 27 (6): 699-705.

Herzog, W., Reed, L.J., 1993. Lines of action and moment arms of the major force-carrying
structures crossing the human knee joint. J Anat. 182 (Pt 2): 213-30.

71

Hirsch, G., Sullivan, L., 1965. Experimental knee-joint fractures. Acta Orthop Scand. 36:
391-399.

J ayaraman, V.M., Sevensma, E.T., Kitagawa, M., Haut, RC, 2001. Effects of anterior-
posterior constraint on injury patterns in the human knee during tibio-femoral joint loading
from axial forces through the tibia. Stapp Car Crash J. 45: 449-468.

Kennedy, J.C., Bailey, W.H., 1968. Experimental tibial-plateau fractures: studies of the
mechanism and a classiﬁcation. J Bone Jt Surg 50(A): 1522-1534.

Kuppa, S. 2002. An overview of knee-thigh-hip injuries in frontal crashes in the United
States. 17th ESV

Li, G., Rudy, T., Allen, C., 1998. Effect of combined axial compressive and anterior tibial
loads on in situ forces in the anterior cruciate ligament: a porcine study. J Orthop Res
16:122-127.

Li, G., Rudy, T.W., Sakane, M., Kanamori, A., Ma, C.B., Woo, S.L.-Y., 1999. The
importance of quadriceps and hamstring muscle loading on knee kinematics and in-situ
forces in the ACL. J Biomech 32: 395-400.

Markolf, K.L., Bargar, W.L., Shoemaker, S.C., Amstutz, HQ, 1981. The role of joint load
in knee stability. J Bone Jt Surg. 63: 570-585.

McConkey, JP., 1986. Anterior cruciate ligament rupture in skiing: A new mechanism of
injury. Am J Sports Med. 14(2): 160-164.

Pandy, M.G., Shelbume KB, 1997. Dependence of cruciate-ligament loading on muscle
forces and external load. J Biomech. 30(10): 1015-1024.

Richards, D.P., Ajernian, S.V., Wiley, JP, Zernicke RF, 1996. Knee joint dynamics
predict patellar tendonitis in elite volleyball players. Am J Sports Med. 24 (5): 676-683.

Speer, K.P., Warren, R.F., Wickiewicz, T.L., Horowitz, L., Henderson, L., 1995.
Obeservations on the injury mechanism of anterior cruciate ligament tears in skiers. Am J
Sports Med. 23 (1): 77-81.

Torzilli, P., Deng, X., Warren, R., 1994. The effect of joint-compression load and
quadreceps muscle force on knee motion in the intact and the anterior cruciate ligament-
sectioned knee. Am J Sports Med. 22: 105-112.

Woo, S.L., Hollis, J .M., Adams, D.J., Lyon, R.M., Takai, S., 1991. Tensile properties of the
human femur-anterior cruciate ligament-tibia complex: The effects of specimen age and
orientation. Am J Sports Med. 19 (3): 217-225.

Yahia, L., Brunet J ., Labella, S., Rivard CH, 1990. A scanning electron microscopic study
of rabbit

72

CHAPTER FOUR

THE EFFECT OF AXIAL LOAD IN THE TIBIA ON THE
RESPONSE OF THE 90° F LEXED KNEE TO BLUNT IMPACTS

WITH A DEF ORMABLE INTERFACE

ABSTRACT

Lower extremity injuries are a frequent outcome of automotive accidents. While
the lower extremity injury criterion is based on fracture of bone, most injuries are of less
severity. Recent studies suggest microscopic, occult fractures may occur in the knee for
impacts with rigid and deformable interfaces due to excessive levels of patello-femoral
contact pressure. One method of reducing this contact pressure for a 90° ﬂexed knee is to
provide a parallel pathway for knee impact loads into the tibial tuberosity. Yet, blunt
loads onto the tibial tuberosity can cause posterior drawer motion of the tibia, leading to
injury or rupture of the posterior cruciate ligament (PCL). Recently studies have shown
that axial loads in the tibia, which are measured during blunt loading on the knee in
typical automobile crashes, can induce anterior drawer motion of the tibia and possibly
help unload the PCL. The purpose of the current study was to explore the effect of
combined anterior knee loading (AKL) and axial tibia loading (ATL), without muscle
tension, on response and injury for the 90° ﬂexed human knee.

In repeated impacts with increasing ATL the stiffness of the knee to an AKL
impact increased. For a 3 kN AKL, the stiffness of the knee increased approximately 26%
when the ATL was increased from 0 kN to 2 kN. For 6 kN and 9 kN AKL, the stiffness
was increased approximately 17% and 20%, respectively, when the ATL was increased

from 0 kN (uniaxial) to 4 kN (biaxial). The posterior tibial drawer was shown to increase

73

with increased AKL and decrease with increased level of ATL at an average of 0.3 mm
per 1 kN ATL for both the 3 kN and 6 kN ATL scenarios. For 9 kN AKL this drawer
displacement was signiﬁcantly reduced for biaxial versus uniaxial impacts, from 7.4114
mm to 5.8106 mm, respectively. Additionally, the percentage of the load carried by the
tibial tuberosity increased with an ATL. For AKL impacts of 3, 6, and 9 kN, the
percentage of load carried by the tibial tuberosity increased from 2.1% (range 0-19%) to
4.9% (0-36%), 2.1% (0-15%) to 6.9% (0-36%), and 8.7% (0-25%) to 12.7% (0-33%),
respectively, between uniaxial and biaxial tests. The biaxial loading scenario also resulted
in a reduction of the patello-femoral contact force by approximately 11.2% versus
uniaxial impacts. Ten knee impacts resulted in PCL tears at an average peak load of
12712.4 kN in biaxial impacts (n=5) and 12013.1 kN for uniaxial impacts (n=5). These
PCL injured specimens had an average age of 62111.3 years. The remaining specimens
(78112.9 years of age) had bone fractures at approximately 9.0131 kN.

This study showed that combinations of compressive ATLs, whose peaks occur at
nearly the same time and magnitude as AKL, have a stiffening effect on the response of
the knee impacting a stiff but deformable interface. Furthermore, ATL can reduce the
posterior drawer of the tibia, which is the basis for PCL injury in the knee. While the
current injury tolerance criterion reﬂect the vulnerability of the PCL to injury by limiting
tibial drawer to 15 mm, the current dummy design does not incorporate the stiffening
effect of an ATL that may occur at the same time as knee contact into an instrument

panel in atypical crash environment.

74

INTRODUCTION

The yearly medical cost of automobile accident injuries in the United States in
2000 was $32.6 billion (Blincoe et al., 2002). These injuries account for approximately
3.2% of the US total medical costs, second only to cancer (Miller et al., 1998). Lower
extremity injuries are a frequent outcome of automotive accidents (Fildes et al., 1997).
Kuppa (2002) documents that the comprehensive cost of lower extremity injuries is $7.64
billion annually, and it makes up a signiﬁcant percentage of the total motor vehicle
trauma. These injuries include 18% of all Abbreviated Injury Scale (AIS) 2+ injuries to
front seat occupants in frontal collisions and 23% of the associated life years lost to
injury. Injuries to the knee-thigh-hip complex account for 52%, or $4 billion, of these
lower extremity injury costs.

The current injury threshold limit for the knee-thigh-hip complex prescribed by
FMVSS 208 and used in NCAP is 10 kN of axial femur load for the 50th percentile male.
This level of femur load has been associated with a 35% probability of AIS 2+ injury in
human cadavers (Morgan et al., 1990). Patella and distal femur fractures were the most
common injuries in this data set. These data compare to those in a recent study of the
1993-1995 National Accident Sampling System (NASS) showing that 17% of knee
injuries in automobile crashes for those years were patella fractures, which occurred at a
frequency similar to that of femur fractures (Atkinson and Atkinson, 2000). In contrast,
tendon and ligament injuries (AIS 2 and 3) account for approximately 2.5 % of all the
knee injuries.

Much of the data reported in the Morgan et al. (1990) study documents

experimental patella and distal femur fractures using rigid impact interfaces. Yet,

75

automotive interior contact points would more typically involve contact on the knee via a
deformable interface. Early cadaver studies have shown that while rigid interfaces into
the knee typically generate patella fracture, deformable interfaces more often produce
femur fractures (Powell et al., 1975, Melvin et 81., 1975, Patrick et al., 1965). Recent
studies from this laboratory, using isolated cadaver knees, conﬁrm the data of earlier
studies on whole human cadavers. While 5 kN impacts, via a rigid interface, result in
transverse patella fracture, no injuries have been documented in the patella or femur
following 5.8 kN impacts via a deformable interface (1.4 MPa crush strength Hexcel)
(Atkinson et al., 1997). In contrast, 5 kN impacts with a slightly stiffer, but deformable,
interface (3.3 MPa crush strength Hexcel) produced occult rrricrocracks underlying the
retropatellar cartilage (Atkinson and Haut, 2001). A recent study shows that these types
of occult microcracks are seen in automobile accident victims suffering hip fractures
(Bealle and Johnson, 2000). The clinical literature suggests these occult microcracks
(bone bruises) occur often in ligament injured sports patients and lead to early
degradation of knee joint cartilage reminiscent of that which occurs in the early stages of
osteoarthritis (Johnson et al., 1998).

Atkinson et al. (1997) also suggest that these occult rrricrocracks of bone
underlying the retropatellar cartilage in the joint are caused by excessive levels of patello-
femoral (PF) joint pressure during blunt impact onto the knee, and they are precursors of
a gross fracture of the patella. One method of reducing PF joint contact forces during a
blunt impact is to provide a parallel pathway of load transmission across the knee, by
increasing the contact area over the knee (Hering and Patrick, 1977). For a typical crash

conﬁguration of a 90° ﬂexed knee this would involve contact on the tibial tuberosity

76

(Figure 1). Yet, blunt loads acting solely on the tibial tuberosity have been shown to
present the possibility of damage to the posterior cruciate ligament (PCL) via rupture or
avulsion from the tibia. Blunt loads of 7 kN onto the 90° ﬂexed knee (contacting the
patella and tibial tuberosity) of seated human cadavers resulted in ﬁve PCL avulsion
fractures and one PCL tear out of eight knees (Viano et al., 1978). Isolated joint tests in
the same study show partial tears of the PCL at a relative translation (drawer) of the tibia
with respect to the femur of 14.4 mm and complete failure of this ligament at 22.6 mm.
Based on these data Mertz (1993) recommends an injury threshold level of 15 mm of
relative translation between the femur and tibia at the knee joint for a 50th percentile

male.

 

Patella

Femur Deformable
Interface

m. l

Tuberosity /

Tibia

 

 

 

 

 

 

 

 

 

LC' \0’
Figure 4.1. Orientation of anterior knee impact on the patella and tibial

tuberosity.

77

Interestingly, while a blunt impact to the front of the knee via the tuberosity of the
tibia can result in posterior translation (drawer) of the tibia relative to the femur, recent
studies show that axial tibia loads can cause anterior translation of the tibia with respect
to the femur (Torzilli et al., 1994, Fleming et al., 2001, Li et al., 1998, Markolf et al.,
1981). In fact, a recent study has shown that approximately 5.8 kN of axial tibia load
(ATL) is sufﬁcient in a 90° ﬂexed, isolated human knee to cause rupture of the anterior
cruciate ligament (ACL) (J ayaraman et al., 2001). This response is primarily thought due
to an inherent anterior-posterior tilt of the tibial plateau. While this type of injury has
only been shown in a few experiments (Funk et al., 2000), the injury is not well
documented in the NASS database per se. Crash simulation data suggests that the level of
ATL needed for ACL rupture may be achieved (Bedewi and Diggs 1999). One possible
explanation for the lack of ﬁeld ACL injury data in auto crashes may be that when ATL
peaks during a crash, the knee may also be in contact the instrument panel (knee bolster),
and anterior loads in the knee (AKL) may also peak at nearly the same time and

magnitude (Bedewi and Diggs, 1999) (Figure 2).

AKL\

  

ATL /

Figure 4.2. Combined loading scenario caused by instrument panel contact and ﬂoor pan

intrusion.

78

The AKL may act to restrain anterior translation of the proximal tibia, preventing ACL
injury. In effect, the knee may be “constrained”. In this situation experimental data from
isolated joints under ﬂexion angles varying from 0-200 from a straight leg document
tibial plateau fractures occurring at approximately 8 kN of ATL (Hirsch and Sullivan,
1965). Based on those tests, Mertz (1993) recommends a proximal tibia axial force limit
of 8 kN to address tibial plateau and split femoral condyle fractures in the 50th percentile
male. The Jayaraman et al. (2001) study documents fracture of the tibial plateau and
femoral condyle at 9.2 kN of ATL for the 90° ﬂexed, isolated tibio-femoral joint.
Banglemaier et al. (1999) dynamically tested 90° ﬂexed, isolated tibio-femoral joints and
document fractures of the femoral notch, femoral condyles, tibial plateau and
combination injuries at an average peak load of 8 kN delivered axially through the tibia.
The objective of the current study was to investigate the effects of combined axial
loads in the tibia (ATL) and anterior knee loads to the patella and tibial tuberosity (AKL)
via a stiff but deformable impact interface using isolated human knee joint preparations
ﬂexed at 90°. The hypotheses of the study were that; (1) ATL presented during blunt
knee impacts would result in anterior translation of the tibia, thereby stiffening the knee’s
response during contact with a deformable interface. (2) The constraint of the knee with
the deformable interface would prevent ACL rupture from ATL large enough to rupture
the ACL in unconstrained tests. (3) The application of an ATL during blunt knee contact
involving the tibial tuberosity would help unload the PF joint due to increased load being
supported in the tibia via its anterior translation. (4) And, the AKL required to generate

PCL rupture, via contact on the tuberosity of the tibia, would be increased by an anterior

79

directed resultant force coming from an axial load component generated by

simultaneously applying an axial load in the tibia.

METHODS

Blunt impact was delivered to ten pairs and one single knee joint aged 71.71146
years. The limbs were procured from university sources (see Acknowledgment) and
stored at —20°C until testing. The joints were selected from donors with no known knee
injuries or signs of surgical intervention during a postmortem evaluation. Twenty-four
hours prior to testing the joints were thawed to room temperature. The preparations were
sectioned 20 cm inferiorly and superiorly to the knee joint. Superﬁcial muscle tissues
were excised from each preparation leaving the articular joint capsule intact. Therefore,
muscle tension was not simulated during the experiments. The femur and tibia/ﬁbula
were cleaned with alcohol and potted in 6.3 cm diameter cylindrical aluminum sleeves
with room temperature curing epoxy (Fibre Strand #6371, Martin Senour Co. Cleveland,
OH). Figure 3 shows that the tibia/ﬁbula was centered and aligned axially within the 10
cm deep cylinder. The femur was also aligned axially, but the posterior surface was
placed adjacent to the edge of a 14 cm deep cylinder and extra epoxy was placed between
the anterior surface of the femoral condyle and an extended piece of the aluminum
cylinder. This extra epoxy supported the femoral condyle under an ATL. Each test
specimen was mounted in the ﬁxture with 90° of ﬂexion. This ﬂexion angle was visually
established by aligning the femur and tibia with a 90° square tool. The femur was

oriented vertically with the knee joint facing up, and the tibia was positioned horizontally

80

 

 

    

 

Epoxy Potting /

Aluminum 1
Cylinder

 

 

\____/

Figure 4.3. Two-dimensional schematic of the 90° ﬂexed knee after dissection and
potting in epoxy-ﬁlled aluminum cylinders.
in the ﬁxture. The femur was not allowed to rotate axially, but axial rotation of the tibia
was allowed via the horizontal hydraulic actuator of the testing machine (Figure 4).
Figure 4 shows how the impact loads were applied simultaneously in two
directions with two servo-controlled hydraulic actuators that were oriented 90° to each
other. This biaxial testing machine consisted of a vertically oriented 5.5 kip actuator
(model #20452, MTS Corp., Eden Prairie, MN) on a 22 kip (model #31221, MT S Corp.)
frame, and a horizontally oriented 11 kip actuator (model # 204.61, MTS Corp.) mounted
to a custom designed frame. The actuators had separate electronic controls (model #4582
Microconsole for the vertical actuator and a #44882 Test Controller for the horizontal

actuator, MTS Corp.). The actuators were programmed to run simultaneously with a

81

 

 

 

 

 

 

 

 

 

 

 

 

 

Vertical Actuator ——>
Accelerometers
Horizontal Actuator
Load Cells \5
Deformable
Interface

 

 

 

    
    

Universal Joint

able

Custom
Table

Figure 4.4. Biaxial impact machine and test set-up for simultaneous AKL and ATL
impacts to the 90° ﬂexed knee.

common source waveform generator (model #45891 Microproﬁler, MT S Corp.) that

generated a haversine load waveform input with a 50 ms time to peak. Prior to each test a

preload of approximately 50 N was applied to the joint in both actuators for biaxial

impacts and only in the vertical actuator for uniaxial impacts.

82

A repetitive loading protocol was followed based on Table l, with the
programmed peak loads increasing for each impact until a gross failure was visualized in
the joint. Specimen pairs followed identical impact protocols up to test 6. After test 6, one
knee was randomly chosen from each pair to be loaded uniaxially, while the opposite
knee was loaded with a sequence of biaxial impacts. The uniaxial knee from each pair
was subjected to impacts consisting of only an AKL, which was increased by increments
of 3 kN until joint failure. The biaxial knee was also impacted with an AKL increasing

by 3 kN increments after test 6, however these impacts were in combination with an ATL

 

 

 

 

 

 

 

 

 

 

 

of 4 kN.
AKL (kN) ATL (kN) 1st knee ATL (kN) 2“d knee
Test 1 3 0 Same
Test 2 3 Preload Same
Test 3 3 2 Same
Test 4 6 0 Same
Test 5 6 2 Same
Test 6 6 4 Same
Test 7 9 0 4
Test 8 12 0 4
Test 9 15 0 4
Test 10 18 0 4

 

 

 

 

 

 

Table 4.1. Impact sequence for biaxial knee tests.
Some specimens were loaded with slightly different impact protocols, as
described next. The ﬁrst two pairs of knees in the study did not follow the pattern
established by Table 1 for test 1-6. Instead, the uniaxial knee from each pair was

subjected to only an AKL, which started at 3 kN and increased by increments of 3 kN

83

until failure. The biaxial knee from each pair was also impacted with a 3 kN incremental
AKL, but these were combined with an ATL which was increased by 2 kN increments up
to 6 kN. Four additional knees were also loaded normally through test 6, however an
additional impact with 6 kN of AKL and 6 kN of ATL was also applied. The biaxial knee
in each of these pairs was loaded with 6 kN of ATL until failure. Finally, there were three
supplementary knees that did not follow the normal protocol, but failure level data was
documented and reported here.

All AKL impacts were delivered with a stiff but deformable interface (3.3 MP8
crush strength aluminum honeycomb #CR [[1 5056, Hexcel Corp., Stamford, CT). The
interface was attached to a 15 cm by 15 cm by 1.25 cm thick steel plate. The honeycomb
was not precrushed. A 5 kip (thousands of pounds) load transducer (model #3210AF-5K,
Interface, Scottsdale, AZ) was attached to the actuator behind the knee impact interface.
ATL was applied to the tibia/ﬁbula cylinder via the horizontal actuator through a
universal joint. An 18 kip load transducer (model # FFL (18/112)u-(3/12)sp, Strainsert
Co., West Conshohocken, PA) was attached to the actuator behind this universal joint.
Load data from both transducers were inertially compensated with accelerometers (model
#353B15,PCB Piezoelectronics, Depew, NY). A LVDT (model #1002 XZ-D, Shaevitz,
Fairﬁeld, NJ) was attached to the base of the femoral potting cylinder and measured the
posterior displacement (drawer) of the tibia relative to the femur. Four channels of data
(load and displacement) from the testing machine, two channels of compensated load
data, and one channel of displacement data from the LVDT were collected at 1000 Hz
and recorded on a personal computer with a 16-bit analog/digital board (model DAS

1600; Computer Boards, Mansﬁeld, MA).

84

Prior to each sequential test, pressure sensitive film (Prescale; Fuji Film Ltd.,
Tokyo, Japan) was manually inserted under the quadriceps tendon into the patello-
femoral joint to measure the magnitude and distribution of contact pressures generated in
the joint during impact. The pressure ﬁlm was placed to ensure that it covered the entire
retropatellar surface. Low (0-10 MP8) and medium (10-50 MP8) range pressure ﬁlms
were stacked together and sealed between two sheets of polyethylene (0.04 mm thick) to
prevent exposure of the ﬁlm to body ﬂuids (Atkinson et al., 1998). A new pressure ﬁlm
packet was placed under the quadriceps tendon and into the patello-femoral joint
immediately prior to each test in the sequence. The ﬁlm was later removed from the
sealed polyethylene packet and digitally scanned at 150 dpi (ScanMaker E6, Microtek
International Inc., Redondo Beach, CA). The ﬁlm was converted to grey scale values
using commercial software (Scion Image 4.0.2; Scion Corp, Frederick, MD). Calibration
tests were performed on similar ﬁlm packets prior to the cadaver tests. Brieﬂy, a
haversine, displacement-controlled waveform was used to generate calibration peak loads
in approximately 50 ms. The calibration ﬁlm packets were loaded between two polished
stainless steel plates. This provided a dynamic calibration for the pressure ﬁlm over the
range of loads used in the current study. As described by Atkinson et al. (1998), the
stacking order of the ﬁlm and the rate of loading are important factors for accurate
calibration of joint contact pressures.

After each impact in the repetitive series of tests, the specimen were examined for
gross fracture of bone by visual inspection of the entire knee joint and palpation of the
patello-femoral joint from under the quadriceps tendon. Ligaments were also inspected

after each impact by a manual laxity evaluation performed on the joint after each test.

85

Following the failure test, each joint was carefully dissected and all injuries were
documented photographically. After potting the knees in epoxy, but before impacting,
medial-lateral radiographs were obtained with 90° of knee ﬂexion. Subsequently, Figure
5 documents how these radiographs were used to approximate the anterior-posterior slope

angle of the tibial plateau in each specimen.

 

Figure 4.5. Lateral radiographic image of the 90° ﬂexed knee showing the anterior-
posterior tibial plateau slope angle.
Additionally, ﬁve knees (from 3 subjects, average age=65) were scanned by computed
tomography (Light Speed CT. 4, General Electric) at 1.25 mm resolution, with a 50%
overlap for increased fracture imaging capability. The knees chosen for CT evaluation
were from the younger specimens that appeared to have high tissue quality. CT scanning
was done prior to impacting. CT scans were also taken following impacts in cases where
load-displacement data appeared to suggest a ligamentous failure that was not obvious
with gross examination of the joint. The focus of the CT scan was to search for occult

microscopic fractures of bone in the tibial plateau or femur. Finally, one knee joint from

86

each pair was also scanned by dual energy X—ray absorptiometry (DEXA, QDR 4500
Hologic Inc., Bedford, MA) to measure the average bone mineral density for the entire
knee (Figure 6). The femoral condyles and tibial plateau of each specimen were scanned

in an anterior-posterior direction using an established technique (Murphy et al., 2001).

 

Figure 4.6. Knee joint BMD measurement region in a DEXA scan from a representative
specimen. The scan area is represented by the trace within the thick rectangle.

The deformable impact interface (Hexcel) from each test and shown in Figure 7
was digitally scanned at 300 dpi. The areas of patella and tibial tuberosity contact were
measured using image analysis software (SigmaScan, Systan Software Inc., Point
Richmond, CA.). The percentage of tibia contact area to patella contact area was then
computed and served as a reference to compute load sharing between the two bone
surfaces.

The results (mean 1 standard deviation) of peak load and peak displacement from

the horizontal and vertical sensors, and the tibial drawer from the LVDT were

87

documented for each test in the study. Peak loads, displacements and stiffnesses for
various test conﬁgurations were compared with one-way, repeated measure ANOVAs.
The percentage of interface contact area on the tibia versus patella was compared (mean,
range) between uniaxial and biaxial specimens with paired t-tests. Specimen age, bone
mineral density and tibia slope angle were compared with unpaired t-tests. Multi-variate
linear regressions were developed to show the relationship between tibial drawer or
patello-femoral contact force and the applied loads (Sigma Stat v2.03, SPSS Inc.,

Chicago, IL). Statistical significance in all tests was set at p<0.05.

 

Patello-femoral contact Tibial tuberosity contact

Figure 4.7. Hexcel deformation from a representative specimen (31382R6).

RESULTS
Sub-Injury Tests

The output load-time response curves for simultaneous AKL and ATL impacts
showed similar loading and unloading regions as Figure 8 with an average time to peak
load of approximately 50 ms. Figure 9 documents the overall shape for all load-

displacement responses during the loading portion of the tests was linear with r2 values

88

7000 —

6000 ~ _AKL
—ATL

 

 

o l l I l
0 0.05 0.1 0.15 0.2
Time (sec)

Figure 4.8. Load versus time curves for test 5 from a representative specimen.
3500 _
3000 _
2500 -

2000-

Load (N)

1500-

1000-

 

 

 

0 l T l I ‘1
0 1 2 3 4 5

Displacement (mm)

Figure 4.9. Load versus displacement loading curves for test 3.

89

above 0.95. The impact response of the isolated knee in biaxial tests produced an
increasing stiffness with increasing ATL in the various sub-failure AKL tests.
Comparisons were made between tests with similar AKL levels (tests 1-3, average AKL
of 2.6102 kN [Table 2] and tests 4-6, average AKL of 5.6102 kN [Table 3]). Figure 10
shows that the AKL impact load-displacement response of test 3 (8311222 N/mm), the
impact with the highest ATL in this group, was signiﬁcantly stiffer than test 1 (6591208
N/mm, p=0.002) and test 2 (7161222 N/mm, p=0.009). The ATL stiffness in test 3 was
8601188 N/mm, but could not be calculated for tests 1 and 2 because no load was
delivered in that direction. In the higher group, the AKL stiffness of test 6 (9211193
N/mm) was signiﬁcantly stiffer than test 4 (7861155 N/mm, p=0.002), but not test 5
(8501188 N/mm, p=0123). In this group there was also a signiﬁcant difference between
test 4 and 5 (p=0.018). However, there was no signiﬁcant difference between the ATL

stiffness responses for test 5 (12991333 N/mm) and test 6 (13141284 N/mm).

 

 

 

 

 

 

 

 

 

 

 

IOWATL

14004
E IR'ObadATL
E1200- l2kNA‘n.
Z n4kNATL T
”10004
m
0
5 a...
1‘5

500.
8
c
X 4001
.§
5 2°°‘
c
< o. .

3kN AKL 6kN AKL 9kN AKL

Figure 4.10. Bar graph of the anterior knee stiffness with increasing ATL at various AKL

levels. * Signiﬁcantly different from uniaxial impacts.

9O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

885191 Tel mm Malﬁja T'n'a mam
lD NrrtH Lcm(N 118307") 81111855081111 Loam 056m) 811mm Dammr) Faoeaxl
313878 1 1735 629 38194 0 0 5:2 223
313371. 2 223 692 2289 634 1.14 583 1577
312378 1 2445 621 377.44 0 0 1.6 3418
31237L 2 2154 312 77807 734 087 259 2115
am 1 2297 535 42951 0 o 313 323
2 2465 515 477.65 32 046 235 3709
3 212 478 5899 1427 1.95 66478 269 2972
382. 1 2437 589 41652 0 0 459 422
2 2485 585 41449 564 061 4:6 283
3 2619 541 48515 1631 1.86 ass 3:5 354
31328 1 2718 315 93414 0 0 2.83 3344
2 2781 284 1071.7 276 025 26 3228
3 284 261 11487 1583 1.7 91887 228 372
3132. 1 an 345 8125 0 0 1.8 2544
2 2706 383 78578 255 025 1.95 2619
3 2736 342 82076 1816 1.72 1595 1.85 3153
31403. 1 2783 305 $45 0 0 012 401)
2 2794 273 11101 22 025 1.4 319
3 226 249 12205 1313 205 831.06 289 :22
31488 1 221 393 70336 0 0 201 4377
2 225 332 92173 :27 045 3:2 322
3 221 29 1029 1478 1.93 67307 332 323
313783 1 223 441 58414 0 0 3(5 324
2 2661 41 82 283 012 234 3199
3 272 374 73134 283 157 10378 23 312
31378. 1 22) 335 81501 0 0 223 342
2 299 36 772.8 294 05 218 323
3 2730 315 92378 212 1.46 11887 1.75 3548
31:23. 1 241) 4:2 56901 0 0 245 341)
2 2483 434 51255 a) 0:2 216 8584
3 2587 387 827 1478 2 67012 28 321
31881 1 2609 413 651.12 0 0 303 482
2 229 am 71327 519 083 312 4344
_ 3 2785 343 8284 1583 1.72 822 245 4184
313883 1 2823 475 58118 0 0 392 288
2 282 426 88151 32 02 319 2441
3 2789 374 5441 1931 137 247 2509
31:81 1 218) 419 631.97 0 0 235 3123
2 283 41 8545 22 02 128 382
_ 3 m1 387 74219 182 1.75 1.45 116
31324 1 278) 2m $124 0 0 183 :82
2 2831 31 98526 :E 0:2 1.75 3510
3 229 27 1031.9 1815 1.2 10254 1.0 2940
31:20 1 212 378 71824 0 0 are 3911
2 545 389 72643 270 025 3:2 3427
3 2718 361 7.261 1718 1.79 71816 218 3463
Anagram 1 537237) 43(11) 82(28): 0(0 0(0) - 28(1.1) 3507(82
2 2616(141) 41 (1.1) 714(22) :m(151) 02(05) - 27(12) 3174003
3 2740(8) 35(08) 831% 182(243) 1.7(02 mum 23(06) 322048

 

 

 

 

 

 

 

Table 4.2. Load, displacement, stiffness and pressure ﬁlm force for tests 1-3, 3 kN AKL

with varying ATL. * Signiﬁcantly different than test 3.

91

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

38mm Test 771818141195 Mal Tb'a . 7158 91658758111
ID Nnber Loed<N 09901111 W Loam 099mm SiﬁressW DaA_er(rmi FacetN
31387Fl 4 5571 808 67471 0 o 7.34 8919
313871. 5 248 891 6054 192 1.54 11036 3376
812378 4 521 844 64876 0 o 356 425
31237L 5 5457 788 71657 2170 22 7742 412 6172
6 5541 672 80864 580 52 5296 32 154
31329 4 5729 525 11138 0 o 495 422
5 5916 464 1281.8 128 123 14709 438 m
6 284 433 13725 3789 255 1381.8 423 408)
31392. 4 5651 6.2 8932 o o 388 475
5 298 589 98833 1917 151 11219 356 4942
6 207 591 98483 3908 277 12847 366 4814
31405 4 520 864 841.2 0 o 32 7326
5 4981 724 7002 1788 1.61 991.38 449 729
6 52 7.42 6942 3901 32 1025 481 712
314058 4 5549 831 2844 o o 549 241
5 5640 598 97549 1774 1.49 1071 52 242
6 207 52 1087.6 3785 301 11087 51 5775
313788 4 5711 727 78948 0 o 586 4844
5 5834 574 10065 2294 133 14493 44 3641
6 5970 814 2524 4418 22 15388 45 452
31378. 4 520 7.07 7882 o 0 52 5963
5 5708 874 521 2230 1.14 1587 42 241
6 5576 61 2325 442 249 15715 32 5919
3132. 4 222 844 2889 o o 854 209
5 524 7.44 71842 2194 1.34 1411.4 45 5813
6 582 7.08 78544 4100 269 1325 373
81332 4 5009 866 57812 0 0 654 5949
5 542 7.4 727.18 2.2 1.3 1422 68 424
6 524 67 2205 4:25 286 14054 572 m
3122 4 281 7.13 7276 o o 523 5912
5 249 884 83758 245 1.01 2565 487 5810
6 529 85 911.79 4817 227 173321 228 545
31m 4 5408 7.13 7229 o o 42 527
5 528 72 7842 2116 1.14 12424 303 25
6 534 62 823 4198 25 14314 34 492
312238 4 5641 52 10062 0 0 401 281
5 586 537 1087.1 2048 121 14784 346 274
6 267 45 1125 421 22 14747 322
31:20. 4 246 7.8 647.46 0 o 609 645
5 248 7.6 2287 151 1.7 90903 448 595
6 5490 7.42 7182 :273 301 11518 477_
4.3895(8) 4 —_‘*_542(22 7.1 (1.1) 788(12) —0(0) 0(0) - 52(12 ——5731(28
5 5521255) 6701) 230831? 212(234) 1.4(04) 122(3331 45(09 @7005)
6 $078) 62(09 21(198) 4083(284) 29(08) 1314284) 40(03) 291(1541

 

 

 

 

 

 

 

Table 4.3. Load, displacement, stiffness and pressure ﬁlm force for tests 4-6, 6 kN AKL
with varying ATL.

* Signiﬁcantly different than test 6. # Signiﬁcantly different than test 4.

92

The amount of posterior translation (drawer) of the tibia relative to the femur was
also measured in this study and is shown in Figure 11. There was a trend for decreasing
tibial drawer with increasing ATL at both levels of AKL. Tibial drawer at both levels of
AKL decreased an average of 0.3 mm per 1 kN of ATL. A multivariate linear regression
analysis of these data yielded the following expression:

Tibial drawer = 0.886 + 0.76 * AKL — 0.31 * ATL, (l)

where both AKL (kN) and ATL (kN) were determined to be signiﬁcant predictors of

 

 

 

 

 

 

 

 

 

 

tibial drawer (m) (Table 4).

0 Test 1 0 Test 4

I Test 2 )1: Test 5

3 Test 3 A Test 6

X 3 kN AKL 6 kN AKL
E ' — 195 A M
3 X XW is A
«3 A _
b A A
e ‘ ‘ Xxx A A
.2
a w
l: “ “ A A
A ‘ F = 0.031
2000 3000 4000 5000

 

Axial Tibia Load (N)

Figure 4.11. Tibial drawer displacement in tests 1-3 and 4-6 with increasing ATL.

 

 

 

 

 

Y-intercept AKL (kN) ATL (kN)
Independent Variable 0.886 0.76 -0.31
Standard Error 0.33 0.09 0.09
P-value <0.001 0.001

 

 

 

 

Table 4.4. Multiple linear regression variables for tibial drawer (Equation 1).

93

 

For test 7 the intent was that one knee from each pair was impacted uniaxially
with 9 kN of AKL, while the opposite knee was impacted biaxially with a 9kN AKL and
a 4 kN ATL. Statistically, the mean values of peak load and displacement for each
protocol were not different (Table 5). Thus, the average peak AKL and displacement
were 8.2i0.4 kN and 9.6il.4 mm, respectively for the combined uniaxially and biaxially
impacted knees. The stiffness responses of each group were also not statistically different
(p=0.097), but on average the biaxial impacted knees tended to be stiffer (997.8i146
N/mm) than the uniaxial impacted knees (831.2i146 N/mm). There was, on the other
hand, a statistically signiﬁcant difference between the peak tibial drawers for the uniaxial

(7.4il.4 mm) versus the biaxial (5.81:0.6 mm, p=0.017) impacted knees.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ Specirmn AnteriorKnee Axial ﬁbial Tibia ‘
ID Load (N) Disp(mm) Stiffn885(N/mm) Load(N) Disp (mm) Drawer(mm)
31387F1 8335 8.55 1025.9 0 0 7.03
31387L 7971 9.66 912.46 3157 2.09 5.93
31237R 8141 12.02 679.44 0 0 5.49
31392L 8570 8.48 1012.5 0 0 6.45
31378L 8353 9.8 837.37 0 0 8.37
31378R 8763 7.82 1151 4846 2.24 6.28
31383R 7235 11.78 653.14 0 0 9.4
31383L 8330 10.15 793.67 4529 2.33 6.44
31382L 8198 9.95 831.94 0 0 6.42
31382R 8517 8.54 1023.6 4853 1.87 5.67
31390L 7821 10.43 777.81 0 0 8.52 ,,
31390R 8342 7.9 1108.4 4511 2.3 4.97
OATL Ave (SD) 8093(290) 10.1 (1.4) 831.2 (146.3) 0(0) 0(0) 7.4(1.4)
EKN ATL Ave (SD)8384(290) 8.8 (1.0) 997.8 (146.0) 4279(654) 2.2 (0.2) 5.8 (0.6)

 

Table 4.5. Load, displacement and stiffness for test 7, 9 kN AKL with either 0 kN or 4
kN ATL in paired specimen. * Signiﬁcantly different than biaxial impacts (4 kN ATL).
Contact Interface Deformation
The contact areas in the deformable interface were measured to estimate the

relative loads applied to the PF joint and to the tibial tuberosity during each test (Figure

94

7). Figure 12 shows a signiﬁcant difference in the percentage of load carried by the tibial
tuberosity in biaxial tests (6.9%, range 0-36%) versus the uniaxial tests (2.1%, range 0-
15%, p=0.034) for an AKL of 6 kN. At the 3 kN and 9 kN levels of AKL the percentages
of tuberosity load in biaxial versus uniaxial tests were 4.9% (O-36%) versus 2.1% (0—
19%) and 12.7% (O-33%) versus 8.7% (0-25%) respectively. However, these differences

were not statistically signiﬁcant.

m

304

    

20-

lelal Tuberoslty Load (%)

 

I.. ..ll.. ﬁll. i"l

Anterior Knee Load

Figure 4.12. Bar graphs of the percentage of load carried by the tibial tuberosity for

impacts at various AKL levels. * Signiﬁcantly different from uniaxial impacts.

Pressure Film Force

The contact pressures and contact areas within the PF joint were documented

using Fuji ﬁlm in tests 1-6 on each specimen. Generally, the pressure distributions were

95

smooth with little or no apparent creasing of the ﬁlm as shown by Figure 13. The PF

force was determined by the merging data from the low and medium pressure ﬁlms

 

 

 

 

 

 

 

Distal
KIT 5.“ “
Lateral ’* '79 Medial
Proximal
Low Pressure Film Medium Pressure Film

 

 

“O C.) ,1.

 

 

 

 

 

 

35.. t 13.,

‘3', I." 5..“

 

 

 

Figure 4.13. Pressure ﬁlm from a representative specimen (31382L6).
(Tables 2 and 3). Figure 14 documents the trend for decreasing PF contact force with
increasing ATL at both levels of AKL, which can also be shown by a multiple linear
regression:
PF Force = 1.43 + 0.77 * AKL — 0.18 * ATL, (2)
where both AKL (kN) and ATL (kN) were determined to be signiﬁcant predictors of PF

contact force (kN) (Table 6):

96

 

 

 

 

 

 

 

 

 

 

o Test1 I Test2
8000 A Tests o Test4
)K Tests A Test6
7000 x 3kN AKL ~ 6kNAKL
E A
'3 6°00 ”(W x r2=_0.039 A AA
g 4000 ‘ A
".- 3000 A i A ‘ r2=0.(T28
.9. A
2 20004 I A
9 I
0..
10004 4
o I I I I I
0 1000 2000 3000 4000 5000

Axial Tibia Load (N)

Figure 4.14. Patello-femoral force from pressure film in tests 1-3 and 4-6 with increasing

axial tibia load.

 

 

 

 

 

Y-intercept AKL (kN) ATL (kN)
Independent Variable 1.43 0.77 -0.18
Standard Error 0.29 0.08 0.08
P-value <0.001 0.03

 

 

 

 

Table 4.6. Multiple linear regression variables for PF force (Equation 2).
Injury Tests

Gross injuries, seen in Figure 15, were in the form of a ligament tear or a bone
fracture and occurred in or after test 8. Ten knee impacts resulted in a complete tear,
partial tear or avulsion of the PCL at an average maximum AKL of 12.4:t2.6 kN and

displacement of 19.7:t6.6 m (Table 5). Among the knees that had PCL trauma, there

was not a statistically signiﬁcant difference between the failure AKL of biaxial (12.7i2.4

97

 

(A) PCL avulsion fracture from posterior (B) Split femoral condyle fracture
tibia

   

Figure 4.15. Representative photographs of injuries to paired human knees (31390).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Specimen Sex Ag- BMD Tibial Plato-u Anterior Knee Axial Tibia Tibia Injury 0
ID (91011142) Ande (dag) Load (N) Disp (mm) Load (N) 0109 (mm) anor (mm)
3130711 M 71 0.010 13 13990 25.52 0 0 19.92 PCL Mia-m T-r
31307L M 71 11 13500 15.01 4052 529 14.40 PCL Avudon
3123714 M 47 14 13031 19.43 0 o 17.9 PCL Ava-1m
31237L M 47 0.550 9.5 13210 23.19 5230 4.00 20.9 PCL MW Tour
m F 92 0148 10 3342 18.49 0 0 16.46 Split Femoru 0001M. Fracture
300990 F 92 7 3100 14.59 095 212 9.24 Split Fm Comm Frann-
31392L M 59 o 031 0 12040 17.41 0 0 10.07 F0110 PCL Tau. Patella Fm
110131 Hm Fraeum. 241011.
3139211 M 59 11 0107 0.70 4243 4.33 0.43 I 1 1
SplltF MWF mum.
3140551 14 01 5 7920 1575 o o 15.0 Pm“. I I
314051. M 01 0.344 7 7742 10.01 3004 000 0.0 “w 1110:0311), "‘1 "m
m
313101. F 01 0.714 0.5 11231 25.0 o 0 24.34 50111 Femoral Condylo Fracture
59111 FomorllCoero men.
3137011 F 01 0 11279 14.1 2945 3.59 7.20 ”I“ 7101.1 21 Fm”.
Split 00111171100101”= moralConrMo
31% M 09 7 10502 40 o 0 32 F 10.2 Invade"
313031. M 09 0.071 7 0090 21.4 3900 217 14.7 59111 Femoral Condyio Frlcmn
31302L F 50 0.5 13203 37.9 0 o 34.5 59'" °°'“"“"”‘°“ If?“ M"
313020 F 50 0.795 10.5 10040 12.4 3024 1.09 9.40 P'm' PCL FT‘“ 1101.1 m
Split. Cornminuted Femoral Condyla
313901. F 70 0954 29 5 o 0 31.17 chmm
313900 F 70 0.500 0 10003 12.47 4430 2.11 9.90 POL Avulsion. F1001. Fm
-< m M 70 0.737 - 0790 24.4 o o . PCL W Tear
03044L M 57 - 14374 32.0 o 0 - Fc1. W100 roar
0304411 M 57 0.754 - 10000 14.1 5000 1.42 131 Fun-1 PCL Tour
I. 4 4 5 _
PCL Ave. (50) 02.1 (11.3) 0.721 (0.11) 10.3 (2.0) 12395 (2010) 19.7 (00) 2243 (2407) 1.5 (1.9) 152 (4.3) 1410mm “L AW”
Fx Ave. (s0) 70.3(12. 9) 0.459(020) 8011. 9) 0050(3093) 206(4 5) 16150873) ”(2‘) "-79” 1101. WWW” Femoral

 

Table 4. 7. Specimen properties, load, displacement, and injury comments for failure

tests. * Signiﬁcantly different than fractured specimens.

98

kN) versus uniaxial (120:3.1 kN) impacted knees. However, the peak displacement of
the AKL actuator was signiﬁcantly less for PCL failure in knees with an ATL (15.6i4.5
mm) versus without an ATL (23.9i5.9mm, p=0.018). Additionally, the peak tibial drawer
decreased from 18.0i1.9 mm without ATL to l3.6i4.6 mm with an ATL in cases of PCL
trauma, but this difference did not rise to the level of statistical signiﬁcance. The
remaining knees (n=1 1) had gross bone fracture. These were in the form of either a split
femoral condyle fracture or a tibial plateau fracture. In four cases these injuries were
combined with an avulsion fracture of the patellar tendon from the distal patella. The
average maximum AKL generated in the fractured specimens was signiﬁcantly lower
(89223.1 kN, p=0.004) than that for the PCL injured specimens. It should be noted that
specimens that showed bone fracture had an average age of 78.3:129 years, while the
PCL injured specimens had a statistically different average age of 62.1ill.3 years
(p=0.001). This may have also been related to bone mineral density, which was
signiﬁcantly higher for PCL injured specimens (072110.11 g/cmz) than bone fractured
specimens (045910.26 g/cmz, p=0.036) (Table 7). Additionally, the tibial plateau slope
angle was 8.0i1.9 degrees for bone fractured specimens, while it was signiﬁcantly
different for PCL injured specimens, being 10.3i2.8 degrees (p=0.016).

The peak tibial drawer for all failure tests was signiﬁcantly lower (1 l.6:l:4.2 mm)
for biaxially loaded knees than for uniaxial knees (23.1:7.6 mm, p=0.002). Additionally,
the AKL actuator displacement at peak load was lower for biaxial loaded knees (14.5:t4.9
mm) than uniaxial loaded knees (26.1:82 mm, p=0.003). These data follow similar
trends from tests 1-6 showing an increased joint stiffness in biaxial versus uniaxial

impacts on the knee. Analysis of the interface deformation revealed that biaxially loaded

99

knees carried 12.7% (range 0-33%) of the AKL on the tibial tuberosity. In contrast, 7.7%
(0-24%) of the AKL was carried by the tibial tuberosity for uniaxial impacts (p=0.097).
Since the AKL was similar for both groups (103:3.6 kN biaxially and 10533.4 kN
uniaxially) in the failure tests, the load generated across the PF joint in these tests was
likely reduced for biaxial versus uniaxial tests. There was a difﬁculty conﬁrming these
data from Fuji ﬁlm in the PF joint because bone fractures within the knee appeared to

create signiﬁcant artifact (Figure 16).

 
   

 

 

 

 

Lateral L ' Medial
1 .3 .ﬂf'fm
_ Proximal _ _
Low Pressure Film Medium Pressure Film

Figure 4.16. Pressure ﬁlm for a fractured specimen (31382L).

Computed Tomography

Computed tomography (CT) imaging of ﬁve knees revealed two cases of
particular interest. In both specimens CT scans were obtained before and after the impact
sequence, but prior to dissection of the knee. Specimen 31390R had a gross injury
consisting of a PCL avulsion which was observed through a laxity test between the tibia
and femur (Figure 15A). A post-test CT scan, shown in Figure 17, documented occult
microfractures in the subchondral bone of the anterior compartment of the tibial plateau

in this specimen. While the gross injuries for specimen 31382R included a partial PCL

(A) Prior to Impact

       

(B) Following Impact

Figure 4.17. CT scan of specimen 31390R prior to and following impacting,
documenting occult microfractures in the anterior compartment of the tibial plateau.
tear and a comminuted tibial plateau fracture shown in Figure 18, a CT scan conducted
prior to the last impact on the joint indicated occult microfractures in the tibial plateau
(Figure 19). The area of these occult microfractures was the location of gross bone

fracture in the tibial plateau following the last impact.

 

Figure 4.18. Photograph of specimen 31382R during dissection, showing gross fracture

of the tibial plateau.

lOl

Prior to Impacting Prior to Final Impact Following Impacting

(a...

3

    

Figure 4.19. CT scans of specimen 31382R documenting occult microfractures

progressing to gross fracture of the anterior tibial plateau.

DISCUSSION

The object of this study was to document relationships between various
combinations of anterior knee loads and axial loads in the tibia of the 90° ﬂexed human
knee and the resulting knee joint stiffness, patello-femoral contact forces and injury
patterns. Hypothesis (1) of the study was that ATL present during an AKL blunt impact
by a deformable interface would result in less posterior translation of the tibia and stiffen
the knee response during contact with this interface. In the current study the posterior
tibial translation increased with an increase in applied AKL. On the other hand, tibial
translation decreased with an increase in ATL at a rate of 0.3 mm per kN of ATL (Figure
11). While both load variables were shown to be signiﬁcant predictors of tibial
translation, it should be noted that there was a high level of variation between samples
and therefore 1'2 values were low. This variation was likely caused by other variables not

considered in the multiple linear regression, such as age, BMD and anatomical geometry.

102

The stiffness of the knee for an AKL impact was increased between 17% and 26% when
an ATL was applied simultaneously (Table 2, 3 and 5). The effect was likely due to an
anterior directed force component that tended to displace the tibia anteriorly under an
ATL. This was largely suggested due to an average tibial plateau slope of approximately
9° among the specimens (Figure 5). Table 7 shows individual specimen tibial plateau
slope values. Previous measurements based on lateral radiographs have shown the tibial
slope of the average knee to be 10:3° (DeJour et al., 1994, Gevien et a1, 1993, Gifﬁn et
al., 2004). Hypothesis (2) of the study was that constraint of the knee with a deformable
interface would prevent ACL rupture due to anterior translation of the tibia for an ATL
previously documented to injure the ACL in an unconstrained knee. While a previous
study documented rupture of the ACL with 58:22.9 kN of ATL (Jayaraman et al., 2001),
no such injuries were documented in the current study for an ATL of as much as 6 kN.
Hypothesis (3) of the study was that application of an ATL during blunt knee impacts
involving the tibial tuberosity would help reduce the patello-femoral joint force due to
increased load supported by the tibial tuberosity. The current study showed that for an
AKL of 6 RN and an ATL of 4 kN the PF contact force was reduced by approximately
11.2% versus specimens with no ATL. And lastly, hypothesis (4) of the current study was
that the AKL needed to cause rupture of the PCL would be increased due to an anteriorly
directed force component generated by an ATL. In specimens that had PCL trauma, the
average maximum AKL was 12.7 kN for biaxial versus 12.0 kN for uniaxial tests, but
these AKL’s were not statistically signiﬁcant in the study.

Knees were impacted at 90° of ﬂexion to simulate the ﬂexion angle during

contact with the IP in simulated car crashes (Atkinson et al., 1999). Additionally, this

103

ﬂexion angle facilitates comparisons with previous experiments using the 90° flexed
human cadaver knee. AKLs were directed parallel with the axis of the femur to simplify
the bending moments in the bones. However, while this loading scenario is comparable to
a majority of previous studies, it may not simulate all occupant orientations in real world
frontal crashes, due to the possibility of abduction or adduction of the hip (Meyer and
Haut, 2003 and Rupp et al., 2003).

Early impacts on the PF joint with a rigid impact interface (Haut, 1989)
document split femoral condyles and transverse and comminuted fractures of the patella
in the knees of 9 human subjects aged 72:11 years at a maximum contact force of
8.5130 kN. The maximum load was increased to 9.5140 kN when a 25 mm thick foam
(Rub-A-Tex R310V) was applied to the interface. Using a similar experimental set-up
and impact protocol with a rigid interface, a later study documented femoral condyle and
patella fractures at approximately 6.73.1.7 kN (n=10) and 6712.1 kN (n=10) for two
groups of specimens aged 48:9 years and 71.7 years, respectively (Atkinson and Haut,
1995). In the current study bone fractures (split femoral condyle, n=9 or tibial plateau,
nﬂ) occurred at a maximum AKL of 8.913.] kN (n=11) in a group of specimens aged
78.3:129 years. In contrast, the bone fracture loads presented in the current study were,
on average, slightly lower that those documented for blunt patellar impacts using a 25
mm thick foam (Haut, 1989). The effect may be due to yet unknown differences in the
impact characteristics of the Hexcel (pressure limiting) and foam interfaces used in each
study and contact of this interface with the tibial tuberosity in the current study.

The characteristics of padded impact interfaces against the 90° flexed knee have

indicated the bone fracture limiting nature of various interfaces (Hayashi et al., 1996). In

104

that study 10 isolated human knees were ﬂexed 90° and impacted by a 20 kg pendulum
with a rigid surface, a 3.1 MPa and a 0.7 MPa crush strength aluminum honeycomb
padding (Hexcel), and a 0.35 MPa crush strength paper honeycomb. Patella and split
femoral condyle fractures were reported with the rigid interface at 17.4 kN. A split
femoral condyle fracture was reported at 15.5 kN for the stiffest, deformable interface.
Subchondral injury of the patella and intercondylar regions of the femur were
documented in 4 knee joints at 10110.3 kN (subject age 61 and 68 years) for the 0.7 MPa
interface. Based on analysis from a 3-D ﬁnite element human knee model and the
experimental data of the study, the optimal interface for protecting the knee against bone
fracture was hypothesized to be 0.62 MPa. No ATL however, was included in the
analysis. The current study would suggest that ATLs could help unload the PF joint to
possibly limit patella and split condylar fractures of the femur. Irnportantly, the Hayashi
et al. (1996) study did not document any injuries to the PCL, in contrast to the current
study.

A study by Herring and Patrick (1977) impacted the anterior knee similarly to the
current study, but used relatively low forces and compliant interfaces and did not
document fracture or PCL trauma in any tests. In these tests loads were measured
biaxially in both the femur and in the distal tibia. Interestingly, for an AKL above a
signiﬁcant load component was generated along the axis of the tibia in the cadaver. For
an AKL of approximately 4 kN there was a resultant ATL of 536186 N. This compares
well to data in the current study for test 2 where a slightly lower AKL of approximately
2.6 kN produced a resultant ATL of 3881151 N, when the tibia was constrained with an

intial preload of 50 N. The previous study also documented that the amount of force

105

 

required to constrain the tibia in Part 572 dummy knee impacts was unexpectedly lower
(174131 N) than in the cadaver. This indicates that the Part 572 dummy knees were not
accurately predicting the response of human knees to AKLs.

The maximum ATL applied in the current study was 6 kN. However, after this
load level caused tibial plateau fracture in two biaxially impacted specimens (31392R and
31405L), the maximum ATL was reduced to 4 kN or 50% of the average fracture load
documented in the Banglmaier et al. (1999) study. Initially, 6 kN of ATL was sought, to
simulate load levels from Jayaraman et al. (2001) where ACL rupture was documented at
5.8 1 2.9 kN of ATL for an unconstrained TF joint. In this study, the authors document
that a force of 1.2 1 0.5 kN is required to constrain anterior drawer of the tibia. With a
static analysis of the knee (assuming a slope of 10° and an ATL of 6 kN) we anticipate
the maximum anterior resultant force to be approximately 1.1 kN. For biaxial impacts
with an ATL of 4 kN, as in the current, study we anticipate a resultant force of 0.71 kN
would available to counteract the AKL applied to the tibial tuberosity (Figure 20). In fact,
in failure experiments the difference between biaxial and uniaxial impacts for PCL
rupture was approximately 0.7 kN. While close to what we expected, the difference
between uniaxial and biaxial experiments was not statistical because of specimen

variations using 4 kN in the current study.

106

 
 
    

Tibio-Femoral
Compression

ATL

l
..I ,

 

 

 

 

Figure 4.20. Orientation of applied and resultant forces for a 90° ﬂexed human knee.

The current study also supports the earlier report by Viano et al. (1978) indicating
injury to the PCL in numerous cases of blunt impact to the 90° ﬂexed knee when contact
occurs with the tibial tuberosity. In this previous study two isolated human knee joint
preparations had PCL rupture for a peak load of 2.6106 kN and posterior directed tibial
drawer of 18.1122 mm. These isolated knee joint tolerance tests involved controlled
translation (subluxation) of the tibia across the femur at 90° ﬂexion until complete joint
failure. Combination PCL and tibial plateau fracture tests (n=5) yielded a peak force of
2.481054 kN and 22.01102 mm (cadaver age 68.8115.l years). In comparison to these
isolated joint tolerance tests, blunt impact tests in the same study with a heavily padded

interface (3.2 kN/cm) on the knees of seated human cadavers (63.7116 years) resulted in

107

various combinations of PCL tears, avulsions, and tibial plateau and intercondylar femur
fractures at 7.0-10.9 kN. In the current study with a 3.3 MPa aluminum honeycomb
interface, PCL injuries occurred in tests having a peak AKL of 12.013] kN without an
ATL (n=5, aged 62.11113 years). The reason for the differences between these studies is
unclear. The higher rate of loading in the Viano et al. (1978) study (26314.8 ms primary
pulse duration) would suggest that PCL injury may be via central (midsubstance) rupture
and at higher loads than for the current study (Crownshield et al., 1976, Noyes et al.,
1974 ). Yet, the predominant mode of PCL failure in the Viano et al. (1978) study using
seated cadavers was avulsion fractures of bone from the tibial plateau. Furthermore, any
inertial motions of the seated cadaver would have tended to increase failure loads for
PCL injury compared to the current study using an isolated knee preparation. For yet
unexplained reasons the differences in peak contact forces between these studies may
have related to the differences in the stiffness of the impact interfaces. Irnportantly, PCL
injury occurred at 18.1122 mm (n=2) of posterior drawer in the isolated knee joint
studies of Viano et al. (1978). This compared to 15214.3 mm (n=10) in the current study.

The current study also noted that the total knee bone density was signiﬁcantly
higher in subjects with PCL injury than a bone fracture. This total knee bone density
considers the density of the distal femur and the proximal tibia/ﬁbula (Figure 6). These
data suggest that younger versus older subjects, in general, may be more susceptible to
ligamentous injury. Although BMD values are particularly relevant in comparing fracture
to non-fracture type injuries within this study, it should be noted that the average bone

mineral densities of all the impacted specimens in this study were generally lower than

108

previous data on healthy, non-osteoporotic, female patients ages 21-48 years (Murphy et
al., 2001).

The natural progression of a PCL injury is the long-term development of
degenerative changes in the medial compartment and patello-femoral joint of the knee (Li
et al., 2002). Long-term degeneration of the knee after PCL injury occurs in 20-60% of
patients, even after posterior stability is improved by PCL reconstruction. One
explanation for the development of chronic disease in the knee may relate to the
appearance of occult microcracks in bone underlying articular cartilage (Johnson et al.,
2000). These compression injuries of the bone have been associated clinically with early
changes in the overlying articular cartilage reminiscent of the disease osteoarthritis
(Johnson et al., 1998, Vellet et al., 1991). A study by Rangger et al. (1998) indicates that
these occult fractures, which are seen in histological sections from patients in automobile
crashes, exist in areas identiﬁed by magnetic resonance imaging (MRI) as “bone bruises”.
In the current study occult microfractures were identiﬁed in the anterior compartment of
the tibial plateau by CT scans in some cases of PCL injury. Examples of automotive
injury cases show similar injuries (Sanders et al., 2000). The author documents cases of
partial and complete PCL rupture due to a dashboard impact and corresponding “bone
bruises” in the anterior compartment of the tibial plateau via MRI. These data may
suggest the potential for a chronic disease in cases where the PCL is excessively tensed,
or ruptured, following an automobile accident. A limitation of the current study, however,
is that the above concern cannot be investigated using the human cadaver subject.

The cadaver model also does not currently incorporate the potential effects of

muscle tensions that may exist in the lower extremity during an automobile crash (Figure

109

20). The potential combination effect of quadriceps and hamstrings muscle tensions on
stiffness and strength of the knee during blunt impact loadings needs further
investigations using cadaver subjects that incorporate various levels of simulated muscle
contractions. Muscle tensions may well show increased stiffness of the knee that could
help explain the relatively low frequencies of knee ligament injuries documented in the

NASS automotive accident databases.

CONCLUSION

The study showed that axial compressive loading of the tibia, which may occur at
nearly the same time and magnitude as anterior knee loads during a frontal automotive
crash, has a stiffening effect on the response of the knee. Furthermore, ATL can also act
to reduce the relative translation between the tibia and femur generated by contact
between a deformable interface and the tibial tuberosity for the 90° ﬂexed knee. The
effect of an ATL is also to carry more load in the tibial tuberosity, helping to unload the
patello-femoral contact forces. This effect can increase, if only slightly, the tolerance of
the knee to PCL injury. Current injury tolerance criteria reﬂect the vulnerability of the
knee to shear displacements by limiting tibia-femur translation to 15 mm. This level of
tibia drawer for PCI injury compared well to the current study. Finally, ACL injury was
not documented in any biaxial specimens, suggesting that instrument panel contact and
the generation of an anterior knee load will help constrain the tibia from moving forward

relative to the femur.

llO

ACKNOWLEDGEMENT

This study was supported by a grant from the Centers for Disease Control and
Prevention (R49/CCR503607). Its contents are solely the responsibility of the authors and
do not necessarily represent the ofﬁcial views of the Centers for Disease Control and
Prevention. The authors wish to gratefully acknowledge Cliff Beckett for technical
assistance in this study, Dr. Todd Fenton for providing the radiographs, Glenn Miller and
Alexis King for assistance with the CT scans and DEXA. We also thank Dean Mueller,
Coordinator of the Anatomical Donations Program, University of Michigan for procuring

human test specimens for this research project.

111

REFERENCES

Atkinson PJ, Haut RC, (1995) Subfracture insult to the human cadaver patellofemoral
joint produces occult injury. J Orthop Res 13:936-944.

Atkinson PJ, Garcia JJ, Altiero NJ, Haut RC. (1997) The inﬂuence of impact interface on
human knee injury: Implications for instrument panel design and the lower extremity
injury criterion, Stapp Conf Proc 41: 167-180.

Atkinson PJ, Newberry WN, Atkinson TS, Haut RC. (1998) A method to increase the
sensitive range of pressure sensitive ﬁlm. J Biomech 31: 855-859.

Atkinson PJ, Benny J, Sambatur K, Gudipaty K, Maripudi V, Hill T. (1999) A parametric
study of vehicle interior geometry, delta-v, and instrument panel stiffness on knee
injury and upper kinetic energy. Stapp Conf Proc 43: 203-215.

Atkinson T, Atkinson P. (2000) Knee injuries in motor vehicle collisions: A study of the
National Accident Sampling System database for the years 1979-1995. Accid Anal
Prev 32 (6): 786.

Atkinson P, Haut RC. (2001) Impact responses of the ﬂexed human knee using a
deformable impact interface. J Biomech Egr 123 (3): 205-211.

Banglmaier R, Dvoracek-Driksna D, Oniang’o T, Haut R. (1999) Axial compressive load
response of the 90° ﬂexed human tibiofemoral joint. Stapp Conf Proc 43: 127-139.

Bealle D, Johnson D. (2000) Subchondral contusion of the knee caused by axial loading
from dashboard impact. J Southern Orthop Assoc 9(1): 13-18.

lBedewi PG, Diggs KH. (1999) Investigating ankle injury mechanisms in offset frontal
collisions utilizing computational modeling and case study data. Stapp Car Crash
Conf 42: 217-242.

Blincoe L, et al. (2002) The economic impact of motor vehicle crashes 2000. NHTSA,
DOT HS 809 446.

Crowninshield RD, Pope MH. (1976) The strength and failure characteristics of rat
medial collateral ligaments. J Trauma 16299-105.

Daniel D et al. (1994) Fate of the ACL—injured patient: A prospective outcome study.
Am J Sports Med 22: 632-644.

Dejour H, Bonnin M. (1994) Tibial translation after anterior cruciate ligament rupture.
Two radiological tests compared. J Bone Joint Surg Br. 76 (5): 745-749.

Ewers BJ, J ayararnan V, Banglmaier R, Haut R. (2000) The effect of loading rate on the

degree of acute injury and chronic conditions in the knee after blunt impact. Stapp
Conf Proc 44: 299-3 14.

112

 

 

Fildes B, Lane L, Vulcan P, Seyer K. (1997) Lower limb injuries to passenger car
occupants. Accid Anal Prev 29: 789-795.

Fleming, B., Renstrom, P., Beynnon, B., and et al., 2001. The effect of weight-bearing
and external loading on anterior cruciate ligament strain. J Biomech 34: 163-170.

Funk JR, Tourret LI, Crandall JR. (2000) Experimentally produced tibial plateau
fractures. IRCOBI Conf: 171-182.

Genin P, Weill G, Julliard R. (1993) The tibial slope. Proposal for a measurement
method. J Radiol. 74: 27-33.

Gifﬁn J, Woo S, Hamer C. (2004) Effects of increasing tibial slope on the biomechanics
of the knee. Am J Sports Med 32 (2): 376-382.

Haut RC. (1989) Contact pressures in the patellofemoral joint during impact loading on
the ﬂexed human knee. J Orth Res 7: 272-280.

Hayashi S, Choi HY, Levin RS, Yang KH, King A1. (1996) Experimental and analytical
study of knee fracture mechanisms in a frontal knee impact. Stapp Car Crash J 40:
160-171.

Hirsch G, Sullivan L. (1965) Experimental knee-joint fractures. Acta Orthop Scand 36:
391-400.

Hering WE, Patrick LM. (1977) Response comparisons of the human cadaver knee and a
Part 572 dummy knee to impacts by crushable materials. Stapp Car Crash Conf 21:
1017-1053.

Jayaraman, V.M., Sevensma, E.T., Kitagawa, M., Haut, R.C., 2001. Effects of anterior-
posterior constraint on injury patterns in the human knee during tibio-femoral joint

loading from axial forces through the tibia. Stapp Car Crash J. 45: 449—468.

Johnson D, Bealle D, Brand J, Cabom D. (2000) The effect of a geographic lateral bone
bruise on knee inﬂammation after acute anterior cruciate ligament rupture. J
Orth Soc for Sports Med 28: 152-154.

Johnson DL, Urban WP, Caborn DN, Vanarthos WJ, Carlson CS. (1998) Articular
cartilage changes seen with magnetic resonance imaging-detected bone bruises

associated with acute anterior cruciate ligament rupture. Am J Sports Med 26 (3):
409-414

Kuppa S. (2002) An overview of knee-thigh-hip injuries in frontal crashes in the United
States. 17‘h ESV

Li, G., Rudy, T., Allen, C., 1998. Effect of combined axial compressive and anterior
tibial loads on in situ forces in the anterior cruciate ligament: a porcine study. J Orth
Res 16:122-127.

113

Li G, Gill T, Defrate L, Zarins B. (2002) Biomechanical consequences of PCL deﬁciency
in the knee under simulated muscle loads-an in vitro experimental study. J Orth Res
20: 887-892

Markolf, K.L., Bargar, W.L., Shoemaker, S.C., Amstutz, HQ, 1981. The role of joint
load in knee stability. J Bone Jt Surg. 63: 570-585.

Melvin J, Stalnaker R, Alem N, Benson J, Mohan D. (1975) Impact response and
tolerance of the lower extremities. Stapp Car Crash Conf Proc 19: 543-559.

Mertz H. (1993) Anthropomorphic Test Devices. Accidental Injury, Biomechanics and
Prevention edited by Nahum A, Melvin J. Springer-Verlag.

Meyer EG and Haut RC. (2003) The effect of impact angle on knee tolerance to rigid
impacts. Stapp Car Crash J 47: 1-19.

Miller M, Osborne J, Gordon W, Hinkin D, Brinker M. (1998) The natural history of
bone bruises: A prospective study of magnetic resonance imaging: detected trabecular

microfractures in patients with isolated medial collateral ligament injuries. Am J
Sports Med 26: 15-19.

Morgan RM, Eppinger RH, Marcus JH, Nichols H. (1990) Human cadaver and hybrid III
responses. International Conference on the Biomechanics of Impacts.

Murphy E, Bresnihan B, FitzGerald O. (2001) Validated measurement of periarticular

bone mineral density at the knee joint by dual energy x ray absorptiometry. Ann
Rheum Dis 60: 8-13.

Noyes FR, Delucas J, Torvik P. (1974) Biomechanics of anterior cruciate ligament
failure: An analysis of strain-rate sensitivity and mechanisms of failure in primates. J
Bone Jt Surg 56 (2): 236-253.

Patrick L, Kroell C, Mertz H. (1965) Forces on the human body in simulated crashes.
Stapp Car Crash Conf Proc 9: 237-259.

Powell W, Ojala S, Advani S, Martin R. (1975) Cadaver femur responses to longitudinal
impacts. Stapp Car Crash Conf Proc 19: 561-579.

Rangger C, Kreczy A. (1998) Bone bruise of the knee, Acta Orthop Scand 69 (3): 291—
294.

Rupp JD, Reed MP, Jeffreys TA, Schneider LW. (2003) Effects of hip posture on the
frontal impact tolerance of the human hip joint. Stapp Car Crash J 47: 2034.

Sanders T, Lawthom K. (2000) Bone contusion patterns of the knee at MR imaging:
Footprint of the mechanism of injury. J Radiographics 20: $135-$151.

114

States JD. (1986) Adult occupant injuries of the lower limb. Proc Symp Biomech: 97-
107.

Torzilli, P., Deng, X., Warren, R., 1994. The effect of joint-compression load and
quadriceps muscle force on knee motion in the intact and the anterior cruciate
ligament-sectioned knee. Am J Sports Med. 22: 105-112.

Vellet AD, Marks PH, Fowler PJ, Munro TG. (1991) Occult posttraumatic osteochondral
lesions of the knee: Prevalence, classification, and short-term sequelae evaluated with
MR imaging. J Radiology 178: 271-276.

Viano DC, Culver CC, Haut RC, Melvin JW, Bender M, Culver RH, Levine RS. (1978)
Bolster impacts to the knee and tibia of human cadavers and an anthropomorphic
dummy. Stapp Car Crash Conf Proc 22: 403-428.

115

CHAPTER FIVE

THE CHRONIC POST-TRAUMATIC EFFECT OF A SINGLE

BLUNT IMPACT IN TISSUES OF THE RABBIT KNEE

ABSTRACT

Blunt joint trauma is a major burden to our society because of its potential for
initiating post-traumatic osteoarthritis (0A), which is associated with chronic debilitation
of patients and long-term medical costs. Clinically, degenerative changes have been
documented in joints two to ﬁve years following severe fractures. In contrast, less severe
injuries may require ten years to manifest changes. To most effectively study the
mechanisms of this disease, in vivo animal models are needed, one of which is the rabbit.
The hypothesis of this study was that blunt trauma to the rabbit TF joint, would result in
accelerated cartilage and subchondral bone changes from those in the non-impacted knee.
The average peak load applied to a group of ten short-term (12 weeks or less) was
approximately 470 N. There were no gross injuries observed at the time of impact or
dissection. Impacted 12 week knee joints had signiﬁcantly less organized calciﬁed
cartilage layer and subchondral bone morphology than non-impacted joints on the medial
side. Additionally, the subchondral plate thickness was increased on both sides for
impacted joints and there was slightly more horizontal splitting of the articular cartilage.
The results of the current study provide insight on how rapidly pathogenesis may develop
in a traumatized joint. However, long-term data is necessary to understand the mechanics

leading up to an end-stage clinical disease in the joint.

116

INTRODUCTION

The number of automotive fatalities is decreasing due to mandated use of seat
belts and airbags (Lund & Ferguson 1995, Viano 1995), however more occupants are
reporting musculoskeletal injuries (Burgess et al. 1995, Dischinger et al. 1992). While the
current lower extremity injury criterion for automotive accidents is based on fracture of
bone (Viano et al. 1977), most injuries are of less severity (Atkinson & Atkinson 2001).
Many clinical studies have revealed a high percentage of patients with early post-
traumatic pain, which eventually can lead to the development of osteoarthritis without a
documented bone fracture (Chapchal 1978, Rassmussen 1972, States 1970). A direct
relationship between trauma and osteoarthritis has been difﬁcult, as diagnosis usually
occurs when the disease reaches its end-stage: complete loss of the articular cartilage
covering a joint. The most common clinical evidence of the disease is radiographic loss
of joint space. Clinically, these changes have been observed in human joints two to ﬁve
years following severe fractures (Wright 1990). In contrast, less severe fractures may
require ten years to manifest changes and may take even longer if the impact does not
generate gross bone fracture.

The ability to diagnose degenerative joint disease in its earliest stages would be
useful in learning to delay or mitigate the development of post-traumatic 0A (Kerin et
al., 2002). Animal studies are necessary to help understand some of the mechanisms of
early joint degeneration. Recent studies have provided more information about the
relationship between a single blunt insult and the pathogenesis of osteoarthritis
(Newberry et al., 1997). It is well documented that a severe, blunt impact will cause

ﬁssures on the surface of articular cartilage (Thompson et a1. 1991, Haut et al. 1992).

117

This damage of the cartilage then accumulates over time and seems to show early signs
of osteoarthritis at six months. Similarly, Donohue et al. (1983) indicated early
histological changes in the articular and calciﬁed cartilage after a blunt impact. They
attempted to show that these changes signify softening of the cartilage, and increased
stress in the underlying bone, producing microcracks and sclerosis that would lead to
osteoarthritis.

Radin et al. (1978) investigated another possible mechanism by using low-level
cyclic loading of the rabbit tibiofemoral (TF) joint. They showed changes in the bone
preceding changes in the overlying cartilage, and at six months documented signiﬁcant
degradation of the joint. In the current study, a blunt force trauma model has been
developed using a rabbit TF joint, which, without generating gross bone fracture may be
used to document early changes in the cartilage, and underlying bone. Based on similar
studies by this laboratory dealing with the patello-femoral joint (Ewers et al. 2000) the
hypothesis of the current study was that a single blunt impact would produce signiﬁcant
changes in the bone and calciﬁed cartilage layer, and be similar to those produced by

Radin following low intensity, cyclic loading of the TF joint.

118

METHODS
Animal Model

The animal model selected for study was the mature (8 month old) Flemish Giant
rabbit. This model has been used for many impact studies in this laboratory (Ewers et al.
2000). A total of 30 animals were used in this study comprising three groups; controls
(non-impacted), low speed, short duration animals and high speed, one year animals. All
animals were exercised for a least two weeks before the impact and throughout the entire
study post impact, except for a one week period following the impact. Exercise consisted
of treadmill hopping for 10 minutes, ﬁve days a week at 0.3 mph. The “short duration”
animals were euthanized between one and 12 weeks post-impact. The results section for
this chapter contains all data collected from the short duration animals. However, one
year and control animals are currently being exercised so there is currently no chronic
data available for these groups. After the one year post impact period expires similar data
will be collected for one year animals. This study has been approved by the Michigan

State University All-University Committee on Animal Use and Care.

Low Speed (Instron) Impact Protocol

A single blunt impact was delivered to one randomly chosen leg of the ten rabbits
in this group. Animals were maintained at a surgical plane of anesthesia using 2%
isoﬂurane. The randomly chosen knee was shaved and the foot wrapped in foam athletic
pre-wrap. A clamp secured the leg while a metal plate connected to a bolt was attached to
the bottom of the foot with ﬁberglass casting tape (J &J Delta-Lite ‘S’). The cast covered

the entire foot except for the ends of the toes and continued proximally to the rnid-calf.

119

The rabbit was laid on its back on a raised platform attached to the table of the servo-
hydraulic testing machine (model # 1331, Instron Corp. Canton, OH). The bolt attached
to the foot was ﬁxed through a hole in the table so that the knee was ﬂexed 90° and the
femur was approximately parallel with the table (Figure 5.1). The knee was positioned so
that it was directly beneath the hydraulic actuator. Then a custom fitting interface was

created with fast drying epoxy to distribute the force over a large area of the knee (Figure

5.2).

 

Figure 5.1. Positioning of the rabbit’s 90° Figure 5.2. Custom epoxy interface in

ﬂexed knee directly beneath the actuator. position on the knee.

120

A single load-controlled haversine waveform impact was delivered to each animal with a
programmed load of 2000 N. The load was programmed to peak in 50 ms, but due to this
fast rate and the relatively compliant response of the tissue actual loads were much lower
than the input value. The average value of the actual load was approximately half the load
required for bone fractures in a pilot study on animal carcasses from another study.
Actual loads were measured with a load transducer (model # lOlOla-2500, Instron Corp.
Canton, OH) and data was collected at 1000 Hz and recorded on a personal computer
with a 16-bit analog/digital board (model DAS 1600: Computer Boards, Mansﬁeld, MA).
Peak load and the time to peak were recorded for each animal. After impact, the animals
were extracted from the testing machine and checked for gross fracture or trauma. The
cast was removed with an autopsy saw with care not to cut the skin. Then the animal was

revived from anesthesia and returned to its cage.

High Speed (Gravity Drop) Impact Protocol

A single blunt impact was delivered to the right leg of ten rabbits using a gravity
accelerated mass. Animals were anesthetized and shaved, as described for the previous
group. The plate, however, was attached to the foot with leather straps that were secured
tightly with Velcro, rather than using casting material. The body and leg positioning was
similar to the previous group with the knee ﬂexed to 90°. Impacts were delivered with a 4
cm square deformable interface (1.2 MPa Hexcel) that insured the impact was equally
distributed over the knee (Figure 5.3). A 1.33 kg mass was dropped from a height of 70
cm to induce a severe level of load to the tibiofemoral joint without causing bone

fractures (Figure 5.4).

121

 

Figure 5.3. Rabbit knee ﬂexed 90° and Figure 5.4. Anesthetised rabbit positioned
secured with Velcro straps positioned for a gravity-accelerated mass impact.

beneath a deformable interface.

The impacting mass was arrested electronically to prevent multiple impacts. A load
transducer (model 31/1432 Sensotec, Columbus, OH) with a 500 lb capacity was attached
behind the impact interface to record the actual load applied to the knee. Experimental
data was collected at 10 kHz on a personal computer. Peak load and time to peak were
recorded for all but 4 animals (electronic problem in triggering of data collection
software). The animal was extracted from the impact ﬁxture and checked for gross

injuries before it was revived from anesthesia and returned to its cage.

122

Tissue Quality Quantification

Following euthanasia of each animal both knees were surgically dissected and all
gross joint trauma was recorded. Additionally the retropatellar, tibial plateau and anterior
and posterior femoral surfaces were stained with India ink and photographed at 25x using
a Wilde microscope and digital camera. The control and impacted tibia plateaus were
placed in phosphate buffered saline and the cartilage on the medial and lateral sides was
mechanically tested in a materials testing machine (Figure 5.5). The structural integrity of
the cartilage was determined with stress-relaxation indentation tests (Newberry et al.,
1997). Brieﬂy, each tibia was placed in a clamp attached to a camera mount (Bogen,
Ramsey, NJ) that was secured to an x-y translational base plate. These features allowed
the tibiae to be positioned so that the indentation tests were perpendicular to a ﬂat
location on the surface. Indentations were performed with a computer controlled stepper
motor (model M-l68.30 Physik Instruments, Waldbrom, Germany). A ﬂat, rigid, non-
porous, 0.5 mm diameter probe was pressed 0.1 mm into the cartilage in 30 ms and
maintained for 150 seconds. The resistive loads were recorded from a load transducer
(model JP-25 Data Instruments, Acton, MA) attached between the probe and the actuator.
Data was collected with a personal computer at 1000 Hz for the ﬁrst second and 20 Hz
for the remainder of the test. The response to indentation was a sharp spike in load
followed by a period of load relaxation. Five minutes following the indentation test the

thickness of the site was measured by depressing a needle into the cartilage.

123

.
4‘
03‘
*.
I”.
“F?
‘.
E11
'1

 

Figure 5.5. Indentation material testing machine.

The stiffness of the cartilage was determined from the results of the indentation and
thickness tests by a calculation of a shear modulus from an assumed elastic layer bonded
to a rigid half space (Hayes et al. 1972) The instantaneous (Gu) and relaxed (Gr) moduli
were calculated based on the load at 80 ms and 150 s, respectively, using equation 1.
These moduli were then normalized to the opposite, non-impacted control joint to
minimize variation between animals.

P(l — V)
4aKw

G = (5.1)

Where P is the measured load, v is Poisson’s ratio (assumed to be 0.5 for Gu and 0.4 Gr),
3 is the indenter radius, w is the penetration depth, and K is a scaling factor that depends

on the thickness and Poisson’s ratio.

124

Histological analysis

Tibial plateaus were placed in 10% buffered formalin for seven days and
decalciﬁed in 20% formic acid for another seven days. Tissue blocks were cut in a media]
to lateral direction on each side of the plateau in the region of tibiofemoral contact. The
blocks were processed for routine parafﬁn embedding. Four sections, eight urn thick,
were cut across each side of the plateau and stained with safranin O and fast green, using
standard techniques (Atkinson et al. 1998). The sections were examined under light
microscopy at 25-400 magniﬁcation, and scored by histomorphometric analysis. A table
of the categories and scoring system is given in Table 5.1. The highest quality slide was
chosen from each side of the plateau by the histological technician (J. Walsh) and this
slide was blindly scored by the technician and two graduate students (E. Meyer, A.
Meram) in all categories. The thickness of the cartilage, calciﬁed cartilage and

subchondral bone were also measured at 40x with a calibrated eyepiece.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mm mm 0 mm W o
SWIM 1 mandamus) 1
mama-ad 2 Foams 2
Focalyuswed 3 0111001000 3
ﬂ 5091mm 4 ! realms 4
WWW None 0 WWW um
1-38041‘000 1 WWW Nam-I o
12an 2 311311 1
34m 3 m 2
“Mann 4 Farah/Baas 3
“mm 4 We 4
mean Mini 0 mama-11109091101 m 0
91311.00; 1 arm 1
mass 2 m 2
Foaly 3 Dark 3
Taaltaaa 4 WWW uits
mm“: Nana 0 WWW Dam 0
301100000 1 50139104909000 1
macros 2 mm 2
8010015810 3 50110831119 3
mm 4 14113038109 4
moan 4 new None 0
WWW None 0 Few 2
011mm 2 Mum 4
11013313190113 4
W810: 4
WWW wits

 

 

Table 5.1. Histomorphometric scoring table

125

Following histomorphometric scoring the slides were digitally photographed at
25x and trabecular bone porosity was measured in a standard area with an image analysis
software package (SigmaScan, Jandel Scientific, San Rafeal, CA) for each picture by

dividing the area of the pores by the total area (Figure 5.6).
Medial . Lateral

 

Articular Cartilage
Subchondral Bone

Trabecular Bone

Epiphyseal Plate

Cortical Bone

Bone Marrow Space

Figure 5.6. Schematic of a histology slide of the rabbit tibial plateau with

regions for porosity measurement of the trabecular bone.
Statistics were performed by paired t-tests (data passed test for normality) to
determine signiﬁcant differences between the groups of impacted and non-impacted

joints (signiﬁcance for p<0.05). All data shown as mean 1 one standard deviation.

RESULTS

Low speed impacted animals were divided into two groups based on the time

duration after impact when they were euthanized. The ﬁrst group consisted of all animals

126

euthanized before 12 weeks, and the second group were euthanized at exactly 12 weeks.
The average peak load achieved in all the impacts was 469.6 1 123 N, and there was not a
statistical difference between the peak loads applied between groups (<12 week animals
429 1 75 N, 12 week animals 510 1 156 N). This load was approximately 50% of the
peak load required to cause gross bone fracture from a pilot study on rabbit cadavers
(Appendix). There were no gross injuries (bone fracture or ligament tears) observed at the
time of impact or dissection in any of the animals.

Gross examination after dissection of the tibio-femoral joint revealed more
cartilage ﬁssures on the medial side then the lateral side, but similar amounts between
impacted and non-impacted knees. There were a few small meniscal tears on the lateral
plateau, however since these were not always on the impacted side, they were probably

not caused by the impact (Figure 5.7).

   

Figure 5.7. Specimen RS0608 gross dissection photographs of cartilage ﬁssures and

small meniscal tears on the tibia plateau.

The cartilage on the medial and lateral sides of the tibia plateau was mechanically
indented to measure the instantaneous and relaxed moduli, as well as the cartilage

thickness. For the 12 week group the impacted and nonimpacted data was combined in a

127

 

 

ratio, to eliminate variations between animals (Figure 5.8). Both the medial and lateral
sides of the impacted joints showed a slight increase in cartilage thickness over controls,
while the mechanical properties of the lateral side decreased and the medial side
increased slightly. There were no statistical differences between medial and lateral sides
or for impacted versus controls, probably due to the small sample size of only ﬁve

animals in this group.

 

1.6 -
I Lateral l Medial

1.4-

 

 

 

 

0-8 ‘ Thickness

Ratio (impacted/control)

 

0.6 ~

 

0.4 J

 

 

 

Figure 5.8. Results from mechanical indentation testing of the tibial plateaus of 12
week animals.

Histologically, there were signiﬁcantly more ﬁssures (p<0.001) and geometry
irregularities (p=0.014) in the cartilage on the medial side than the lateral side. The
articular cartilage was also thicker on the medial side in both impacted and non-impacted
knees. In addition, impacted 12 week animals had a less distinct tidemark and less
organized subchondral bone morphology on the medial side (p=0.012) (Figure 5.9). The

subchondral bone plate thickness was statistically increased on the medial (p=0.02) and

128

 

lateral (p=0.01) impacted sides (Figure 5.10). These animals also had slightly more
horizontal splitting of the articular cartilage in the impacted rabbit joints. The
histomorphometric scoring system showed statistically more pathologic cells in the
articular cartilage on the medial impacted side by the presence of excessive clones and

clusters (p=0.024).

 

2 00 - Medial Impacted
I Medial Non-impacted
I Lateral Impacted

I Lateral Non-impacted

Histomorphometric Score

 

 

Subchondral Bone
Morphology

Articular Cartilage

Cells Calcmed Cart1lage

 

 

 

Figure 5.9. Results of histomorphometric scoring of 12 week rabbits.

* Indicates there was a statistical difference from the non-impacted knees.

    

 

,7. \‘

' ’..1‘l '
o

.117 3 e

Non-impacted . 5 RSO97O LtMed , Impacted 2512809 ORtMetd

Figure 5.10. Increased subchondral bone plate thickness and horizontal splitting in the

articular cartilage following impact.

Trabecular bone porosity was measured for a constant area in all 12 week animals

(Figure 5.11). While there was not a signiﬁcant difference between impacted and non-

129

impacted knees on either the medial or lateral side, the impacted medial side had the
highest porosity of 47.7 1 5.9% and the non-impacted side was lower with 39.7 1 3.2%
(p=0.06). In contrast, the porosity of the impacted lateral side was lower (36.9 1 7.5%)
than the non-impacted side (43.3 1 12.9%), but these data were not statistically

signiﬁcant.

60.0% -
50.0% -
40.0% 1
30.0% a

20.0% .

Bone Porosity

10.0% -

 

 

0.0% -

 

Medial Medial Non- Lateral Lateral Non-
lmpacted impacted impacted impacted

Figure 5.11. Trabecular bone porocity of 12 week impacted and non-impacted knees.
* Statistically different than Non-impacted

DISCUSSION

Blunt trauma to the TF joint resulted in an accelerated disease process in the joint
by subchondral bone remodeling, horizontal splitting in the deep layer of articular
cartilage, and the presence of chondrocyte clusters. This study provides evidence. that a
single impact may induce degenerative changes similar to those seen by Radin et al.
(1984) with low level, cyclic TF compression. The previous study of Radin et al.,

documented a slight trend for decreased subchondral bone porosity with time, which they

130

 

 

concluded resulted in an increase in the stiffness of this region underlying the articular
cartilage. Another study (Dedrick et al. 1997) that investigated 0A progression in a
canine model showed the opposite effect, documenting an increase in the subchondral
bone porosity. Our study showed that the impacted subchondral bone became denser and
thickened at 12 weeks. Subchondral bone plate thickening was also documented in the
impacted patellofemoral (PF) joint by Ewers et al. (2002) at 7.5 months and Mazieres et
al. (1987) by the third month. This is in agreement with the results of the current study.
However, in contrast to the Radin et al. (1984) and Mazieres et al. (1987) studies, the
animals in our experiment were regularly exercised. Exercise has been shown to have a
beneﬁcial effect on the impacted joint in the PF model (Weaver, 2001). In the previous
study degenerative changes in the impacted PF joints were mitigated (or slowed) in the
exercise group versus a cage-activity group.

In the current study degenerative changes were also present in the articular
cartilage on the medial side of the TF joint by the presence of chondrocyte clones and
clusters, a phenomenon indicative of early 0A. The results of the current study provide
insight on how rapidly pathogenesis may deve10p in a traumatized TF joint. However,
data must be collected from the long-term animals after one year to understand the
mechanics leading up to an end-stage clinical disease in the joint. Future studies should
also investigate the effect of other variables such as exercise and various drug therapies

on long-term degradation of diarthrodial joints following a single blunt impact.

131

 

ar——— _ Ame-yum..—

REFERENCES

Atkinson PJ, Walsh JA, Haut RC. (1998) The Human Patella: A Comparison of Three
Preparation Methods. J Histotechnology 21: 151-153.

Burgess A, Dischinger P, O'Quinn T, Schmindhauser C. (1995) Lower Extremity Injuries
in Drivers of Airbag-equipped Automobiles: Clinical and Crash Reconstruction
Correlations. J Trauma 38: 509-516.

Chaphal G. (1978) Post-traumatic Osteoartritis After Injury of the Knee and Hip Joint.
Reconstr Surg Trauma 16: 87-94.

Dedrick DK, Goulet RW, O’Connor BL, Brandt KD. (1997) Preliminary Report:
Increased Porosity of the Subchondral Plate in an Accelerated Canine Model of
Osteoarthritis. Osteo and Cart 5: 71-74.

Dishinger P, Cushing B, Kems T. (1992) Lower Extremity Fractures in Motor Vehicle
Collisions: Inﬂuence of Direction, Impact and Seatbelt Use. AAAM 36: 319-326.

Donohue JM, Buss D, Oegema TR, Thompson RC. (1983) The Effects of Indirect Blunt
Trauma on Adult Canine Articular Cartilage. J Bone Jt Surg 65-A: 948-957.

Ewers BJ, Weaver BT, Haut RC. (2002) Impact Orientation Can Signiﬁcantly Affect the
Outcome to the Rabbit Patellofemoral Joint. J Biomech 35: 1591-1598.

Haut RC, Ide TM, Walsh JC, Robbins JL, DeCamp CW. (1992) Studies on the Blunt
Impact Response of Articular Cartilage: Animal Experiments, Human Cadaver
Experiments, and Mathematical Modeling. IPTBS 1: 13-26.

Kerin A, Patwari P, Kuettner K, Cole A, Grodzinsky A. (2002) Molecular Basis of
Osteoarthritis: Biomechanical Aspects. Cell Mol Life Sci 59: 27-35.

Lund A, Ferguson S. (1995) Driver Fatalities in 1985-1993 Cars with Airbags. J Trauma
38: 469-475.

Mazieres B, Blankaert A, Thiechart M. (1987) Experimental Post-contusive
Osteoarthritis of the Knee: Quantitative Microscopic Study of the Patella and the
Femoral Condyles. J Rheum 14: 119-122.

Newberry WN, Zukosky DK, Haut RC. (1997) Subfracture Insult to a Knee Joint Causes
Alterations in the Bone and in the Functional Stiffness of Overlying Cartilage. J
Orthop Res 15: 450-455.

Radin EL, Parker HG, Pugh JW, Steinberg RS, Paul IL, Rose RM. (1973) Response of
Joints to Impact Loading 1H: Relationship Between Trabecular Microfractures and
Cartilage Degradation. J Biomech 6: 51-57.

132

 

 

Radin EL, Martin RB, Burr DB, Caterson B, Boyd RD, Goodwin C. (1984) Effects if
Mechanical Loading on the Tissues of the Rabbit Knee. J Orthop Res 2: 221-234.

Rasmussen PS. (1972) Tibial Condylar Fractures as a Cause of Degenerative Arthritis.
Acta Orthop Scand 43: 566-575.

States JD. (1970) Traumatic arthritis- A Medical and Legal Dilemma. AAAM Proc 14:
21-28.

Thompson RC, Oegema TR, Lewis JL, Wallace L. (1991) Osteoarthritic Changes After
Acute Transarticular Load: An Animal Model. J Bone Jt Surg 73-A: 1990-1001.

Viano DC. (1977) Considerations for a Femur Injury Criterion. Stapp Conf Proc 21: 445-
473.

Viano DC. (1995) Restraint Effectiveness, Availability and Use in Fatal Crashes:
Implications to Injury Control. J Trauma 38: 538-546.

Weaver BT. (2001) The Analysis of Tissue Response Following A Single Rigid Blunt
Impact in an in vivo Animal Model: Regular Exercise is Beneﬁcial in a Stable Joint
After Trauma. MS Thesis: 33-50.

Wright V. (1990) Post-traumatic osteoarthritis- A medico-legal mineﬁeld. J Rheum 29:
474-47 8.

133

 

 

CONCLUSION

This thesis continues the efforts of the Orthopeadic Biomechanics Laboratory to
understand mechanisms of osteoarthrosis and investigate methods of diagnosis,
intervention, and prevention of this disease. Other investigators have shown that post-
traumatic joint degradation is a signiﬁcant problem in our society. However, the
progression mechanisms are not very well understood, especially in cases with no gross
fracture of bone.

Current automotive industry standards for crash tests of new vehicles require that
femur load not exceed 10 kN. This injury criterion is based on gross bone fracture that
was documented in human cadaver biomechanical experiments (Melvin et al. 1975,
Patrick et al. 1965, Powell et al. 1975) similar to Chapter 1 in the current study. The
purpose of doing experiments similar to previous studies is that one injury mechanism,
which was suggested in those early, broad studies, may be isolated and investigated in
higher detail. Another motivation is to look for additional occult injuries that may occur
before gross bone fracture, but were not documented previously. These occult injuries
such as cartilage ﬁssures and rrricrofractures of the underlying bone are thought to be
precursors to chronic joint degradation. If a explicit correlation was revealed between
these occult injuries and development of osteoarthritis then it would suggest that the
automotive industry should reevaluate their injury criterion for one that could better
protect passengers against this disease.

Chapter 2 documented that the patella is more vulnerable to fracture with a rigid
interface when the knee was impacted obliquely versus axially with respect to the femur.

Additionally, the orientation of the fracture pattern changed from a horizontal linear or

134

 

5”,.

 

comminuted fracture for axial impacts to a vertical fracture across the patella for oblique
impacts. In two cases of oblique impact on the knee the patella dislocated laterally prior
to fracture. Since clinical studies often show an association of patella fractures and
dislocations with the onset of chronic disease in the knee and a large portion of drivers
have been shown to sit with their legs abducted, these patients appear to have increased
long-term risk from oblique than axial impacts on the knee.

Chapter 3 documented that ACL rupture could occur in the human knee prior to
gross bone fracture for axial tibia load when displacement between the tibia and femur is
not constrained. Anterior and lateral displacements of the tibia with respect to the femur
as well as internal rotation of the tibia were recorded at the time of ACL failure. Since the
compressive loads in the tibia were in the range of what could be generated during a jump
landing, this may be a mechanism in many “non-contact” ACL tears.

Chapter 4 documented that axial tibia load, similar to what was investigated in
Chapter 3, has a stiffening effect on the response of the knee to anterior knee impacts. At
subfracture level loads the translation between the tibia and femur and the patello-femoral
contact force were reduced. At failure level loads it took a slightly higher anterior knee
load to cause PCL trauma for knees with a simultaneous axial tibial load. This load
increase was comparable to the load that was required to constrain TF motion in a
previous study (Jayaraman et a1. 2001). Since axial compressive loads in the tibia are
documented in crash simulations at the same time and magnitude as anterior knee loads,
the knee may actually have a stiffer response than previously thought.

Chapter 5 documented the short-term degenerative changes of a rabbit tibio-

femoral joint to blunt impact. Histologically, chondrocyte clones and clusters were seen

135

 

 

 

in the medial articular cartilage of impacted knees by 12 weeks. There was also
degeneration of the calciﬁed cartilage layer and subchondral bone morphology in both
medial and lateral sides. Additional studies are in progress for one-year post impact.
These will better document the mechanics of chronic joint degradation leading up to an

end-stage clinical disease of the joint.

 

 

 

136

   

11111111111101111111111